Correlator3D User Manual

Correlator3DTM User Manual
Table of Contents
1. Important Notes .....................................................................................................................................5
2. Overview ................................................................................................................................................6
3. System Requirements .............................................................................................................................7
4. Data Requirements .................................................................................................................................8

4.1 Sensor Types ............................................................................................................................... 8
4.2 Input Image Formats .................................................................................................................. 10
4.3 DEM Format............................................................................................................................... 10
4.4 Orthophoto Format..................................................................................................................... 10
4.5 Vectors ....................................................................................................................................... 11
4.6 Coordinate Systems................................................................................................................... 11
5. Visual Interface ....................................................................................................................................13
5.1 Multilingual User Interface.......................................................................................................... 13
5.2 Project Tree ............................................................................................................................... 13
5.2.1 Images .................................................................................................................................14
5.2.2 15
5.2.3 GCP .....................................................................................................................................15
5.2.4 AT ........................................................................................................................................15
5.2.5 DEM .....................................................................................................................................15
5.2.6 Point Clouds ........................................................................................................................16
5.2.7 3D Models............................................................................................................................17
5.2.8 Orthos ..................................................................................................................................17
5.2.9 Mosaic .................................................................................................................................17
5.2.10 Vectors.................................................................................................................................17
5.2.11 Reference Orthos ................................................................................................................18
5.2.12 Contours ..............................................................................................................................18
5.2.13 Online Data ..........................................................................................................................18
5.3 Properties................................................................................................................................... 18
5.4 Zooming and panning ................................................................................................................ 18
5.5 Image Band Setup ..................................................................................................................... 19
5.6 Picture-in-Picture 3D Viewing .................................................................................................... 20
5.7 Scale .......................................................................................................................................... 20
5.8 Status Bar .................................................................................................................................. 21
5.9 Profile ......................................................................................................................................... 21
5.10 Volume ....................................................................................................................................... 22
6. Project Creation ...................................................................................................................................24
6.1 UAV / Medium / Large Format ................................................................................................... 25
6.1.1 Step 1 - Image Selection......................................................................................................25
6.1.2 Step 2 – Camera Parameters ..............................................................................................25
6.1.3 Step 3 – Project Parameters ................................................................................................26
6.2 Satellite ...................................................................................................................................... 26
6.2.2 Step 2 - Image Pairing .........................................................................................................27
6.2.3 Step 3 – Project Parameters ................................................................................................27
6.3 ADS ........................................................................................................................................... 27
6.3.2 Image Pairing.......................................................................................................................28
6.3.3 Project Parameters ..............................................................................................................28
6.4 A3 .............................................................................................................................................. 28
© 2019 SimActive Inc. All rights reserved. 2

6.4.2 Project Parameters ..............................................................................................................28

7. Data Generation ...................................................................................................................................29
7.1 Aerial Triangulation .................................................................................................................... 29
7.1.1 Tie Point Extraction ..............................................................................................................29
7.1.2 GCP Creation ......................................................................................................................30
7.1.3 Tie Point Editing ...................................................................................................................30
7.1.4 Bundle Adjustment ...............................................................................................................32
7.1.5 Quality Assessment .............................................................................................................34
7.1.6 Image Filtering .....................................................................................................................35
7.1.7 Coarse DEM Extraction .......................................................................................................36
7.2 DSM Generation ........................................................................................................................ 36
7.3 DTM Extraction .......................................................................................................................... 37
7.4 DEM Editing ............................................................................................................................... 39
7.4.1 Standard Selection ..............................................................................................................39
7.4.2 Append a Selection ..............................................................................................................41
7.4.3 Touch Up .............................................................................................................................41
7.4.4 Crop .....................................................................................................................................41
7.4.5 Delete ..................................................................................................................................42
7.4.6 Set Elevation........................................................................................................................42
7.4.7 Filter .....................................................................................................................................42
7.4.8 Delete and Fill ......................................................................................................................43
7.4.9 Topological Fill .....................................................................................................................43
7.4.10 DTM Extraction ....................................................................................................................43
7.4.11 Paste DEM...........................................................................................................................44
7.4.12 DEM History (Undo/Redo) ...................................................................................................44
7.5 Orthorectification ........................................................................................................................ 45
7.6 Mosaic Creation ......................................................................................................................... 46
7.7 Mosaic Editing............................................................................................................................ 48
7.7.1 Export Seamlines.................................................................................................................49
7.7.2 Adjusting the Colors .............................................................................................................49
7.7.3 Seamline Editing Mode ........................................................................................................50
7.8 3D Model Generation ................................................................................................................. 53
8. Automated Workflow ..........................................................................................................................55
9. Feature Extraction ................................................................................................................................57

9.1 Extraction Modes ....................................................................................................................... 57
9.1.1 Building Extraction Mode .....................................................................................................58
9.1.2 Road Extraction Mode .........................................................................................................60
9.1.3 Waterbody Extraction Mode.................................................................................................60
9.1.4 Forest Extraction Mode ........................................................................................................60
9.2 Attributes .................................................................................................................................... 61
9.3 Real-time True Orthorectification ............................................................................................... 61
10. Tools .....................................................................................................................................................62
10.1 Import EO Data .......................................................................................................................... 62
10.2 Seed DEM Creation ................................................................................................................... 63
10.3 Fiducial Detection ...................................................................................................................... 64
10.4 GCP Creation............................................................................................................................. 66
10.4.1 Creating and Importing GCPs ..............................................................................................66
10.4.2 GCP Editing Dialog ..............................................................................................................66
10.5 DEM Merging Tool ..................................................................................................................... 69
10.6 Export EO and Camera Parameters .......................................................................................... 70
10.7 Export Mosaic ............................................................................................................................ 70
10.8 Export DEM................................................................................................................................ 72
10.8.1 Export DEM Tiles .................................................................................................................72
10.8.2 Export LIDAR LAS and ESRI Shapefile ...............................................................................73

10.8.3 Export with Geoid ................................................................................................................73

10.9 DEM Inspection.......................................................................................................................... 74
10.10 Change Detection ...................................................................................................................... 75
10.11 Point Cloud Colorization ............................................................................................................ 76
10.12 DEM Contour Extraction ............................................................................................................ 76
11. Script Mode ..........................................................................................................................................78
11.1 Command Prompt ...................................................................................................................... 78
12. Troubleshooting ...................................................................................................................................80
12.1 Log Window and Log File........................................................................................................... 80
12.2 Good Practice ............................................................................................................................ 80
12.3 Temporary Files ......................................................................................................................... 80
13. Appendix A ...........................................................................................................................................82
13.2 Tile Definition File ...................................................................................................................... 83
13.3 Script Format Description........................................................................................................... 84

1. Important Notes
This symbol indicates an important notice. Failing to follow the notice

could lead to unexpected or incorrect results.
This symbol indicates a tip or helpful information. Following this

advice will lead to better results and performance.
This manual uses the following abbreviations as summarized in Table 1:
DEM Digital Elevation Model

DSM Digital Surface Model
DTM Digital Terrain Model
GSD Ground Sampling Distance
LUT Look-Up Table
EO Exterior Orientation
GCP Ground Control Points
RPC Rational Polynomial Coefficient
Table 1: Abbreviations.

2. Overview
Correlator3D™ software is a patented end-to-end photogrammetry solution for the
generation of high-quality geospatial data from satellite and aerial imagery, including UAVs.
It builds on unique computer vision algorithms that significantly differ from traditional
photogrammetry techniques. Powered by GPU technology and multi-core CPUs,
Correlator3D™ ensures matchless processing power to support rapid production of large
datasets.
Correlator3D™ consists of the following modules:
- Aerial Triangulation: exterior orientation correction and camera calibration

- DSM Generation: generation of digital surface models and point clouds
- DTM Extraction: extraction of digital terrain models
- DEM Editing: manual editing of DEMs
- Orthorectification: creation of orthophotos and true orthos
- Mosaic Creation: stitching orthophotos together creating a mosaic
- Mosaic Editing: manual editing of mosaic seamlines
- 3D Model Generation: generation of textured 3D models
- Feature Extraction: semi-automated collection of 3D features
The software also provides other tools including:
Automated Workflow: automatic execution of Correlator3D™ workflow

GCP File Creation: semi-automatic tool for creating and editing GCPs
NDVI Map Creation: compute the normalized difference vegetation index
Change Detection: identification of 3D changes
DEM Merging: stitching of DEM patches
DEM Inspection: generation of statistics about DEM quality
Point Cloud Colorization: colorization of a point cloud using a mosaic
DEM Contour Extraction: extract contour lines from a DEM
DSM DTM Mosaic 3D Model
Figure 1: Sample geospatial data.
The following sections describe, in detail, the different functionalities, uses, and guidelines
of the software, to ensure optimal results.

3. System Requirements
Correlator3D™ has been designed to operate on standard PCs. However, it consists of
state-of-the-art software and as such, requires a recent computer. A great portion of the
processing is performed directly on the graphics card, which is part of what allows
Correlator3D™ to generate geospatial data so rapidly. The supported cards must feature
an NVIDIA® GeForce graphics processor.
It is therefore recommended to use Correlator3D™ on a PC that meets or exceeds the

specifications in Table 2:
Component Description
CPU Intel® Core™ i5 processor or higher
GPU Graphics card with an NVIDIA® GeForce 9 GPU (G92) or higher, with 1GB
RAM or higher (note that some Quadro, Tesla, and Tegra cards also feature
a GeForce GPU), or with an AMD Radeon™ R9 GPU or higher, with 1GB
RAM or higher.
RAM 8GB or higher
OS Windows 7, 8 and 10. Correlator3D™ is a 64-bit application.
Table 2: Recommended specifications for running Correlator3D™.

4. Data Requirements
Correlator3D™ requires different types of data to be able to perform its processing and
can produce several kinds of outputs. As a result, it expects and generates data using
specific data types. Table 3 summarizes the supported file formats.
Data Type Supported Formats File Extensions

DEM SimActive Model Format (SAMF) .smf2, .smf
GeoTIFF .tif
Arc/Info ASCII Grid .asc
3D Constraint ESRI Shapefile .shp
Microstation DGN .dgn
Image Types TIFF / BigTIFF / GeoTIFF .tif
JPEG Compliant .jpg
JPEG2000 .jp2
Windows Bitmap .bmp
NITF .ntf
Mosaic Definition File SimActive Format .sbd
Tile Definition SimActive Format .tdf
Ground Control Points SimActive Format .gcp
Orthorectified Images GeoTIFF (BigTIFF) .tif,
TIFF (BigTIFF) World Format .tif, .tfw
Support File for ADS SOCET SET support file .sup
RPC for Satellite RPC Model file .rpc, .rpb, .xml, .txt
Image Tiles for Satellite Tile Model file .til
DIMAP .xml
SimActive Script SimActive Script Format .spt, .txt
Geoid GTX Format .gtx
Project File SimActive Format .c3d
3D Features ESRI Shapefile .shp
3D Model OBJ Format .obj
SimActive 3D Model Format .tmf
Contours ESRI Shapefile .shp
AutoCAD 2000 DXF .dxf
Interior/Exterior SimActive Format .ieo
Orientation File
Table 3: Supported file formats.
4.1 Sensor Types

Correlator3D™ supports the following types of sensors:
 UAV / Medium / Large format camera systems
 Satellite sensors using the RPC model
 Leica’s ADS pushbroom sensors
 VisionMap A3 camera systems

These four sensor types function in a similar manner, but require different parameters
when creating a project. The following flowchart describes the workflow of
Correlator3D™. Note that the workflow is sensor-dependent.
UAV/Medium/Large Format Satellite ADS VMA3
Project Creation Project Creation Project Creation Project Creation
Aerial Triangulation
3D Model Generation
DSM Generation
Feature Extraction
DTM Extraction
DEM Editing
Orthorectification
Mosaic Creation
Mosaic Editing
Figure 2: Correlator3D™ workflow.

4.2 Input Image Formats

Supported formats are summarized in Table 4.
Image Type Data Format

TIFF, BigTIFF and GeoTIFF 8, 16 or 32 bits per band
1 to 32 bands
JPEG in TIFF 8 or 12 bits per band
1, 3 or 4 bands
LZW in TIFF 8, 16 or 32 bits per band
1 to 32 bands
JPEG 8 bits per band
1 or 3 channels
JPEG2000 8 or 16 bits per band
1, 3 or 4 bands
NITF 8 or 16 bits per band
1, 3 or 4 bands
Windows Bitmap 8 bits per band
1, 3 or 4 bands
Table 4: Supported input image formats
TIFF and GeoTIFF images are supported in both interleaved (RGBRGB) and pixel ordered
(RRGGBB) formats.
4.3 DEM Format

The DEM should have square cells for both the GeoTIFF and Arc/Info ASCII Grid formats.
This means that cell dimensions in X and Y should always be the same. Also, elevation
values in GeoTIFF files should be expressed as 32-bit floating point numbers.
4.4 Orthophoto Format

Orthophotos can be generated in the Tiled TIFF format with overviews to maximize system
performance during their subsequent exploitation using GIS software. However, the user
should keep in mind that in general, more basic image viewing software do not support
images in the Tiled TIFF format with overviews. Orthophotos are produced in one of the
following formats:
8, 16 or 32 bits per band

1 to 32 bands

4.5 Vectors
Existing vectors can be supplied to the software for different purposes. They can be used
as 3D constraints for the DSM generation and DTM extraction modules to force specific
elevation values in the output DEM. They can also be used as an input to the
orthorectification process to generate true orthos. Vectors can also be used as a reference
in the software. For example, some projects can be divided in multiple blocks. Having an
overlay displaying a reference grid can then be useful to guide the editing process.
Vector files must be provided in the ESRI Shapefile or Microstation DGN and contain
points, lines or regions. If vectors are to be used as a 3D constraint but that the features
do not have elevation data (z-component), users will be prompted to specify a default
elevation value or to specify the name of the attribute to get the elevation from (e.g.
“HEIGHT”).
Figure 3: Vector elevation dialog.
4.6 Coordinate Systems

Correlator3D™ requires the projection to be specified when creating a project. The user
specifies the projection details as shown in Figure 4.

Figure 4: Projection dialog.
Four options are presented to the user. A predefined list of the most common projections
is available. Otherwise, the user may search for a specific projection system by entering
the corresponding keywords or EPSG code. The user can also specify a custom projection
or specify a Proj.4 definition. If using the Proj.4 option, please refer to this website for
specific values: http://spatialreference.org/
Although mandatory for satellite, ADS and A3 sensors, the projection is optional for
UAV/medium/large format sensors, and can be ignored by pressing the “Ignore” button as
shown in Figure 4. Note that the rotation system supported by Correlator3D™ is
photogrammetric (omega, phi and kappa) and that the geographic coordinate system (i.e.
latitude/longitude) and navigational angles (i.e. roll, pitch and yaw) are converted when
importing EO data during project creation. If the user chooses to ignore the projection,
then no projection information will be saved in the GeoTIFF files.
Proj.4 values representing a geographic coordinate system (latitude/longitude) are not

supported.

5. Visual Interface
Correlator3D™ is designed for ease of use, featuring a simple visual interface. The main
window serves to load, save, and display DEMs, mosaics, orthos, images, GCPs, vectors
and AT results. Additionally, it provides access to the different software functionalities and
tools. At the core enabling this user experience is the project tree, further described in the
corresponding section.
Figure 5: Correlator3D™ main window.
5.1 Multilingual User Interface

Correlator3D™ can be operated in different languages, including English, French,
Portuguese, Spanish, Japanese, and Chinese. The user can select the language at any time
by accessing the corresponding option in the General Settings dialog from the File menu.
Note that the language change is only applied once the software is restarted.
5.2 Project Tree

The project tree displays the project information as a hierarchical structure on the user
window. The tree is used to perform multiple functions: modify the visibility of an overlay
using a checkbox, save results, add data to the project, etc. It allows the user to seamlessly
transition between data within each project. When the cursor is placed above an item listed
in the project tree, contextual buttons appear. The following subsections describe each
aspect of the project tree.

5.2.1 Images
At the top of the project tree are the “Images”. This item is filled with parameters given
during the project creation. When selecting the “Images”, the camera(s) used in the project
appear with their corresponding image footprints if a DEM is loaded. It is possible to have
more than one camera within each project, and consequently, within the project tree.
Double-clicking on a camera will allow the user to see its associated flight line(s). The user
has the option to select which flight lines or images are to be processed in corresponding
steps using the contextual buttons. With this, flight line and images can be disabled; they
will not be considered while generating a DSM or orthos. Moreover, the checkboxes are
used to toggle visibility in the main window of the corresponding item (image, flight lines,
or complete image set).
In the AT module, all images included in the project are considered; enable/disable have no
impact and cannot be changed in this mode.
When a DEM is loaded, the image footprints on the ground are displayed. This function
can be turned on and off at the user's preference.
When images are selected using the tree, the image positions are displayed as arrows
overlaid on the DEM. This is very useful when ensuring that both the imagery and 3D data
span the same region. Note that each arrow indicates the top of the image. Flight lines
and their direction are also displayed as dashed lines with arrows.
When the cursor is placed over an arrow, a tooltip shows the name of the corresponding image.
The user can also right-click on the image position to display the image in a separate window.
The image will be rotated following its kappa-angle value. Panning is performed by clicking and
dragging with the middle (wheel) mouse button while zooming can be performed using the mouse
wheel. Note that several images can be displayed simultaneously.
Within the aerial triangulation module, the arrows may be replaced by color disks to allow a
quick assessment of the quality of the bundle adjustment process. Please refer to section 7.1.5
for details.

Figure 6: Image window.
5.2.2
5.2.3 GCP
An existing GCP file can be added to the project directly using the project tree. There is
also the option to remove the GCP file from a project. In order to create a GCP file, the
user can go into the AT module. Similar to the aforementioned option, the project tree can
be used to toggle the GCP visibility using the checkboxes.
Correlator3D also allows the user to specify check points. GCPs and check points are
stored in the same GCP file, but the project tree displays them in two distinct subcategories
of GCPs. GCPs and check points are only considered when performing bundle adjustment,
but can be used to visually assess quality of the DSM, DTM, and mosaic. Please refer to
sections 7.1.2 and 10.4 for further details.
5.2.4 AT
Please refer to section 7.1 for details.
5.2.5 DEM
The project tree simplifies working with DEMs as a DEM can be loaded directly onto it.
Select “Add a DEM to project”, by placing the cursor over the “DEM” item in the project
tree. More than one DEM can be a part of the project. However, at any given time, only
one DEM can be displayed. The ability to toggle between multiple DEMs is possible

through the project tree using the checkbox for each DEM. It is also possible to remove a
DEM by clicking on the “X” when placing your mouse over the respective DEM.
Once the data is read, it will be displayed in the window. Two display styles are available
for viewing elevation data: color and grayscale.
If the color display style is selected, elevation values are mapped from blue to red (through
green and yellow) according to the minimum and maximum values found in the data: low
elevation values will be displayed in blue or green, middle values in yellow and high values
in red.
If the grayscale display style is selected, elevation values are linearly mapped from black to
white according to the minimum and maximum values found in the data: low elevation
values will be displayed in dark colors while high values will appear in light colors.
The display style can be changed by selecting “Display Style->Color” or “Display Style-
>Grayscale” in the “View” menu.
Colors can be mapped from the minimum and maximum elevation values visible on the
screen. The latter case is called “stretch” mode and can be enabled or disabled by selecting
“Display Style->Stretch” in the “View” menu.
Subtle variations in elevation values can be highlighted by selecting the “Shade DEM
variations” feature, which simulates lighting and shadows. This feature can be enabled or
disabled by selecting “Display Style->Shade” in the “View” menu.
When a DEM file is loaded, a legend is visible at the left of the main window. This legend
shows the color mapping currently used for displaying the elevation data. Minimum and
maximum values are displayed with a line indicating the “0” value, if relevant. When the
cursor is moved over the DEM, a horizontal bar moves up and down over the legend
providing a visual cue for the elevation under the cursor. Precise elevation and coordinates
of the location under the cursor are available in the right part of the status bar. The legend
can be shown or hidden by selecting “Legend” in the “View” menu.
DEMs can be saved by clicking the contextual button in the tree.
5.2.6 Point Clouds
Point clouds are displayed under the “Point Cloud” section of the tree. The user can toggle
their visibility by using the checkbox. Points clouds can be also be added by selecting “Add
a point cloud to project”, by placing the cursor over the “DEM” item in the project tree. It
is also possible to directly delete a point cloud from the project by hovering over the
corresponding item and clicking on the “X”.

5.2.7 3D Models
3D models are displayed under the “3D Model” section of the tree. The user can toggle
their visibility by using the checkbox. 3D models can be also be added by selecting “Add a
3D model to project”, by placing the cursor over the “3D Model” item in the project tree.
It is also possible to directly delete a 3D model from the project by hovering over the
corresponding item and clicking on the “X”.
The user can also modify the way the 3D model is being displayed. If the model has a
texture, its visibility can be toggled. It can also be displayed as a wireframe. These options
can be accessed in the “3D Model Display Style” section of the “View” menu.
5.2.8 Orthos
With the project tree, the user can toggle the visibility of each generated ortho with its
checkbox. It is possible to directly delete all the orthos from the project by hovering over
the “Orthos” item on the project tree and clicking on the “X”. The user can also specify a
new or existing folder for orthos. In order to do this, place the mouse over the “Orthos”
and select the “+” icon to open or set a new ortho folder. Note that generating orthos
without specifying a new folder will delete the previously stored orthos from the folder.
The user has the option to select which orthos are to be processed in mosaic creation using
the contextual buttons. With this option, orthos can be disabled; they will not be
considered while generating a mosaic.
5.2.9 Mosaic
The user can toggle the visibility of the mosaic using the project tree by using the checkbox.
It is possible to directly delete the mosaic from the project by hovering over the “Mosaic”
on the project tree and clicking on the “X”. The user can also specify a new or existing
folder for the mosaic. In order to do this, place the mouse over the “Mosaic” item and select
the “+” icon to open or set a new mosaic folder. Note that generating the mosaic without
specifying a new folder will delete the previously stored mosaic from the folder.
5.2.10 Vectors
Vectors in a project are divided into two categories: Features, and References. Features
are associated to 3D constraints and user selections and can be edited in the DEM Editing,
Volume Calculation and Feature Extraction modules. The user can toggle the visibility of
the different vectors using the corresponding checkbox. References are used to guide the
user and serve to ease any aspect of the production, mostly editing. They can contain any
type of vector, including project division, waterbodies, etc.

5.2.11 Reference Orthos
Reference orthos are orthos that were created outside of a project and are used for quality
assessment. The AT module can also use the reference orthos along with a DEM as
horizontal and vertical reference data when GCPs are not available.
The functionality of the tree view with respect to the reference orthos is the same as the
one described in section 5.2.6. Note that disabled reference orthos will not be considered
while performing a bundle adjustment with reference imagery.
5.2.12 Contours
The user can toggle the visibility of the generated contour lines using the project tree and
its checkbox. Existing contours can be added by placing the cursor over the “Contour” item
in the project tree and selecting “Open contour”. Note that more than one contour can be
part of the project. It is also possible to remove a contour by clicking on the “X” when
placing your mouse over the respective contour item. Contours can be saved by clicking
the contextual button in the tree.
When a contour is selected using the tree, it is displayed as lines in the main window.
Hovering the cursor over a specific line will pop up the corresponding elevation. It is also
possible to click on a specific line to highlight it.
5.2.13 Online Data
If a projection was specified during the project creation, the user can toggle the visibility
of online data. These data are automatically downloaded from the internet and can help
assess the geolocation of the project. Online data can be either satellite imagery, maps,
terrain data, or hybrid. These options can be selected from the “Online Data Style Option”
in the “View” menu.
5.3 Properties
The project tree includes a window that displays the properties of a selected object (e.g.
GCP, camera, images, DEM, orthophotos). The tool is beneficial as it assists in the
assessment of metadata, such as the camera calibration, image position, GCP type, etc. The
information the property window displays is dependent on the type of object selected.
5.4 Zooming and panning

The main window of Correlator3D™ supports advanced zooming and panning capabilities.
Data caching and automatically triggered refreshes are performed to offer both rapid and
fluid response, while displaying the most accurate information available. Fast results are

obtained even for huge projects at low memory costs. When working with large files, hard
disk performance can limit the speed at which information can be displayed. In such
situations, Correlator3D™ will temporarily display a less accurate version of the data while
it is being loaded. The accurate version will be displayed shortly afterward, as soon as
available.
This allows the user to continue panning and zooming while data loading is performed and
ensures a fast response at any given moment.
Panning is performed by clicking and dragging with the left or middle (wheel) mouse
buttons.
Note that during DEM editing, only the middle (wheel) button is supported for panning.
Zooming can be performed in multiple ways. The simplest and most intuitive way is to use
the mouse wheel. If the mouse does not include a wheel, the options “Zoom->Zoom In”
and “Zoom->Zoom Out” in the “View” menu will provide the same results.
The “Zoom to region” feature lets the user specify a rectangular area. Zoom and pan
operations are then computed to show the selected area as large as possible and centered
onscreen. The “Zoom to region” feature is available by selecting “Zoom->Zoom to region”
At any time, the user can return to the default view by selecting “Zoom->Home” in the
“View” menu. The default view displays the entire project centered onscreen.
5.5 Image Band Setup

The Image Band Setup dialog allows the user to select which band of the image to use for
each color channel (red, green and blue). This allows to analyze multi-spectral imagery in
many ways. To view just a single band, the same band can be selected for all three channels.
Default black and white bands are also available to provide other combinations. Note that
the configuration can be set independently for each layer of the project (input imagery,
orthos, mosaic, etc.).
Figure 7: Image band setup window.

5.6 Picture-in-Picture 3D Viewing

Correlator3D™ features picture-in-picture (PiP) 3D viewing capability. The PiP brings a 3D
rendered view of projects through an inset window, visible at all times to streamline
processing and editing. The interactive viewing capability allows display of image locations
and associated viewing pyramids, as well as DSMs, DTMs, orthophotos and mosaics.
Ground control points, check points, and aerial triangulation results are displayed in 3D to
facilitate quality control. The PiP allows a 3D visual assessment of data and results
throughout the entire production workflow, including real-time modifications displayed
during DEM editing.
The PiP window always reflects in 3D what is currently displayed in 2D in the main view.
For example, zooming in the main window will result in cropping the data shown in the 3D
view. Therefore, users willing to inspect a feature in more detail should first zoom in the
main window, and then interact with the PiP window.
At any time, the 2D and 3D views can be swapped by pressing the corresponding button
in the toolbar.
The size of the PiP window can be set to small, medium, or large in the in “View->3D View”
menu. Users can also have the 3D view in its one window by selecting “3D View->Window”
Rotation in the PiP window is performed by clicking and dragging with the left mouse
button. Zooming can be performed by using the mouse wheel, while panning is done by
clicking and dragging with the middle (wheel) mouse button. The left button of the mouse
can be used to select an object like an image pyramid to see its footprint.
Users can always reset the view point to its original state using the corresponding button
in the toolbar.
Elevation values of objects being displayed in the 3D view can be exaggerated or

suppressed to facilitate data analysis. To achieve this, users can repeatedly press the
corresponding buttons in the toolbar.
5.7 Scale
The user can also zoom the current view to a specified scale. For example, to make
1 centimeter on the screen to be equivalent to 1 kilometer of ground truth, one would
enter a scale of 100,000 (this is the number of centimeters in a kilometer).
A scale indicator is located near the bottom of the legend. The indicator can be shown or
hidden by selecting “Scale” in the “View” menu. The scale is dynamically adjusted during a
zoom operation.

5.8 Status Bar

The status bar provides various types of information, including the sensor type, the DEM
horizontal resolution, the coordinates of the cursor and the elevation at the cursor’s
location. The sensor type is either “UAV/Medium/Large Format”, “Satellite”, “ADS” or
“VMA3” and indicates the current project type. The DEM resolution displays the resolution
of the DEM written in the file. Unlike the scale, it is invariant to zooming operations.
Coordinates are refreshed while the cursor is moved over the DEM and the elevation under
the cursor is displayed only when available. The status bar can be shown or hidden by
selecting “Status Bar” in the “View” menu.
5.9 Profile
The Profile tool generates a vertical profile of the DEM over a user-specified path. The
path consists of a series of connected linear segments. Once a path is specified,
Correlator3D™ opens a new window containing the vertical profile. The profile is
generated by sampling the DEM and contains two different representations
simultaneously.
The first representation is an area chart shown in light blue. It shows the real elevation
value for each sample based on a nearest neighbor approach for samples falling between
two values of the DEM. It provides a discrete representation that is well-suited for
visualizing surfaces that are highly non-continuous (e.g. urban areas).
The second representation is a smooth dark blue line super-imposed over the previous
representation. This time, a bilinear interpolation technique is used to obtain the sample
values. The representation is better suited for visualizing areas where the terrain is
continuous (e.g. rural areas).
In the profile windows, vertical blue lines show the delimitation of each linear segment of
the path.
When the mouse is moved over the profile, a blue circle located on the interpolated profile
follows the mouse motion. XY coordinates and elevation values associated with this blue
circle are automatically updated and shown in the lower right corner of the profile window.
A similar circle appears and moves synchronously in Correlator3D™’s main window over
the user-specified path to show the exact location corresponding to the mouse position in
the profile view.
A “stretch” button located in the status bar enables or disables the stretch mode. When
enabled (default), the profile is stretched to take the entire size of the window, but the
aspect ratio is not respected. When disabled, square aspect ratio is used, ensuring that the
same number of pixels in the x or y dimensions correspond to the same physical dimension
in DEM units.
By right-clicking on the profile, the user can save an image of the profile in a file or to the
clipboard.

Figure 8: Profile window.
5.10 Volume
The Volume tool measures a volume over a user-specified selection and a reference
elevation. The user begins by selecting the area in which the volume is to be calculated.
The functionality related to the selection is similar to the DEM Editing module. Please refer
to section 7.4 for details.
Once a path is specified, the user clicks on the corresponding button to open the Volume
window and selects a reference elevation for the volume to be computed. There are five
types of reference elevations:
- Interpolate DEM at polygon boundary: obtain volume by interpolating DEM values

present at the chosen boundary
- Interpolate polygon elevation: obtain volume by interpolating elevation values
defined in each polygon vertex
- Use fixed elevation: obtain volume from a specified height
- Remove constant offset (dig): obtain volume by removing an offset value
- Add constant offset (fill): obtain volume by adding an offset value
- Use reference DEM: obtain volume by selecting a reference DEM
These options give the user flexibility in the volume type they can calculate.
The following statistics appear in the properties windows on the left once the process is
complete:

- Volume (Cut): the volume measured between the reference elevation, and the
terrain above
- Volume (Fill): the volume measured between the reference elevation, and the
terrain below
- Terrain area: the area covered by the selection, considering the terrain morphology.
This can be visualized as the size of the sheet you would require to drape the terrain
- Orthogonal area: the area covered by the selection, not considering the terrain
elevation
- Volume coverage: percentage representing the coverage of the selection used to
measure the volume, with no consideration for the undefined elevation values. The
user should ensure that this value is always close to 100%. The value may indicate
that the reference DEM being used does not overlap completely with the current
DEM, or that the software was not able to completely interpolate at selected
boundary (because too many vertices were falling outside the model).
Figure 9: Volume window.

6. Project Creation
The simplistic design and interactive user interface is made possible through the creation,
and use, of projects. A project within Correlator3D™ consists of a file with information
about the location of all the input and output data: EO data, camera calibration, AT data,
projection, images, DEMs, GCP, orthos, mosaic and reference orthos. The purpose of the
project is to centralize information for ease of use. Part of this information must be
specified during project creation.
The project creation wizard is used to create a project in Correlator3D™. As the wizard
incorporates all supported sensor types, the user does not have to be concerned with the
specific sensor type for the full photogrammetry workflow. A project type must be simply
selected from one of five options, and proceed with the required steps.
Once a project has been created, it can be closed and accessed at a later time. A project
can be moved from one location to another, as long as the file hierarchy is preserved on
disk (e.g. one folder with the project file and all images, orthos).
To save the main file of a project, the user can select “Save Project As” within the File
menu. This will not change or move the other files, only the main one.
There are five types of projects differentiated by the sensor type and image size. These
include: UAV, Medium/Large Format, Satellite, ADS, and A3. The following sections
describe the steps to create a project for each of the listed project types. As the UAV and
Medium/Large Format project types only differ in image size, they are grouped together.
Figure 10: Project type.

6.1 UAV / Medium / Large Format

The UAV/medium/large format imagery refers to digital images and scanned films. The
difference with each is based on the image size. The UAV format is used for digital images
smaller than or equal to 40MP. The medium/large format is used for images greater than
40MP, and for scanned films.
UAV/medium/large sensor the user should ensure that the overlap between adjacent
images is appropriate to ensure optimal results. The endlap (i.e. the overlap between
successive images) should be at least 60% of the image size, but also should be as close as
possible to that value. In terms of the sidelap requirements (i.e. the overlap between images
from adjacent flight lines), it is recommended to use a value between 20% and 30%. Note
that for low-altitude flights, the endlap should be approximately 80% and the sidelap 50%
to minimize occlusion effects from one image to the other.
If images were collected with an endlap greater than 75%, half of the images can be disabled
(i.e. one out of every two images) from the image list. This will not only reduce the processing
time, but will also lead to more accurate results. Note that for image orthorectification, the
overlap of the resulting orthorectified images can be specified.
6.1.1 Step 1 - Image Selection
The user selects the images to be processed in the project. This is done by adding the
images and their corresponding EO data in the wizard. Images can be added individually,
or through a complete folder using the appropriate button. Once they are added, the
software automatically requests EO data (please refer to section 10.1 for further details).
If EXIF information is present, the software can use it to generate the EO data, though it is
not mandatory. The user has the option to import a text file, or a legacy SimActive EO file.
Projection information will be asked for after this dialog (please refer to section 4.6 for
further details).
More than one camera can be added to a project. If there is more than one camera, those
cameras can either belong to the same group (e.g. a multi-head camera systems), or to
different groups (e.g. an area was flown by two different cameras that are not attached
together). This can be performed through the drop-down menu under “Camera” and “<Add
a new camera>”. The camera name must be inputted and the group selected based on the
aforementioned description.
The user has the ability to remove all selected images using the “Clear” button, choosing a
new set of images instead. Specific images can be removed from the dataset using the “X”
button on the right of the list of selected images.
6.1.2 Step 2 – Camera Parameters
Camera parameters for all cameras of the project must be specified. The user can import a
legacy SimActive camera file. Alternatively, the software can detect the values

automatically, including the focal length and pixel size. “Digital – Polynomial” is the
suggested type for non-metric cameras similar to those normally used on UAVs. More
information on different camera calibration types can be found in section 13.1.1.1. If the
user is working with scanned films, the fiducial detection tool will be shown (please refer
to section 10.3 for further details).
Although parameters can vary from one camera to the other, only one calibration type is
supported for a given project. For scanned films, if the dimensions of images in pixels
vary, then the smallest width and height values observed should be specified. Note that
these minimum width and height values may come from two different images.
6.1.3 Step 3 – Project Parameters
In this step, the user defines the project location by clicking on the “…” button. There is the
option to specify a projection, however, it cannot be different than the one selected for
the images in Step 1.
The other project parameters are optional. The “DEM” option allows the user to specify
(e.g. SRTM) a reference DEM.
With the “GCP File” option, the user can specify a file containing GCP and check point
information. The file can be an existing one or it can be created by importing data from a
text file (please refer to section 10.4 for further details).
The “Reference Orthos” option lets the user browse and select a folder containing the
reference orthos.
Both GCPs and reference orthos can be specified during project creation. However, when
performing bundle adjustment, the user will need to select a single ground reference.
6.2 Satellite
The metadata of a satellite image is stored in its corresponding RPC file. This file comes in
several formats, all of which are supported by Corrleator3D™. When creating a project, the
user must specify the input RPC files of the images to be processed. Correlator3D™
automatically parses these files to recover the satellite parameters. For large satellite
images, the images may be tiled and referenced by a TIL file. This TIL file will be required
as an input. Only one (1) RPC file is needed per TIL file. The project creation wizard will
automatically suggest an image file when an RPC file is selected first.
The user has the ability to remove all selected images using the “Clear” button, choosing a
new set of images instead. Specific images can be removed from the dataset using the “X”
button on the right of the list of selected images.

6.2.2 Step 2 - Image Pairing
In order to generate a DSM using overlapping satellite images, the user must specify image
pairs. Each pair will be used to generate a corresponding DSM, with the results merged at
the end of the process.
To construct a pair, select each of the images to be paired using the drop-down menu, and
click on “Add”. A star (*) before the name of the image in the drop-down menu indicates
that the image has not been used in any pair thus far.
This step can be skipped if no DSM is required.
6.2.3 Step 3 – Project Parameters
In this step, the user defines the project location by clicking on the “…” button. It is
mandatory to specify a projection. The projection can be specified by selecting an existing
DEM with a valid projection in it.
The other project parameters are optional. The “DEM” option allows the user to specify
(e.g. SRTM) a reference DEM.
With the “GCP File” option, the user can specify a file containing GCP and check point
information. The file can be an existing one or it can be created by importing data from a
text file (please refer to section 10.4 for further details).
The “Reference Orthos” option lets the user browse and select a folder containing the
reference orthos.
Both GCPs and reference orthos can be specified during project creation. When performing
bundle adjustment, the user will need to choose which ground reference to use as they cannot
be used simultaneously.
6.3 ADS
The metadata of an ADS image is stored in a standard SOCET SET support file. This support
file is an ASCII file that uses keywords. When creating a project, the user must specify the
support files of the images to be processed. Correlator3D™ automatically parses these files
to recover the names of the images, the camera calibration files, and the exterior
orientation data. Therefore, the user must make sure that all file names and paths specified
in the support file are valid.
In the ADS support files, paths containing spaces are not valid.
Correlator3D™ only processes level 1 (L1) imagery for ADS sensors.

6.3.2 Image Pairing
In order to generate a DSM using ADS imagery, the user must specify image pairs. Each
pair will be used to generate a corresponding DSM, with the results merged at the end of
the process.
To construct a pair, select each of the images to be paired using the drop-down menu, and
click on “Add”. A star (*) before the name of the image in the drop-down menu indicates
that the image has not been used in any pair thus far. This step can be skipped if no DSM
is required.
Better results in ADS mode will be achieved if the angle between the sensors is small (i.e.
forward-nadir or nadir-backward). Using the pair forward-backward is likely to generate noisy
or erroneous results due to occlusions (objects visible in one image might not be visible in the
other).
6.3.3 Project Parameters
In this step, the user defines the project location by clicking on the “...” button. It is
mandatory to specify a projection for an ADS project. The projection can be specified by
selecting a DEM with a valid projection; it must correspond to the one specified in the
support files.
6.4 A3
Each VMA3 frame is stored in its own folder along with files containing metadata. When
creating a project, the user must specify a VMA3 frame folder containing all frames to be
processed.
Correlator3D™ only processes Super Large Frame (SLF) imagery for VMA3 sensors.
6.4.2 Project Parameters
In this step, the user defines the project location by clicking on the “...” button. It is
mandatory to specify a projection for an ADS project. The projection can be specified by
selecting a DEM with a valid projection in it.

7. Data Generation
7.1 Aerial Triangulation

Correlator3D™ includes an aerial triangulation (AT) module for UAV/medium/large format
sensors and satellite sensors, which automatically improves exterior orientation values and
camera calibration parameters from the image data. It is therefore recommended to
perform this process on every project to ensure the quality of the geospatial products
generated from the imagery. The software requires as a minimum overlapping aerial
images, coarse GPS measurements, and basic camera parameters.
It is worthwhile to spend time on the AT step, as it will help save significant processing
time on subsequent steps.
To enter this mode, the user must select “Aerial Triangulation” in the “Module” menu. Aerial
triangulation should be performed in the following order:
1. Tie Point Extraction

2. GCP Creation (if required)
3. Tie Point Editing (if required)
4. Bundle Adjustment
5. Quality Assessment
6. Image List Filtering (if required)
7.1.1 Tie Point Extraction
The first step is to extract tie points, which is performed automatically. A tie point is an
image feature that is visible in two or more images.
Correlator3D offers two types of tie point extraction: standard and exhaustive. The
standard extraction type is suitable for most projects, while the exhaustive type is better
fitted for challenging datasets with suboptimal inputs.
Exhaustive tie point extraction is usually slower than standard tie point extraction. It is
suggested to use the exhaustive type only when the standard one does not generate optimal
results (e.g. several images that are not connected to others).
If reference orthos are present, the tie point extraction process will also extract points in
those images, even if their ground sampling distance is different. Note that these
observations will be used during bundle adjustment only if the Ground Reference option is
set to “Reference Orthos”.
Compared to bundle adjustment, the tie point extraction process is the most time
consuming. If the user needs to perform numerous bundle adjustments on the same
project, then tie points do not need to be recalculated each time.

For best performance, the user should ensure to have both a GPU and CPU OpenCL
enabled devices. Most graphics card drivers include OpenCL support. However, special
drivers may need to be installed to enable OpenCL on the CPU. The CPU manufacturer
can be consulted for more details.
7.1.2 GCP Creation
Next, the user must decide whether to use ground control points (GCP). Although not
required for relative accuracy, ground control points are necessary for absolute accuracy.
The quality of the ground control points will have a strong impact on the results. Ground
control points have two benefits. First, they ensure that there is no global spatial offset (i.e.
xyz translation) in the refined values. Second, they provide another means to validate the
residual error of the refined values.
The user can also create check points. These are special GCPs that will not be considered
during bundle adjustment, but for which statistics and reports will be generated. Check
points are used to assess the quality of the bundle adjustment using an independent data
set.
The GCP Editing tool helps the user locate GCPs in the images by selecting the best image
candidates, and estimating where the GCPs should fall in those images. Correlator3D™
relies on the exterior orientation parameters and camera calibration data located in the IEO
file. Upon entering the GCP Editing tool, the software will verify the average residual error
and suggest the user to perform an estimation of IEO parameters if this error is too large.
This will add an “Estimate” AT step to the project tree. When GCP creation is completed
and bundle adjustment is performed, the “Estimate” step will be automatically removed.
Therefore, the “Estimate” step has no impact on the bundle adjustment process. The GCP
Editing tool is described in Section 10.4.
7.1.3 Tie Point Editing
After the initial automatic tie point extraction step has been performed, the software will
display links created between images. Hovering the mouse over any link displays the
number of tie points present between those images.
The user can right-click on any link to visually inspect or edit the tie points between two
images, ensuring features match in both images. They can also right-click on the image
position to display the image itself, with all its feature points displayed as an overlay.
Under some circumstances, it may not be possible to automatically extract tie points
between some images. This is more likely to occur when the overlap is too low or the
quality of images is not sufficient. The user can then manually add tie points to the project.
A link can be added between two images by clicking on the corresponding button, by
clicking and holding the left button on the first image and by dragging the link to the second
image. The tie point editing dialog will appear with the two corresponding images. The user

can zoom in using the mouse wheel, and pan by clicking and dragging with the center
button (wheel).
Figure 11: Tie points between two images.
At this point, the objective is to add as many tie points as possible. To do so, the user must
right-click on one feature in one image, and drag it to the corresponding feature in the
second image. The software will then automatically try to refine the match, and will display
the Tie Point Editing window. In the window, the user can either validate or refine the
position of the feature in each image.
A tie point is more beneficial when the feature is identified in multiple images. Due to this,
user should always consider using such features, distinguished as colored stars in the
images. If the user clicks on one of these stars, then drags to the corresponding feature in
the other image, the software will display the Tie Point Editing window. Multiple images
containing the feature will be displayed. The new observation for the tie point will be
identified with its name in blue.
As the user enters more tie points, the software will adjust the relative orientation of the
two images to facilitate the task. Also, when moving the cursor in one of the two images,
the software will provide a predictor (shown as a multicolored circle) in the other image to
give a hint of where the feature should approximately appear. This helps facilitate the
process.
At all times, the user can select a tie point to edit by either right-clicking on it, or using the
corresponding button.
The user can also select a tie point to delete it using the corresponding button.

There is no specific rule on how many tie points should be added between two images,
although 5 or 6 is a good approximate.
Note also that the goal is not necessarily insuring that there is a link between every
corresponding image. For example, adding a link between a pair of images can tie two flight
lines together, but it is not necessary to link each adjacent images in the flight lines. Instead,
the links can be added on image pairs that are well dispersed (e.g. one image pair in the
beginning, middle, and end of the flight lines). This can save the user time, while ensuring
that the overall project is connected.
7.1.4 Bundle Adjustment
Bundle adjustment derives adjusted exterior orientation and camera data. Correlator3D™
can be used to perform a boresight calibration or a full AT adjustment. In the boresight
calibration process, the software computes the error component that is common to all
exterior orientation estimates. This corresponds to the relative position/orientation of the
camera to the GPS/IMU unit. If a full AT adjustment is selected, then the software still
automatically performs a boresight calibration beforehand but also computes a unique
correction for each individual image.
The software can also be used to perform a camera calibration in parallel to the EO
adjustment. The software adjusts camera parameters based on the camera type used.
These values may include focal length, principal point, LUT values, as well as radial and
decentering distortion parameters. Various camera models are supported and described in
Appendix A.
When available, the user can use either ground control points or reference orthos to ensure
that there is no global spatial offset (i.e. XYZ translation) in the refined values.
When using ground control points, Correlator3D™ takes into account user error when
identifying the GCP location in the images and also accounts for possible error in the XYZ
coordinates. In the bundle adjustment settings, the user may specify uncertainty values
for the GCP. The "Pixel Radius" is the presumed error when defining the GCP locations in
the image using the GCP Editing tool. For example, a value of 0.5 means that the user
predicts having introduced an error of half a pixel when defining the GCP location in the
image.
The bundle adjustment settings let the user specify a general uncertainty for the "XY" and
"Z" components of the GCPs. For example, an uncertainty value of 1m for the "XY" field
will allow a 1m fluctuation to the XY values of the GCPs, whereas a value of 0m implies the
user has absolute confidence in the GCPs. This will force the software to comply with the
GCPs exactly when performing an EO adjustment. This setting only applies to GCPs that
do not have per-point uncertainty values.

Per-point uncertainty values are specified using the GCP Editing tool. Different XYZ
uncertainties can be specified for each GCP. When available, these values override the
aforementioned general uncertainty values.
When GCPs are used, final residual analysis and statistics will also be computed for the
check points.
If reference orthos are used, tie points detected in these images will provide horizontal (XY)
anchoring to the project. If a DEM is currently loaded in the user interface, it will provide
vertical (Z) anchoring. While optional, the use of a reference DEM is recommended.
Bundle adjustment can be run in two (2) modes. These are called "Constrained" and
"Unconstrained" and apply to the camera calibration and EO adjustment settings. The EO
adjustment setting also allows for a boresight adjustment only or for direct georeferencing.
The unconstrained mode performs all necessary parameter allocations for the user. It is
designed to provide the fastest AT solution. The constrained mode requires specification
of the bundle block settings, providing more control over the project.
In the constrained mode, the user can lock one or more parameters for bundle adjustment.
These locked parameters can come either from the EO values and/or the camera values.
For example, if the GPS and camera data are known to be accurate, the user may choose
to lock them and only process the IMU portion for bundle adjustment. This would force
the software to perform bundle adjustment only on the IMU portion of the EO values.
In the constrained mode, the user is also required to define a search space for the easting
(x), northing (y), elevation (z), omega, phi and kappa values. The search space defines the
assumed uncertainty regarding the accuracy of the input data. For example, specifying an
uncertainty of 5m for easting and an uncertainty of 2° for omega, restricts the software to
adjust the easting and omega parameters from their current positions as defined by the
input to a maximum of 5m and 2° off respectively, in the output.
Note, in a full AT adjustment, the uncertainty value bounds only the full AT portion of the
adjustment, meaning only the unique correction for each image will be bound by the
uncertainty value. The boresight portion of the adjustment, which is automatically
performed during any full AT adjustment, will not be bound at all. For example, if a 5m
value in X was chosen as the uncertainty for a full AT adjustment, then as a first step, a
common boresight correction of any appropriate magnitude (as determined by the
software) will be applied to all images followed by a unique correction to each individual
image of up to 5m in X. The user may change this behavior by unchecking the "Perform
Boresight Calibration" option. With this option unchecked, the software will skip the
boresight calibration process, making sure the uncertainty range is respected during the
adjustment.
The "Direct Georeferencing" option should only be used when the GPS data are known to
be accurate (e.g. RTK navigation systems). In that case, the software will only adjust the
orientation parameters.

The search space must be appropriately large to compensate for the EO errors. Large
average residuals may imply that the search space was too small and that results are not
optimal.
In the case of satellite sensors, the AT process will calculate corrections on RPC values to
improve the accuracy of the georeference. The software requires overlapping satellite
images and RPC data. GCPs may also be used to ensure absolute accuracy. The EO
refinement process will automatically adjust the RPC values and output the RMS pixel
error.
Finally, an option to remove images having a tie point residual error higher than the
specified threshold is provided. This option is equivalent to the Image Filtering functionality
described in section 7.1.6 except that it is performed through the bundle adjustment
process. Selecting “Automatic” will make the software to calculate the average residual
error for all images, and use a threshold corresponding to five times this value.
7.1.5 Quality Assessment
Once the bundle adjustment has been performed, the user is presented with the overall tie
point residual error expressed as colored circles centered on each image. There are four
different colors:
- Green: when the residual error is less than 0.75 pixel

- Yellow: when the residual error is between 0.75 and 1.5 pixel
- Red: when the residual error is larger than 1.5 pixel
- Blue: when it was not possible to correct the position/orientation of the image (e.g.
no tie point)
Note that it is normal for the residual errors to be red in the initial step. It simply means
that bundle adjustment is necessary to bring the values down to the green range. Tie points,
ground control points and check points, when available, are also displayed using the same
color coding. By hovering the mouse over a point or an image, the user can see the
associated residual error in the right hand side of the status bar.
At any time during AT, the user can switch between the color-coded circles and the kappa
arrows by selecting the corresponding option in the “View” menu. The “IEO Display Style”
option applies to images while the “GCP Display Style” option applies to GCPs. For GCPs,
a third display style is also available: Spatial Residuals.
The spatial residual is the difference between the specified coordinates of the GCP and the
theoretical coordinates that are triangulated using the current IEO data. When the “Spatial
Residuals” display style is selected, each GCP and check point is represented as a set of
two arrows. The red arrow shows the vertical residual error. A red arrow pointing up
indicates that the specified GCP altitude is higher than what is triangulated from the IEO
data. The blue arrow shows the planar residual error. A blue arrow pointing north-east
indicates that the specified GCP coordinates are north-east of what is triangulated from
the IEO data.

In the software interface, an AT List is displayed on the left hand side detailing the AT steps
that have been performed. As more bundle adjustments are performed, the list grows
displaying all iterations. This allows flexibility to back track in case of an undesirable
iteration. Specifically, the last bundle adjustment operation may be undone or the user may
choose to come back to results of a specific iteration.
Figure 12: Bundle adjustment results.
To provide the user with feedback, detailed aerial triangulation reports are generated
outlining the statistics on all operations performed. The user has access to a general
summary, or more detailed reports, such as boresight corrections, image and camera
calibration adjustments and tie point, image and GCP residuals. At the end of any AT
process, the software automatically displays a PDF report summarizing the statistics as well
as the adjustment performed. For convenience, a link to the folder containing more detailed
AT reports is also provided at the bottom left corner.
The PDF report can also be accessed from the project tree using the contextual button for
a specific bundle adjustment step.
7.1.6 Image Filtering
Before continuing with the rest of the photogrammetric workflow, the user should ensure
that incorrect images are removed from the project. This will avoid having noise or other
artifacts in their results. Therefore, the option to perform “Image Filtering” on the right-
hand side is available. This tool creates a new AT step in which images having a tie point
residual error higher than the specified threshold will be removed. The user can see the
images being filtered in real-time as the threshold value is modified. Note that this
operation can also be done automatically through bundle adjustment using the
corresponding setting.

7.1.7 Coarse DEM Extraction
Once all iterations have been performed and the user is satisfied with the results, the
updated project can be saved. At this point, the user can create a coarse DEM with the
corresponding button. This coarse DEM can be useful if the user wants to assess the quality
of the AT. For example, images could be orthorectified using this DEM to make sure that
their positions are correct.
7.2 DSM Generation

One of Correlator3D™’s main capabilities is the generation of digital surface models (DSM)
from imagery. The software requires overlapping aerial images and GPS/IMU data. The
user must specify the desired output horizontal resolution, or simply request the software
to calculate an optimal value given the input dataset. This feature takes into account the
project parameters (i.e. image resolution, altitude, focal length, etc.) as well as the required
processing time and then determines the best compromise. The user can also request the
software to calculate the maximal value given the input dataset, or a coarse value for better
processing times.
Once a horizontal resolution is specified, the user must select a vertical accuracy from the
corresponding slider bar. Lower accuracies will lead to better processing times while higher
ones will give the best results, but with longer processing time.
The user can generate a point cloud by selecting the corresponding option. It will be
generated in parallel with the DSM. Once the point cloud is created, it can be colorized
using the Point Cloud Colorizing tool. Please refer to section 10.11 for details.
Images are processed pair-by-pair in the order that they appear in each flight line. The
required time is thus directly proportional to the number of images. Processing can be
performed on multiple machines by taking the exact same project and disabling images
through flight lines based on the respective divisions. DSM results can then be merged
using the DEM Merging Tool (see Section 10.5).
An optional 3D constraint file can be provided to Correlator3D™. This file can contain
points, lines or polygons with elevation (3D) data. Using data values located in the
constraint file may result in two benefits. First, it may help the DSM generation process to
resolve potential ambiguities, such as occluded regions. Second, it will force the output
DSM to have the provided elevation values at the location of the constraining elements.
If a 3D constraint is provided, Correlator3D™ will force the output values in the DSM to
be consistent with the provided constraint. However, if the constraint is not consistent
with the rest of the dataset, it may lead to visual artifacts in the DSM. If such a situation
appears, the user should make sure that the 3D constraint is in the same coordinate
system (e.g. same datum) and that all input data are valid (e.g. camera calibration or
exterior orientation parameters).
The software generates a DSM representing all objects in the scene (terrain, vegetation
and man-made structures). No manual post-processing is necessary to remove pits or

spikes: once generated, the DSM is ready to use. Correlator3D™ will generate high quality
DSMs across most types of surfaces. However, because of their inconsistent appearance
from one image to another, water bodies will generally lead to noisy DSM regions.
The following steps must be followed to perform DSM generation:
1. Select “DSM generation…” in the “Processes” menu.

2. If a 3D constraint is used:
a. Check “3D Constraint”.
b. Specify the file containing the 3D constraint using the corresponding drop-
down menu. The constraint will be displayed in the main window overlaid
on the DEM.
3. Specify the desired resolution. Press the “Optimal”, “Maximal” of the “Coarse”
button to let the program calculate an optimal, maximal or coarse value. Once a
resolution is specified, select a vertical accuracy from the corresponding slider bar.
4. If a point cloud is required, check “Generate Point Cloud”.
5. Click on the “Process” button. Real-time feedback will be provided on the overall
progress of processing.
Save the generated DSM by selecting “Save DEM As…” using the contextual
button in the project tree. If a point cloud was generated, save it by selecting “Save
Point Cloud as…” using the contextual button in the project tree.
Figure 13: DSM generation dialog.
7.3 DTM Extraction

Correlator3D™ features a module to extract a digital terrain model (DTM) from a digital
surface model (DSM). Such models are particularly useful for generating orthophotos, as
they avoid visual artifacts (e.g. along building edges) caused by using DSMs. The software
analyses the DSM and automatically removes structures lying on the ground.
In addition to the input DSM, an optional constraint file can be provided to Correlator3D™.
This file can contain points, lines or polygons with elevation (3D) data. It will force the

output DTM to have the provided elevation values at the location of the constraining
elements. Alternatively, the user has the option to provide an exclusion zone. In such a
case, the software will extract the DTM for all areas minus that of the exclusion zone.
If a constraint is provided, Correlator3D™ will force the output values in the DTM to be
consistent with the provided constraint. However, if the constraint is not consistent with
the rest of the dataset, it may lead to visual artifacts in the DTM. If such a situation
appears, the user should make sure that the input DEM and 3D constraint are in the same
coordinate system (e.g. same datum) and that all input data are valid (e.g. camera
calibration or exterior orientation parameters).
The user can also output into a DEM the features that were removed from the DSM. Note
that only the DSM features will appear in that DEM. All other elevation values will be set
as undefined.
To create a DTM, follow these steps:
1. Select the DSM you wish to use for DTM extraction from the project tree.
2. Select “DTM extraction…” in the “Processes” menu.
3. If a constraint is used:
a. Check “Constraint”.
b. Specify the type. Either "Elevation" or "Exclusion Zone".
c. Choose the file containing the constraint using the corresponding drop-
down menu. The constraint will be displayed in the main window overlaid
on the DEM.
4. If the DSM features are required:
a. Check “Output DSM Features”.
b. Specify the output DEM file using the corresponding drop-down menu.
progress of processing. Once the process is done, the DTM will appear in the main
window.
6. Save the generated DTM by selecting “Save DEM As…” using the contextual
button in the project tree.
Figure 14: DTM extraction dialog.

7.4 DEM Editing

While the DSM/DTM generation modules produce highly accurate results, the user may
still desire to alter results according to their specific requirements. The Correlator3D™
DEM Editing module features nine (10) easy to use, yet powerful editing tools:
– Crop an area
– Delete an area
– Force the elevation to a user-specified value
– Offset the elevation by a user-specified value
– Apply a median or mean filter of user-specified size
– “Delete and fill” i.e. delete an area and analyze neighboring areas to fill the
deleted region.
– “Topological Fill” i.e. delete and fill an area while preserving the topology of
the neighboring terrain.
– DTM Extraction for a user specified area
– Touch Up small areas
– Paste an external DEM onto the existing DEM
The aforementioned tools process a user-specified polygonal selection. The time required
to perform the operations depends on the operation type and the size / complexity of the
selection. The user can also display orthorectified images, a mosaic or any other type of
reference data as an overlay to the DEM using the project tree. This can facilitate the
selection of features like water bodies.
To edit a DEM with a user-defined selection the following steps must be performed:
1. Enter the DEM editing module, by selecting “DEM Editing…” in the “Edit”
menu. The editing toolbar will appear at the right of the window.
2. Create a selection.
3. Select an operation to perform in the Edit menu or the DEM Editing toolbar.
4. Repeat step 2 and 3, as required
5. Save the edited DEM by selecting “Save DEM As…” using the contextual
There are three modes of selection that are mutually exclusive:
– Standard selection
– Append a selection
– Touch up
7.4.1 Standard Selection
The standard selection is the most common way of selecting an area for editing. It is used
when several operations need to be performed sequentially on several locations. A new
selection can be created at any time by clicking on the left mouse button. Doing so
automatically deselects the previous selection and starts a new one. Additional left-clicks

are necessary to make the selection. To close the selection, the user must click in the
vicinity of the initial point or simply right-click.
During a selection, pan operations are supported using the middle (wheel) mouse button.
Placing the cursor at the edge of the display area automatically pans the DEM.
Note that zoom operations are not permitted during a selection.
Correlator3D™ supports single non-self-intersecting polygonal selections. As such,

Correlator3D™ will refuse to add a point to or close the selection if it makes the polygon to self-
intersect.
When the selection is finalized, a contextual toolbar is displayed on the top left corner of
the main view. With this toolbar, the user has the option to edit the selection with different
tools before doing an operation on the region.
The first one, the hand tool, is used to move, add, or delete selection points. A point can
be moved by clicking on its vicinity with the left button of the mouse and dragging it at a
new location. Conversely, it can be removed by clicking on the right button. New points
can be added by clicking on the right button of the mouse anywhere on the selection.
Depending on the scale factor, Correlator3D™ will adjust automatically the level of details of
the selection. Therefore, if the user zooms in the selection, additional selection points will be
shown.
The second tool, the scissors tool, is used to remove a series of selection points and
replacing them with new ones. A first point must be selected with clicking on the left button
of the mouse, and then a second one to determine the series of points to be removed. New
points can then be added to follow a new path, starting from the last point to the first one.
Note that elevation values for each vertex of the polygon can also be manually set. Use
this function to select any vertex and change the elevation value.
The last operation on a selection can always be undone by pressing the corresponding
button in the mosaic editing toolbar or using the keyboard shortcut “Ctrl-Z.”
Correlator3D™ also offers automatic tools to modify a selection. If instead of a polygonal

selection the user wishes to select a feature, the Feature Selection option should be used.
This option will snap automatically to the boundary of the feature contained within the
original selection. The user has then the option to edit the selection by moving existing
points or creating new ones.
The user can also snap the selection based on elevation values. This is particularly useful
to select undesired spikes or other features to be removed quickly. When clicking on this
button, a dialog will be displayed to select the criteria to be used to snap the selection. For
example, the user could choose to select everything above a certain elevation.

Once the selection is completed, the user can either select the DEM operation to perform,
or save the selection before by clicking on the “Save Selection” button from the contextual
toolbar, or by using the “Enter” keyboard shortcut.
A selection can also be cleared by clicking on the “Delete Selection” button from the
contextual toolbar or by using the “Escape” keyboard shortcut.
Finally, if the user wants to edit the selection after saving it, clicking on the selection will
bring back the contextual toolbar, and editing will be possible. Clicking the “Discard
Changes” button will bring back the original selection.
7.4.2 Append a Selection

Once a first selection has been created, the user can toggle the “Append a Selection” mode.
This mode will automatically add a new feature file in the project tree, allowing the user to
save the selection to disk (by using the corresponding contextual button next to it). It then
becomes possible to add other selections to this file. This can be particularly useful if the
same operation has to be applied simultaneously on several locations. It can also be useful
to define exclusion zones or other types of constraint that can be used during other
processes like the DTM Extraction.
This mode provides the exact same functionality than the Standard Selection, meaning that
a contextual toolbar will be displayed for each new selection that it being appended. The
user can click on the “Save Selection” button from the contextual toolbar to add this new
selection to the file.
Note that, just like in the “Standard Selection” mode, it is always possible to edit one of the
selections by clicking on it.
7.4.3 Touch Up
The Touch Up tool works in a similar manner than the Delete and Fill tool. Instead of
requiring a manual polygonal selection, a small fixed-size square selection is automatically
performed around the location clicked by the user. The tool requires less user interaction
and the internal processing is optimized for higher performances. Thus, it is a very fast tool
to correct small areas.
7.4.4 Crop
Correlator3D™ features a polygonal crop tool. Once a polygonal selection is made, the
Crop tool will discard any elevation values outside the selection. The DEM will also be
resized to fit the size of the bounding box of the selection.
To use the Crop tool, select “Crop” from the “Edit” menu.

7.4.5 Delete
The Delete tool discards all elevation values inside the selection, leaving any other value
intact.
To use the Delete tool, select “Delete” from the “Edit” menu.
7.4.6 Set Elevation

The Set and Offset Elevation tools are accessible from the same dialog.
Figure 15: Set and offset elevation dialog.
The following is a description of the required settings:
Operation: Decides between a Set Elevation or Offset Elevation function.
Value: New elevation value for the Set Elevation tool or offset value for the Offset
Elevation tool.
The Set Elevation operation sets the selected DEM cells to the specified elevation, while
the Offset Elevation operation offsets the selected DEM cells by the specified elevation.
The elevation can be a positive or negative fractional value and is specified in the DEM
units (meters or feet).
To use the Set Elevation or the Offset Elevation tool, select “Set Elevation…” in the “Edit”
menu.
7.4.7 Filter
The filter tool applies a mean or a median filter to the selected DEM cells. Cells outside the
selection and cells located on the border will be used to compute filtered elevation values.
However, these cells will remain unaffected by the process. Settings can be adjusted using
a dialog.

Figure 16: Filter dialog.
The following is a description of the required settings:
Filter Type: Decides between mean and median filters.
Filter Size: Diameter of the filter, specified in DEM units (meters or feet). The specified
value must be an odd multiple of the DEM resolution.
To use the Filter tool, select “Filter…” in the “Edit” menu.
7.4.8 Delete and Fill

Delete and Fill is probably the most useful and powerful editing tool provided by
Correlator3D™. With a single click, all the selected DEM cells are deleted and replaced by
new elevation values obtained from an analysis and interpolation of the area surrounding
the selection. It requires no parameters to be specified and provides excellent results in
most situations.
To use the Delete and Fill tool, select “Delete and Fill” from the “Edit” menu.
7.4.9 Topological Fill

The Topological Fill tool provides similar functionality to the Delete and Fill tool, except
that it analyzes the topology of the neighboring terrain to better interpolate the area. For
example, it can be used to remove a feature located on a slope.
To use the Topological Fill tool, select “Topological Fill” from the “Edit” menu.
7.4.10 DTM Extraction

The DTM Extraction tool extracts the DTM for the selected DEM cells. It uses the same
algorithms than the DTM extraction process and settings can be adjusted using a similar
dialog. Please refer to Section 7.3 for further details.

Figure 17: DTM extraction dialog.
To use the DTM Extraction tool, select “DTM Extraction” from the “Edit” menu.
7.4.11 Paste DEM

The Paste DEM tool replaces regions of the current DEM with a provided DEM. The user
can adjust the height tolerance parameter that defines the difference in height between
the two DEMs for which a smooth transition will be calculated. Hence, elevation values
showing a difference smaller than this value will be averaged in the final DEM.
Figure 18: Paste DEM dialog.
The Paste DEM tool will not consider any current selection as the input DEM will be pasted
entirely.
To use the Paste DEM tool, select “Paste DEM” from the “Edit” menu.
7.4.12 DEM History (Undo/Redo)

The history of each modification made to the DEM is automatically saved. This history is
displayed below the project tree as a list of all operations that were performed to the DEM.
Clicking on one of the operations will undo all subsequent operations. To better guide the
user on what operations should be undone, Correlator3D™ provides the following visual
feedback: hovering the cursor above any item of this list will display all the modified regions
so far as an overlay to the DEM. Regions shaded in blue with a number in it correspond to
the regions that will not be affected by the “Undo” operation if the user clicks on this item.
Regions shaded in red correspond to regions for which the modifications will be canceled.
Note that it is always possible to redo the operations if no operation is made after the
undo.

7.5 Orthorectification
Correlator3D™ can generate orthorectified images. The user can specify the desired
output resolution, or simply request the software to calculate the optimal or coarse value
for better processing times. This feature automatically determines the ground sampling
distance (GSD) of the input imagery.
Correlator3D™ can generate true orthos if a DSM and/or a constraint file is used. In such
a case, the orthorectification process will correct for the distortion of buildings and other
3D structures, resulting in a true ortho. Occluded areas which have no image content, e.g.
building shadows, will be set to black in the final orthorectified image. Note that if a
constraint file is inputted, it must contain polygons with elevation (3D) data defining the
structures.
For frame-based sensors, the user can also specify the overlap of the resulting
orthophotos. By default, the software will suggest the maximal overlap, i.e. the actual
overlap between images. The user can also select the optimal overlap, which corresponds
to 30%.
Choosing a smaller overlap will result in a faster orthorectification, smaller orthophotos on disk
and ultimately faster mosaic creation. Note however that reducing the overlap between
orthophotos will also reduce the area where a mosaic seamline can go through and may
therefore impact results.
For UAV/medium/large format sensors, the user can crop the image width and height to
any user-defined value. This allows for forward and sidelap control, which may be desirable
as a strategy to extract the most nadir image portion from the orthophoto.
The better the quality of the inputs, the higher the quality of the resulting final orthophotos.
Therefore, it is recommended first to perform an aerial triangulation1, a DSM generation and a
DTM extraction and then use the DTM for producing orthorectified images.
The output format of the orthorectified images is by default tiled with overviews. This
should not be changed if the user prefers using the Mosaic Editing module afterward to
obtain the highest performance.
A real-time overview of the orthophotos will be displayed during processing. Each time an
orthorectified image is generated, it will appear as an overlay to the currently loaded DEM.
To create orthorectified images, the following operations should be performed:
1. Select the DEM you wish to use for orthorectification from the project tree.
2. Select “Orthorectification…” in the “Processes” menu.
3. If true orthorectification is required:
a. Check “DSM-based” if a DSM is being used for orthorectification.
1 Not supported for ADS and VMA3 sensors.

b. If a 3D constraint file is used, check "Constraint File" and specify the file
using the corresponding drop-down menu. The constraint will be displayed
in the main window overlaid on the DEM.
4. Specify the desired output horizontal resolution or press the “Optimal” or “Coarse”
button to let the program calculate an optimal or coarse value.
5. In the "Settings", select "Overlap" to specify the desired output overlap
percentage. Press the “Optimal” or the “Maximal” button to let the program
calculate the optimal or maximal value. To crop the image width and height, select
"Crop" and specify the desired percentages.
6. Specify the format for the orthorectified images.
Figure 19: Orthorectification dialog.
7.6 Mosaic Creation

The mosaic creation module allows for the merging of individual orthorectified images to
obtain a unique image covering the entire project area. The process is entirely automated.
Correlator3D™ will select which portion of which image must be included in the final
mosaic. Local intensity adjustments and seamline generation are performed to provide
smooth and seamless transitions between adjacent images composing the mosaic.
Before starting the mosaic operation, the user can specify how many blocks to divide the
mosaic into at minimum. A block is a smaller section or part of the overall mosaic. The
purpose of this is to support the generation of large mosaics but also to allow multiple users
to edit different blocks simultaneously after the mosaic has completed. The number of
blocks is therefore dependent on the user’s choice and project size. The software will
create the blocks according to the user's selection, but may need to perform further block
divisions as necessary.
Selecting the Bound Mosaic Area option allows to generate a mosaic for only a portion of
the project, as defined by a user created polygons stored in a tile definition file or a
Shapefile. The user can also specify the size of the feathering zone depending on project

type. The feathering zone is the region along a seamline where blending will be applied to
ensure a smooth transition from one image to the other. For urban areas, a smaller value
is recommended in order to avoid ghosting effects along smaller objects like cars. In other
areas like open fields, the value should be larger since the content is very similar from one
image to the other. The size of the feathering zone is specified in pixels. Note that when
editing a mosaic, a different value can be specified for each seamline.
Fully automated color balancing can be applied during the mosaic creation process. This
balancing includes:
Brightness and contrast adjustments of each image

Inter-image equalization
Independent channel analysis for hue equalization
Photometric constraints to maintain global information of brightness, contrast
and hue
Some projects may include regions where colors are very different from one image to the
other (e.g. water bodies). In order to avoid these regions to have too much effect on the
color balancing results, the user can specify an exclusion zone. This file can contain several
polygons depicting the regions to be ignored.
To profit from the most nadir portion of the orthophotos, the user can select the “Nadir
Optimization” option. This allows improving the overall mosaic quality while increasing
processing speed. Note however that it will reduce the size of areas where seamlines can
be calculated and moved during mosaic editing.
The output of the mosaicing process consists of a single image divided in multiple blocks,
themselves divided into multiple tiles. Files required to manually edit the mosaic seamlines
and colors are automatically generated during mosaicing and are stored in subfolders to
the mosaic folder. The software has therefore its own structure to store the files. Users
who desire to specify their own mosaic tiling and format have to export the mosaic, either
during the mosaic creation process, or during mosaic editing. This functionality is described
in section 10.7.
To create a mosaic, follow these steps:
1. Make sure the project includes orthos.

2. Select "Mosaic Creation…" in the "Process" menu.
3. If only some portions of the orthos need to mosaiced, check "Bound Mosaic Area"
and specify the file using the corresponding drop-down menu.
4. Specify the size of the feathering zone in pixels.
5. Specify the minimum number of blocks
6. If color balancing is required, check "Color Balancing".
7. If nadir optimization is required, check "Nadir Optimization".
8. If an exclusion zone is specified, check "Exclusion Zone File" and specify the file
using the corresponding drop-down menu.

9. If the mosaic needs to be exported after the process, check "Export Mosaic after
Completion". A dialog will be displayed before the process to specify export
options.
10. Click on the "Process" button. Real-time feedback will be provided on the overall
Figure 20: Mosaic creation dialog.
7.7 Mosaic Editing

The mosaic editing module allows the auto-generated mosaic to be interactively adjusted
according to specific user requirements. It provides a smooth interaction with the data and
allows real-time updating during modifications, without having to regenerate the entire
mosaic completely. This design reduces the number of manual operations required,
significantly increasing the overall productivity.
Correlator3D™ also allows for limitless users to simultaneously edit a mosaic without
conflicts or duplication of any data. Therefore, multiple users can access the same mosaic
location, reducing significantly storage requirements and leading to fast and efficient
operation suited to production environments.
To edit a mosaic, please follow these steps:
1. Make sure the project includes a mosaic and the orthos used to generate this
mosaic.
2. Select “Mosaic Editing…” in the “Processes” menu. The mosaic and the
corresponding blocks will then appear in the main window.
Upon entering mosaic editing mode, users are presented with a view of the entire mosaic
and a block overlay. The block overlay reveals how the mosaic is split up. This view enables
users to apply global operations to the entire mosaic area including exporting seamlines,
exporting the mosaic and manually adjusting the colors.
Users can also edit the auto-generated seamlines by selecting which block(s) to work on.
Blocks are selected by clicking on them. While each mosaic block can be edited by a single
user, different mosaic blocks can be edited simultaneously by multiple users. In order to

provide multiple users seamline editing capability, Correlator3D™ provides an automatic

locking system. Blocks already in use cannot be selected and will be shaded in red to
indicate they are locked by another user. Otherwise, if no blocks are in use, then they will
all be shaded in green to indicate they are available.
Please note, while users do not need to manually select them, all blocks must be available (i.e.
not locked) to export seamlines, export the mosaic or manually adjust colors.
7.7.1 Export Seamlines

Users can export the seamlines to an ESRI Shape file by pressing the corresponding button
in the toolbar.
7.7.2 Adjusting the Colors

Users can adjust the global colors of the mosaic using the Color Adjustment tool to enhance
its visual appearance. This tool allows modifying the levels of the mosaic, i.e. making
adjustments to the brightness, contrast, and gamma for each individual b or for all channels
together.
In the Channel drop-down list, users can select a color channel to edit. Depending on the
number of channels of the imagery, users can choose the "Master" channel to edit the red,
green, blue and alpha channels all together, or choose "Red", "Green", "Blue" or "Alpha" to
edit the channels individually. Note that selecting the "Alpha" channel will also allow users
to display this channel in the main window. For grayscale mosaics, only one channel can
be selected.
The original histogram of the mosaic for the selected channel is displayed in blue in the
dialog, while the new histogram to be applied is overlapping in red. A preview of the
adjusted mosaic is displayed on the mosaic in the main window. This option can be toggled
at all times in the color adjustment dialog.
To increase the contrast of the image, users must drag the input levels sliders, which show
the image’s brightness values. To darken the darkest values, the left slider must be dragged
to the right; all values below this value become zero (black). To lighten the lightest values,
the right slider must be dragged to the left; all values above this value become 255 (white)
or 65,535 for 16-bit imagery. To change the value of medium grey, the gamma slider must
be dragged left or right; the range for of the gamma value is 0.1 to 10.0, with 1 being in the
middle.
To decrease the contrast, users must drag the output levels sliders. To lighten the darkest
pixels, the left slider must be dragged to the right. To darken the lightest pixels, the right
slider must be dragged to the left.
For 16-bit imagery, users can activate the "Stretch" mode of the main window. Doing so when
the Color Adjustment dialog is displayed will automatically stretch the histogram of the Master
channel and therefore provide a quick way of adjusting the colors over the 16-bit range.

Adjustments to the mosaic on disk will not be applied until the "Process" button is clicked.
Note that once the process is started, it cannot be stopped since all mosaic tiles on disk
must be adjusted with the same parameters.
Figure 21: Color adjustment dialog.
7.7.3 Seamline Editing Mode
After the blocks to be edited have been selected, press the "Seamline Editing" button to
enter Seamline Editing Mode.
Note that users can toggle back to the mosaic overview to select different blocks by
pressing the "Block Section" button.
To ensure smooth operation and to minimize memory requirements, it is recommended to select

the fewest blocks possible when entering into seamline editing mode.
There are two viewing modes in which users can interact with the project. In both modes,
users can zoom in, zoom out or pan the images the same way they interact with a DEM.
The first mode displays the chosen blocks and the corresponding seamlines and is used to
visually inspect this mosaic section and locate seamlines to be edited.
The second mode is used to edit a particular seamline. Users can access this mode by
clicking on the seamline to be edited. Correlator3D™ will then load the images
corresponding to each side of the seamline. One side is displayed as a bright image and
corresponds to the mosaic region that was generated prior the addition of the new
orthorectified image. The other side is displayed as a dark image and corresponds to the
new orthorectified image that is being added to the mosaic. The overlap region between
both images is displayed with the original color.

In order to accelerate the loading time, only orthorectified images corresponding to the current
view region are displayed. Therefore, when editing a seamline, some images may be missing if
users pan outside the region.
Correlator3D™ displays seamlines according to a particular color scheme. A color is

assigned to each segment of a seamline depending on its quality, i.e. as determined
automatically during the mosaic creation. If the software is confident that the seamline
follows a path that minimizes the artifacts in the final mosaic, it will be green. If not, its
color will be red and if the confidence level is in between, the seamline will be yellow. Note
seamlines that have been previously edited will be blue. Users can therefore rely on the
color scheme to have a first appreciation of the mosaic and determine which seamlines
should be edited first.
For 16-bit imagery, users can also stretch the histogram to improve brightness and contrast
during mosaic editing. It can be enabled or disabled by selecting “Display Style->Stretch”
in the “View” menu. Note that enabling this mode will not modify the colors of the mosaic
on disk.
Users can also display the mosaic tile definition by toggling the corresponding button. The
name of each tile as well as its boundaries will be displayed as an overlay to the mosaic. If
a tile overlay was not provided by the user, then this option will instead display the
software's own internal tile definition for the blocks.
7.7.3.1 Interacting with Seamlines

Once a seamline has been selected, users can modify its path using two different tools.
These tools can be accessed by clicking on their corresponding buttons in the mosaic
editing tool bar or from the edit menu. Users can also switch from one tool to the other by
pressing the “Tab” key on the keyboard.
The first one, the hand tool, is used to move, add, or delete seamline points. A point can be
moved by clicking on its vicinity with the left button of the mouse and dragging it at a new
location. Conversely, it can be removed by clicking on the right button. New points can be
added by clicking on the right button of the mouse anywhere on the seamline.
Depending on the scale factor, Correlator3D™ will adjust automatically the level of details of
the seamlines. Therefore, if users zoom in the seamlines, additional seamlines points will be
shown.
The second tool, the scissors tool, is used to remove a series of seamlines points and
replacing them with new ones. A first point must be selected with clicking on the left button
of the mouse, and then a second one to determine the series of points to be removed. New
points can then be added to follow a new path, starting from the last point to the first one.
When moving points of a seamline, a red region is displayed around the seamline. This region
corresponds to the feathering zone, i.e. the region where a progressive transition is made
between an image and another. Users should take into account this region when defining the

path of a seamline since it may result into blurry results in the final mosaic. For example, the
feathering zone should not pass nearby an object if this object does not appear in both images.
When modifying the path of a seamline, Correlator3D™will automatically update the image to
represent what the mosaic should look like, without feathering and color balancing being
applied.
If there is a seamline traversing multiple blocks, then only the section of the seamline in the
currently loaded blocks will be available for editing. Furthermore, modifying the point at the
boundary of the block will not be possible. Likewise, changing the feathering size will not be
possible unless all the blocks containing the seamline are loaded.
The last operation on a seamline can always be undone by pressing the corresponding
button in the mosaic editing toolbar or using the keyboard shortcut “Ctrl-Z.”
Users can also modify the size of the feathering zone for a specific seamline. Once the
new size has been set, it will first modify the size of the red region that is displayed around
a seamline when moving its points. Note that modifying the size of the feathering will have
a great impact on the processing time when updating the mosaic since it has to be applied
on its full length. Also note that the new size only affects the currently selected seamline,
and that a different size can be specified on every seamline.
Note, the feathering size can only be changed for an entire seamline. Hence, if a block contains
only part of that seamline, then the feathering size cannot be adjusted. Users will need to load
all blocks containing a seamline in order to change its feathering size.
7.7.3.2 Updating the Mosaic

Once users are satisfied with a seamline, modifications must be saved by pressing the
corresponding button in the mosaic editing toolbar or the keyboard shortcut “Enter”. The
display will then automatically switch back to the mosaic viewing mode. Each modified
segment of the seamline will be displayed in blue, and the region of the mosaic that will be
affected by the modifications will be highlighted in cyan.
Users can then update the mosaic by pressing the corresponding button in the mosaic
editing toolbar. Correlator3D™ will only process the new region, with feathering and color
balancing being applied. This functionality removes the necessity of regenerating the whole
mosaic and allows users to get the final result in a matter of seconds.
Modifications of a seamline can also be discarded by pressing the corresponding button in

the toolbar or the keyboard shortcut “Esc”. The display will then automatically switch back
to the mosaic viewing mode with the original seamline being displayed.
Table 5 summarizes the keyboard shortcuts that can be used when editing seamlines.

Action Keyboard Shortcut

Undo the last operation on a seamline Ctrl-Z
Change current editing tool Tab
Save modifications to a seamline Enter
Discard modifications to a seamline Esc
Table 5: Keyboard shortcuts for mosaic editing.
7.8 3D Model Generation

Correlator3D™ can generate photorealistic 3D models. In a standard workflow, the process
will create a mesh from a DSM, then uses input images to texturize it. Other processing
scenarios are possible. For example, if an existing 3D model is selected from the project
tree, this model will be used as a seed: the mesh will remain the same or will be simplified
to meet the new size requirement. If the “Generate Texture” option is selected, a new
texture will be generated.
One important aspect of a 3D model is its size. For standard third-party 3D viewers, the
3D model that is loaded must fit entirely into the memory of the GPU for visualizing it
smoothly. As a result, Correlator3D™ provides a slider bar for the user to choose the size.
A coarse 3D model will take a minimum of resources when viewing it, but will contain less
mesh complexity and texture details. Conversely, a 3D model with a maximum size will
have a greater workload on the GPU but should have more quality. The user can choose
the size of the resulting 3D model using the slider bar, or ask for the “Optimal” value, which
corresponds to a compromise between quality and size. Note that the range of the slider
bar will be adapted to the available GPU memory on the computer used for processing.
Depending on the selected options, the “Notes” section of the dialog will be updated to
provide a precise description of the process and its expected outputs. It will indicate which
model is going to be used as an input to the process (a DSM or a 3D model), the
approximate number of triangles in the final mesh, the number and size of the texture files
(if a texture is to be generated), and finally the expected GPU memory workload of the
resulting 3D model.
The following steps must be followed to perform 3D model generation:
1. Select a DSM (or a 3D model) from the project tree.

2. Select “3D Model Generation…” in the “Processes” menu.
3. If a mesh needs to be created from the DSM, check “Create Mesh from DSM”.
Otherwise, the mesh from the current 3D model will be used as an input
4. If textured is required, check “Generate Texture”.
5. Specify the model size using the corresponding slider bar. Press the “Optimal” to
let the program calculate an optimal value.
7. Save the generated 3D model by selecting “Save 3D Model As…” using the
contextual button in the project tree.

Figure 22: 3D Model generation dialog.

8. Automated Workflow
The automated workflow is an automatic solution to the Correlator3D™ data production
process. This tool is used to perform the complete Correlator3D™ workflow:
Aerial Triangulation
DSM Generation
DTM Extraction
Orthorectification
Mosaic Creation
3D Model Generation
To start processing, the user must first select a processing profile. Five profiles are available
corresponding to common processing needs:
Quick Project Overview: generates all outputs at coarse resolutions.

True Ortho: generates all outputs at optimal resolutions and a true
orthomosaic.
Classic Photogrammetric: generates all outputs at optimal resolutions and
a DTM-based orthomosaic.
Fast Ortho: generates coarse DEM and an orthomosaic at an optimal
resolution.
3D Model: generates a 3D model.
Each profile has its own set of parameters and options. For specific requirements, the user
can customize the options of a profile. It is important to note that each processing step is
optional. By simply selecting or de-selecting a checkbox, the user can perform desired
actions. For example, the user can start the automated workflow from DTM up to mosaic
by using only the appropriate checkboxes.
When selecting the “Generate script”, the user can create a script with the chosen actions
and parameters and skip the processing. The script is saved to a text file (Project.spt) in the
project folder. This script can then be adapted to specific requirements, where other
variables and parameters can be changed from the default settings. The user can then
execute the script from the file menu. Further details can be found in section 11.

Figure 23: Automated workflow dialog.

9. Feature Extraction
The feature extraction module enables the user to extract 3D features from a DSM by
creating 3D polygons. Automatic contour detection of buildings and other features enable
fast footprint collection. Intelligent analysis aids the user while editing features, using the
3D information in the DSM to automatically calculate how well planes fit the segmented
polygon and providing visual feedback. Feedback can also be provided through real-time
true orthorectification to achieve pixel accuracy. This semi-automatic design saves time
over traditional vector extraction methods.
To extract features, the following operations should be performed:
1. Select the DEM you wish to use for feature extraction from the project tree.
2. Select “Feature Extraction…” in the “Mode” menu.
3. Check the Edit/Append box to edit an existing vector file. Specify the vector file
using the corresponding drop-down menu. The vector file must be an existing ESRI
shapefile containing 3D information.
4. Check the "Real-time True Ortho" box to enable this feature. Specify the DTM file
5. Click on the “Process” button to start extracting 3D features. The resulting vector
information will be saved in a new vector file once the user chooses the save
option in the feature extraction menu.
Figure 24: Feature extraction dialog.
If loading an existing vector file containing building footprints, the user will need to select/edit
each building to assign a 3D polygon.
9.1 Extraction Modes

There are four modes in which the user can interact with the DSM. These four modes are
based on feature type which includes buildings, roads, water bodies and forests. The
default mode is buildings. The user can switch modes by clicking on the Process menu and
selecting the mode of choice.

9.1.1 Building Extraction Mode
Should the user wish to extract buildings must first ensure they are in the building
extraction mode. A building is selected in 2 steps. First, pan and zoom to the area of
interest. Second, make a rough selection around the building by clicking around the building
perimeter. The software will automatically squeeze the selection around the contour of the
building as shown in Figure 25.
The user can go back to the original selection by selecting "Undo" in the "Edit" menu.
Figure 25: Contour around a building.
The user then has the option to modify the contour according to their requirements using
the hand tool. The hand tool is used to move, add, or delete points. A point can be moved
by left-clicking and dragging it to a new location. Conversely, it can be removed by right-
clicking with the mouse. New points can be added by right-clicking anywhere on the
contour.
Holding the control (Ctrl) button while moving a line segment will restrict the movement of that
segment in such a way that the line segment moves either in a direction parallel to itself or in a
direction along itself (collinear with itself). A special case occurs when moving an interior line
segment contained within the building footprint. In this case, the line segment will also resize
itself to fit within the building footprint. Note, however, the line segment may not resize itself if
the building footprint is irregular.
The edge merging option is enabled by default. When enabled, edges will be merged
together as long as they are collinear and represent the same line.
When using the hand tool to add a point, that point may “disappear” if it was added to a straight
line. This is because the edge merging tool is enabled. The edge merging tool “sees” this point as
splitting a line segment into two collinear line segments and consequently merges them. Hence,
the edge merging tool should be disabled to split a straight line into two parts.
All angles of a selected building contour that approach ninety (90) degrees can be rectified.
The rectify operation changes these angles to be as close to ninety (90) degrees as possible.

The software automatically extracts 3D information from the DSM and determines the z-
value of each vertex of the polygon bounded by the contour. The polygon starts off as a
plane representing the building. In most cases a building’s roof is more complex than a
plane. For these situations, the “Add a surface” tool can be used to add further details to
the building polygon. The ‘Add a surface’ tool is used to segment the building polygon into
more planes to better represent the roof geometry. To segment the plane, left click and
hold on any contour line of a building and drag the mouse to see a dotted line indicating
the plane segmentation as shown in Figure 26.
Figure 26: Plane segmentation indicated by the dotted line.
The newly created dotted line indicates the building polygon is divided into two sections.
The software automatically recalculates the new plane to be applied to each section and
determines the z-value of each vertex. It uses the 3D information in the DSM for this
calculation. The fact that the line is dotted indicates that the two new planes touch at their
boundaries (e.g. two planes that create a roof slope). If the line was solid it would indicate
that the boundary of one plane meets the other plane in an area other than the boundary,
creating a discontinuity in elevation (e.g. when a small building is connected to a taller
building). In other words, the software automatically uses the building contour (the outer
most border of the building) and the dotted/ solid lines to fit planes within those
boundaries. The user should segment the building polygon to create as many planes as
necessary to represent all the various slopes in the roof geometry of a particular building.
In addition to segmenting planes by clicking on the edges, points can likewise be dragged
to divide the polygon into different slopes. New points can also be added by right-clicking
on the feature lines.
If the surface merging option is turned on, planes that are coplanar and that represent the
same surface will be automatically merged together as the user is modifying the building.
Visual feedback is provided based on how well planes fit within their surfaces. In particular,
as the user is segmenting the polygon to better define roof geometry, the software
automatically calculates how well a particular surface can be represented by a single plane.
If a plane poorly represents a surface, then that plane is shaded red. Likewise, if a plane
represents a surface only marginally well, then it is shaded yellow. Otherwise, the plane fits
a surface well and is shaded green.

The best fit function uses the information in the DSM to improve the position of the planes.
The criterion for moving vertices is based on which configuration will result in the best fit
for the planes. Note this might not necessarily correspond to the best visual positioning of
the contour. If the position of the points is already optimal, then the best fit function will
not perform any further changes.
Note that elevation values for each vertex can also be manually set. Use this function to
select any vertex and change the elevation value.
9.1.2 Road Extraction Mode
This mode allows the user to extract road features. The user traces out the road using the
mouse. For each vertex of the polygon, the median elevation within a fixed boundary inside
the polygon is taken as the road height. This accounts for any banking or slopes the road
may traverse.
9.1.3 Waterbody Extraction Mode
This mode allows the user to extract water surfaces. The user traces out the water
boundary using the mouse. For each vertex of the polygon, the median elevation within a
fixed boundary outside the polygon is taken as the water height. This means that there
must be some ground clearing around the waterbody in order for the polygon to take on
appropriate values.
9.1.4 Forest Extraction Mode
This mode allows the user to extract forests. As in building extraction, this mode uses
contour fitting. Hence, the user needs only to make a rough selection around the forest
area and the software will automatically squeeze the selection to the forest boundary. The
elevation applied will be a local average at each vertex.
Table 6 summarizes the keyboard shortcuts that can be used when extracting features.
Action Keyboard Shortcut

Switch between Move a Point and Add a Surface Tab
Surface Merging Ctrl + U
Edge Merging Ctrl + E
Table 6: Shortcuts for feature extraction tools

9.2 Attributes
Attributes may be applied to each feature. To apply an attribute to a feature, the user needs
to first select that feature and enter corresponding information. New fields can also be
created to suit the user’s requirements. Moreover, an XML file can be provided containing
a user’s desired attribute list. The information in this XML file will be loaded automatically.
To specify the XML file, please select "Feature Attribute File" in "General Settings" which
can be accessed from the "File" menu.
9.3 Real-time True Orthorectification

A true ortho can be created for a small region in real-time around a selected feature. To
allow for true orthorectification, the necessary inputs need to be specified when entering
feature extraction mode. The user then proceeds to the feature extraction process. At any
time, the user can enable true ortho generation by toggling the "Display real-time true
ortho" button. Note, as the user continues editing the selected feature, the true ortho will
be automatically updated to reflect the changes. This enables the user to iteratively refine
the polygon for the selected feature until the true ortho is a perfect fit.

10. Tools
10.1 Import EO Data
Correlator3D™ features a tool to import EO data. The tool allows to load EO data stored
in a file (ASCII format) and to arrange it according to the SimActive format. In the original
file, the EO data must be ordered by columns and the tool assigns specific columns to
particular EO values (i.e. image name, easting, northing, elevation, omega, phi and kappa).
The interface also allows some rows or columns to be ignored.
The user must specify in which units the angular values are represented (degrees, radians
or gradians) and the tool will convert them, if necessary, to degrees. An angular offset can
also be added to kappa values by using the “Camera Mount Rotation” option. In addition,
if values in the “ImageName” column do not include a file extension (e.g. tif, jpg or bmp),
then the tool can automatically add one using the “Add Extension to Image Name” option.
The user must select a position and orientation system. The software operates using the
"Easting/Northing" position and the "Omega/Phi/Kappa" orientation systems. However, a
"Lat/Long" position and/or "Roll/Pitch/Yaw" orientation system may be specified. The tool
will then perform the necessary conversion. Note, the user must specify a projection to
allow conversion between different position and orientation systems. The orientation
system can also be set to "Automatic". This will set omega and phi angle values to zero, and
the kappa will be derived from the image positions.
Note, that the interface also allows the user to edit a particular cell value and to remove
undesired lines.
To create an EO file, the user should perform the following operations:
1. Specify an ASCII file by pressing the corresponding "..." button.

2. Specify the delimiter(s) to be used by the software to recognize the columns in
which EO data are represented.
3. Specify the input angular units.
4. If required, specify a camera mount rotation to add an angular offset to the kappa
values.
5. If required, specify an extension to be added to image names.
6. Select the appropriate position and orientation system.
7. If necessary, select the projection.
8. Assign EO data values to the appropriate columns by pressing on the buttons
located at the top of the displayed ASCII file.
9. Remove any undesired rows by clicking on the corresponding check box located
at the beginning of a line.

Figure 27: EO file creation dialog.
10.2 Seed DEM Creation

Correlator3D™ offers a tool to automatically create a seed DEM. The software creates a
DEM with the boundaries computed from the provided project and with every elevation
point set at the specified terrain height.
To create a seed DEM, the user should perform the following operations:
1. Select “Seed DEM Creation…” in the “Processes” menu.

2. Specify the average terrain height.
3. Click on the “Process” button. Real-time feedback on the overall progress of
processing will be provided. Once the process is done, the seed DEM will appear
in the main window.
Figure 28: Seem DEM creation dialog.

10.3 Fiducial Detection

Correlator3D™ features a tool for automatically detecting fiducials. The tool is useful for
extracting the position of marks (in pixels), which are necessary for using any processing
module of the software with scanned films. The fiducial detection tool is shown to the
user when creating a project with scanned films.
Correlator3D™ supports fiducial marks that consist of two lines intersected in a

perpendicular manner. These two lines can be continuous or not and the corresponding
cross mark can be oriented in any way. Also, note that fiducials are assumed to be bright
on a black background.
Fiducial detection is performed in four steps:
1. The user enters the required information in the fiducial detection dialog.
2. The user manually selects the fiducial marks for the first image.
3. Correlator3D™ automatically detects the fiducial marks for the remaining images.
4. Correlator3D™ automatically validates the detection results and prompts the user
to manually select the fiducial marks for faulty images.
Detection results can be verified visually using images that the software automatically
generates (these can be found in the temporary folder as described in Section 12.3). The
figure below shows an example of a successful fiducial detection. Note that the center of
the green mark represents the location where the software located a fiducial mark.
Figure 29: Sample fiducial detection results.
Fiducial marks that are incorrectly detected can be easily identified by looking at these
images. The following figure presents such a situation. In that case, the user should
manually correct the resulting exterior orientation file. Note that a warning will be reported
in the log file (Section 12.1) when there was an error in the detection process.
Figure 30: Sample fiducial detection results with one mark incorrectly identified.

The user can skip the automatic detection phase and manually select the fiducial marks.
To use the fiducial detection tool, the user should perform the following operations:
1. Drag each of the four images until a fiducial mark appears in the center of each
frame under the green marks.
2. Check the "Skip automatic detection" if all fiducial marks need to be selected
manually.
3. Click on the “OK” button when all fiducial marks are correctly selected.
4. Automatic detection is performed and real-time feedback on the overall progress
of processing is provided.
5. Automatic detection of faulty images is performed. If necessary (or if the "Skip
automatic detection" was checked), the fiducial detection dialog for manual
selection appears and these steps must be performed:
a. Drag the faulty fiducial marks under the corresponding green marks.
b. Click the “Next” button.
c. Repeat steps a) and b) until process is completed.
When using the fiducial detection dialog for manual selection, the size of the fiducial
marks can be adjusted by using the “+”, “-“ and “1:1” buttons. The fiducial marks are
selected by dragging their image portion over the intersection of green lines, either by
using the mouse or the arrow keys on the keyboard. If the arrow keys are used, “Shift”
and “Ctrl” keys can serve for coarser, but faster positioning, while the arrow keys alone
are better suited for fine-tuning purposes.
Figure 31: Fiducial detection dialogs for manual selection. On the left, all fiducial marks
were correctly selected. On the right, a faulty detection must be corrected.

10.4 GCP Creation

Correlator3D™ features a tool to create GCPs for UAV/medium/large format and satellite
sensors. The GCPs are used during the aerial triangulation process. The tool enables the
user to specify pixel coordinates of ground control points in a visual environment. GCPs
may be created by specifying their positional coordinates. The software uses these inputs
to automatically determine which images contain the specified GCP. These images are
then displayed in the visual interface and the user is requested to pinpoint the locations of
ground control points.
The GCP creation tool is only available in the AT module.
10.4.1 Creating and Importing GCPs
GCPs can be created using the project tree or imported from a text file. To create a GCP,
the user must hover the cursor over the GCP portion of the project tree, and press the
“Add a new point” contextual button. When GCPs are already specified in a text file (in
ASCII format), they can be imported using the project creation wizard or directly in the AT
module. The file should be organized in columns such that the tool can assign a specific
column to a GCP field (i.e. GCP ID, X, Y, and Z). The interface also allows ignoring some
rows or columns. Imported GCPs are then displayed in the project tree where they must
be selected for editing in order to add observation data.
The following operations must be performed to import a GCP file:
1. In aerial triangulation mode, select “Import GCP File…” in the “Processes” menu.
2. Specify an ASCII file by pressing the corresponding "..." button.
3. Specify the delimiter(s) to be used by the software to recognize the columns in
which GCP data are represented.
4. Specify the projection system of the coordinates to be imported.
5. Choose if the new GCPs should be added to the existing ones or should replace
them.
6. Assign GCP data values to the appropriate columns by pressing on the buttons
located at the top of the displayed ASCII file.
7. Remove any undesired rows by clicking on the corresponding check box located
at the beginning of a line.
8. Click on the “Import” button.
9. Add observation data to each GCP by editing them using the corresponding button
in the project tree.
10.4.2 GCP Editing Dialog
The GCP Editing dialog is used to edit GCP properties, such as its ID, description and
coordinates, and to tag it in the images. The dialog is divided into four sections:

- Information: this section is located at the top of the window and contains the GCP
properties. For convenience, it can be collapsed using the blue triangle button in
the upper-right corner.
- Observations: this section occupies the largest part of the window. This is where
the user can adjust the location of the GCP in the images.
- Ribbon: this section is located under the observation section. It contains an
overview of all the observations of the GCP along with other images in which the
GCP is likely to be found.
- Zoom and Navigation: this section can be found at the bottom of the dialog. It
contains controls to help navigate the images.
Figure 32: GCP Editing window.
There are four (4) GCP types which are "GCP", "GCP_XY", "GCP_Z" and “check point”. The
first three types are used to distinguish between varying forms of positional certainty. For
example, the user may need to specify a GCP for which they are only certain about the Z
component. In this case, GCP type "GCP_Z" will be used. Note, the user will still have to
specify an approximate X and Y component. The effect of this is that the software will now
treat this GCP accordingly in the aerial triangulation process, placing a high importance on
the Z component, while allowing room for error in the X and Y components. Check points
are GCPs that are not taken into consideration when performing bundle adjustment. They
are generally used for quality assessment.
Certain Uncertain
GCP XYZ None
GCP_XY XY Z
GCP_Z Z XY
Table 7: GCP type description.

The user has the option to add uncertainty information. As described in in section 7.1.4,
when the uncertainty values are left blank, the general uncertainty values will be applied
during bundle adjustment. If an uncertainty value is specified in the GCP Editing dialog, it
will overwrite the general uncertainty values for this specific point. This way, the user can
specify a global uncertainty for all points and redefine it on a point-by-point basis. If
uncertainty is used, it must be specified for all three coordinates (X, Y, and Z).
Once coordinates are specified, Correlator3D™ will project the point in each image and
populate the ribbon with the images that most likely contain the GCP. The software uses
knowledge on already identified GCPs and current IEO parameters to select the best image
candidates. Consequently, each time a new GCP is added to the project, Correlator3D™
becomes better at selecting candidate images.
The observation section of the dialog will display up to three images depending on the
window size. These are used to adjust the GCP location in each image. Initially, an image
candidate has an “unspecified” state. This state indicates that the user has not yet validated
nor discarded the image. To validate an image, the user must adjust the GCP location in
that image by clicking and dragging the image until the green aim is on top the GCP.
If the GCP is already at the proper location, the image can be validated by pressing the
green aim button on the top-right corner of the image.
An image can also be discarded by checking the red discard button next to the green aim
button.
Unchecking both the aim button and the discard button will revert the image back to the
unspecified state. Images in the disabled or unspecified state will not be used during bundle
adjustment as there is no functional difference between both. The disabled state exists to
let the user differentiate images that have been rejected from those that require attention.
The ribbon is used to navigate around the image candidates. The blue rectangle indicates
which images are currently displayed in the observation section of the window. To change
the images being edited, the user can drag and drop the ribbon at the proper location. An
alternate option is to use the arrows on the left and right side of the ribbon or those under
the ribbon.
When the ribbon is dragged completely to the left (therefore showing the rightmost part
of the ribbon) a special “+” button is visible. Clicking this button will add new image
candidates to the ribbon.
At any time, the user can jump to a specific image by selecting it in the combo-box located
above the left image in the observation section of the dialog. If necessary, the image will
be added to the ribbon.
Each time the location of the GCP is adjusted in an image, the software recalculates the
estimated location of the GCP in the images marked as “unspecified”. The ribbon is updated
in real time to reflect these adjustments. This means that each new observation for the
GCP improves the estimated location of the GCP in the other images.

Scrolling the ribbon is a fast and easy way to identify and validate the GCP observations.
The zoom options located in the lower-left of the window are used to adjust the zoom
factor applied to the images displayed in the observation portion of the dialog. Cancel and
save buttons are located in the lower-right of the window. Clicking “Save and Close” will
update the GCP. Clicking “Cancel” will discard the changes and close the window.
To edit a GCP, the user should perform the following operations:
1. Fill the required ID (name), type, projection and XYZ coordinates properties. If
required, add a description and an uncertainty.
2. Mark which images are to be used with the GCP.
a. Use the ribbon to navigate between the images. Use the “+” button to add
images if necessary.
b. Adjust each image to tag the exact GCP location. Use the zoom options to
help find the exact location of the GCP.
3. After all the observations have been added click the “Save and Close” button to
update the GCP file.
10.5 DEM Merging Tool

Correlator3D™ DEM merging tool allows combining DSM or DTM results that were
created separately. It is useful to process a project in multiple parts (e.g. at different times
or in parallel on multiple machines) and merge different tiles of DEMs together afterward.
It can also help to recover intermediate results from individual pairs when, for some reason,
the DSM process is interrupted. Correlator3D™ implements an intelligent data fusion
approach that leads to consistent data sets, even when there are minor discrepancies along
adjacent tile boundaries.
The DEM merging tool can be accessed from the Tools menu and does not require any
DEM to be loaded in the main window. It can merge files stored in different formats and
horizontal resolutions. Note that when merging DEM sets having different horizontal
resolutions, the resulting data will automatically inherit the finest resolution. Also, if there
is no overlap between adjacent data sets, Correlator3D™ will set the corresponding
elevation values as undefined. An option to fill gaps in the resulting DEM is also available.
This option allows the user to get a DEM in which small areas where elevation values were
left undefined are filled with interpolated terrain data. The final merged data will be
displayed in the main window and can then be saved by the user.
To merge DEMs, the user should perform the following operations:
1. Select “DEM Merging…” in the “Tools” menu.

2. Specify the files to be merged by clicking on the “Add Files…” button. A file
browser will appear. Select the files to be added and click “Open”. The list will
then be updated. Note that all files can be removed from the list (e.g. in case of an
error) by pressing on the “Clear List” button.

3. Click on the “Merge” button. Real-time feedback will be provided on the overall
progress of processing. Once the process is done, the resulting DEM will appear
in the main window.
4. Save the merged DEM by selecting “Save DEM As…” in the “File” menu.
Figure 33: DEM merging dialog.
10.6 Export EO and Camera Parameters

Correlator3D™ can export the EO information and camera parameters. In order to perform
this, the user must specify a folder in which the information is saved. In the case of a frame-
based project, one EO file and one calibration file per camera will be created. For satellite
projects, one RPC file will be produced for each image. These files will be exported as
simple text files, which can be used later on in the software.
10.7 Export Mosaic

Users can export the final mosaic using a tile definition file and selecting the type of mosaic
they want to create. Either a standard mosaic or a NDVI (Normalized Difference Vegetation
Index) map can be created. If the latter is selected, red and infrared bands must be available
in the original mosaic. Users can specify different settings for the resulting NDVI map by
clicking on the corresponding button, including the type of color map to be used, the
number of classes (colors), as well as the bounds of the index.
Users can either specify a fixed size for the tiles, individual tile size and coordinates through
a tile definition file or ask for a single tile. If a fixed tile size is specified in the dialog, each
tile will have that size, except for tiles located on the border of the mosaic. In such cases,
the size of the tile is simply reduced to remove unnecessary undefined areas. The tiles are
created to be juxtaposed one after the other with no overlap between two adjacent tiles.
The naming of the tiles describes the spatial relation between each other. The name
convention is as follows: Name_LineNumber_ColumnNumber.tif.
The position and orientation of the input orthorectified images can lead to situations where
some tiles only contain undefined data (completely black areas). When specifying a fixed tile size

in the Export Mosaic dialog, if such a situation occurs, the corresponding tiles are not created by
Correlator3D™. This behavior is part of an optimization effort to reduce memory and disk
requirements while improving overall performance.
If a tile definition file is used, then the tile size specified will be respected, even if the
resulting area contains unnecessary undefined areas. Moreover, each tile will be created
even if they do not contain any data at all.
Users can also export only a portion of the mosaic using the "Crop Area" option, as defined
by a user created polygon stored in a DGN file or a Shapefile.
To export a mosaic, the following steps must be performed:
1. If a Crop Area is used, specify the constraint file using the corresponding drop-
down menu.
2. If a Tile Definition File is used, specify the Tile Definition File using the
corresponding drop-down menu.
3. Specify the Mosaic Folder that will contain the mosaic using the corresponding
drop-down menu.
4. If no tile definition file is used:
a. Specify the name for the mosaic.
b. Specify the tile size.
5. Specify the format for the mosaic tiles.
6. Click on the “Process” button. The mosaic will then be exported using the
parameters specified.
Figure 34: Export mosaic dialog.

10.8 Export DEM

10.8.1 Export DEM Tiles
Correlator3D™ can divide a DEM in multiple tiles. The user can either specify a fixed size
for the tiles or specify individual tile sizes and coordinates through a tile definition file. If
a fixed tile size is specified in the dialog, each tile will have that size, except for tiles located
on the border of the DEM. In such cases, the size of the tile is simply reduced to remove
unnecessary undefined areas. The tiles are created to be juxtaposed one after the other
with no overlap between two adjacent tiles. The naming of the tiles describes the spatial
relation between each other. The name convention is as follows:
Name_LineNumber_ColumnNumber.tif.
The shape of the input DEM can lead to situations where some tiles only contain undefined data
(completely black areas). If such a situation occurs when specifying a fixed tile size in the Export
DEM Tiles dialog, the corresponding tiles are not created by Correlator3D™. This behavior is part
of an optimization effort to reduce memory and disk requirements while improving overall
performance.
If a tile definition file is used, the tile size specified will be respected, even if the resulting
area contains unnecessary undefined areas. Moreover, each tile will be created even if they
do not contain any data at all.
The format of the output DEM tiles can be specified in the dialog. If a definition file is used,
the format is specified through the extension of the DEM tiles.
The tool can be accessed from the File menu and requires a DEM to be loaded from the
project tree.
To export DEM tiles, the following operations should be performed:
1. Select a DEM from the project tree. The DEM will be displayed in the main
window.
2. Select “Export” -> “DEM” -> “DEM Tiles…” in the “File” menu.
3. If a tile definition file is used
a. Check “Tile Definition”.
b. Specify the tile definition file by pressing on the corresponding "..." button.
.
4. Specify the folder that will contain the tiles by pressing on the corresponding "..."
button.
5. If no tile definition file is used
a. Specify the name for the mosaic.
b. Specify the tile size.
6. Specify the format for the DEM tiles.

Figure 35: Export DEM tiles dialog.
10.8.2 Export LIDAR LAS and ESRI Shapefile

Correlator3D™ can export a DEM into a LIDAR LAS file or an ESRI Shapefile. Using these
formats, elevation grids will be exported as a series of 3D points. The tools can be accessed
from the File menu and requires a DEM to be loaded in the main window.
To export a DEM into the LIDAR LAS or ESRI Shapefile format, the user should perform
the following operations:
window.
2. Select “Export DEM…” in the “File” menu.
3. Select “Export” -> “DEM” -> “LIDAR LAS File…” or “ESRI Shapefile” in the “File”
menu.
4. Specify the name and the location of the file to be exported.
10.8.3 Export with Geoid

Correlator3D™ can export a DEM on which a geoid will be applied. The geoid file must be
provided in the GTX file format.
The tool can be accessed from the File menu and requires a DEM to be loaded in the main
window.
To export a DEM with a geoid, the user should perform the following operations:
window.
2. Select “Export” -> “DEM” -> “With Geoid…” in the “File” menu.
3. Specify the geoid file by pressing on the corresponding "..." button.
4. Specify the output DEM by pressing on the corresponding "..." button.

Figure 36: Export with geoid dialog.
10.9 DEM Inspection

Correlator3D™ DEM inspection tool is used to generate statistics about the quality of a
DEM against reference data, i.e. elevation points provided in a ShapeFile or a SimActive
GCP file. The software automatically computes the differences in elevation and provides
statistics including bias error, standard deviation, RMS error and maximum error.
Statistics are displayed in a window at the end of the process. They can be saved as a text file.
To export DEM tiles, the following operations should be performed:
1. Select “DEM Inspection...” in the “Tools” menu.

2. Specify the DEM to be inspected using the corresponding drop-down menu.
3. Specify the reference data using the corresponding drop-down menu.
Figure 37: DEM inspection dialog.

10.10 Change Detection

Traditionally, change detection is performed by comparing two images. This approach is
not robust in general since only changes in the appearance of the scene are detected and
such changes do not necessarily imply structural changes. Correlator3D™ allows change
detection by comparing DEMs instead of images. Such a comparison is possible since the
software can quickly and automatically process 3D data.
Correlator3D™ requires two DEMs, which should be generated from two different image
sets of the same area. It performs a comparative analysis and determines in 3D the
importance of changes. The software outputs change detection results as a DEM with
values corresponding to the differences in elevation between the two DEMs. The results
can be optionally outputted as color information superimposed on a mosaic of
orthophotos. This mosaic gets converted to grayscale and structural changes are
highlighted using a specific color scheme: negative elevation changes (e.g. objects that have
disappeared) are displayed in tones of yellow to red according to their importance (small
to large changes respectively) while positive elevation changes (e.g. new objects) are shown
in tones of dark to light blue (small to large changes respectively). A detection range in
meters is used and only changes that are within the specified range (vertically) will appear.
To perform change detection, please follow these steps:
1. Select “Change detection…” in the “Tools” menu.

2. Specify the reference DEM (i.e. the DEM as it appears before changes) using the
3. Specify the comparison DEM (i.e. the DEM as it appears after changes) using the
4. If a change detection mosaic is required:
a. Check “Display Changes on Mosaic”.
b. Specify the “Input Mosaic File” of the input mosaic using the corresponding
drop-down menu. This file can be either a mosaic definition file or a tile
definition file. Change detection results will be overlaid on the mosaic after
the processing is completed. This mosaic can correspond either to the
images taken before or after the changes.
c. Specify the folder where the change detection mosaic will be generated
If new 3D structures are mostly expected then the mosaic corresponding to the image
data taken after changes should be used. If it is anticipated that structures have in
general disappeared, then the input mosaic should consist of the image data collected
before changes have occurred.
5. Specify the detection range. This implies that only changes within that range
(vertically) will be detected.
progress of processing. Once the process is done, the change detection DEM will
appear in the main window.
7. Save the DEM by “Save DEM As…” in the “File” menu.

Figure 38: Change detection dialog.
10.11 Point Cloud Colorization

The Correlator3D™ point cloud colorization tool is used to colorize a point cloud using a
mosaic. The software automatically determines correspondence between 3D points and
pixels.
When colorizing a point cloud, it is strongly recommended to use an appropriate mosaic. As an

example, if a point cloud was generated for a dense urban area, then the mosaic should be
generated using true orthos, i.e. orthos generated on a DSM. Otherwise, discrepancies could
be observed between points and colors.
To perform point cloud colorization, the user must follow the listed steps:
1. Select the point cloud to be colorized from the project tree.

2. Select “Point Cloud Colorization” in the “Tools” menu.
3. Save the modified point cloud by selecting “Save point cloud As…” using the
contextual button in the project tree. Note that if the point cloud is saved under a
different name than its corresponding DEM, it will be excluded from the project
tree.
10.12 DEM Contour Extraction

The Correlator3D™ DEM contour extraction tool is used to generate equally spaced
contour lines from the currently loaded DEM and save them in a vector file. The contour
lines are extracted based on two parameters: the interval, which corresponds to the
elevation interval between contour lines, and the offset, which corresponds to the offset
from zero relative to which to interpret intervals.
To perform DEM contour extraction, the user can follow the listed steps:
1. Select the DEM you wish to use for contour extraction from the project tree.

2. Select “DEM Contour Extraction…” in the “Tools” menu.

3. Specify the contour interval and offset.
4. If a specific elevation attribute is required in the vector file, check the “Elevation
Attribute” option, and specify the name of the attribute.
6. Save the generated contour by selecting “Save contour” using the contextual
Figure 39: DEM contour extraction.

11. Script Mode

Correlator3D™ includes a scripting mode for advanced users who desire to fully automate
their processing tasks. A script simply consists of a plain text file containing a series of
instructions that the software will execute sequentially. The syntax is very simple: each
operation starts with the process name and all necessary input/output parameters are then
specified. Here is an example for a DSM generation process:
<DSMGeneration>
{
InputEO C:\MyProject\IEO\Step_1\Step_1.ieo
Resolution Optimal
VertAccuracy Precise
OutputDEM C:\MyProject\DEM\DSM.asc
}
Note that the order in which the process parameters are specified is not important. The
user can perform as many operations as desired. Input files that do not exist can be
specified as long as previous operations create them. For DSM generation and
orthorectification, the optimal resolution or overlap can be set by entering "Optimal" (not
case sensitive) in the corresponding field. Similarly, the user can set the maximal resolution
by entering “Maximal” or “Coarse” in the corresponding field. A description of the format
of script files is included in Section 0.
Once a script file is created, it can be run by selecting “Run Script…” in the “File” menu.
Real-time feedback on the overall progress of processing is provided during execution.
Please refer to Appendix A for the script format description.
11.1 Command Prompt

Instead of using the GUI to interact with Correlator3D™, the user may wish to use a
command prompt to run individual scripts and batch files quickly. Navigate to the
Correlator3D folder and type “C3D” to display the possible options. For example, the
following command:
C:\Program Files\SimActive\Correlator3D\C3D
will display the possible options on the command prompt screen provided the software
was installed in that directory.
To run a script:
C:\Program Files\SimActive\Correlator3D\C3D /f PathToMyScript
The “/f” must always precede the path to the script. The list of commands is summarized
in Table 8:

Command Action
/f Operator preceding path to the script or batch file
/v Outputs all messages
/m Outputs only the main messages
/p Outputs progress percentage to the screen
/r Saves message content to a file
Table 8: Command line operations summary.
As another example, consider the following command:
C3D.exe /f PathToMyScript.spt /v /p /r FileLog.txt 2>> Errors.log
This will make C3D.exe process the script file "PathToMyScript.spt". Since the “/v”
command was used, all the levels of information will be outputted. Normally, this
information would be displayed onscreen. However, the “/r” redirects this information to
the file “FileLog.txt”. The “/p” command outputs the progress of each operation in the
script. This information is only displayed onscreen and will not be redirected by the /r”
command. Lastly, the “2>>” is a batch command that will append error messages if any to
the file "Errors.log".

12. Troubleshooting
12.1 Log Window and Log File
Correlator3D™ includes detailed warning and error messages to facilitate troubleshooting.
In addition, all major operations performed during processing are displayed in the log
window. This window is fully dockable, allowing the user to customize their Correlator3D™
environment. Warning and error messages are color-coded in yellow and red respectively.
Figure 40: Log window.
The content of the log window is automatically saved in a log file. This can be particularly
helpful if, for some reason, the processing is interrupted. The file is located in the folder
“My Documents\SimActive\Correlator3D\Logs” and named “Correlator3D_Log.xml”. The
file can also be accessed directly from the user interface by selecting “Log File” from the
“Help” menu. A new log file is created each time the software is launched and logs
information until the program is closed. Copies of the twenty most recent log files are
automatically stored in the same folder.
12.2 Good Practice

As a first step, the user should perform a test on a few images before processing an entire
project. This will ensure that there is no problem with the input data. It will also allow
verifying that both the results and the required processing time meet expectations. Input
parameters (depending on the module used) can be varied, and small scale tests can be
repeated until satisfactory results are produced. Note that this recommendation does not
hold for the aerial triangulation module, as it requires as much input data as possible.
12.3 Temporary Files

Different temporary files are created during data processing. In the case of the DSM
generation process, each individual file, which is stored in the SimActive model format (.smf
files), contains DSM generated from a given pair. Temporary files are named according to
the first image of the pair processed. At the end of the DSM generation process, these
individual files are automatically merged together to produce a single DSM.
If the DSM generation process is interrupted in the middle of a large batch process, then
these files can be retrieved and merged manually using the DEM merging tool. The user

can change the location of the temporary folder by selecting “General Settings” from the
“Settings” menu and changing its location. For best software performance, this folder
should be located on a separate local hard drive, different from the one where input/output
data are stored.
In the case of the fiducial detection process, a visual representation of the results is also
created. For each input photo, an image showing the four detected marks is stored for
visual validation. These images are named according to input image.

13. Appendix A
13.1.1.1 Camera Calibration Types
Correlator3D™ supports four types of camera calibrations.
1. Digital Sensors
The first calibration type, designed for metric cameras, simply consists of providing the
following:
Image width in pixels

Image height in pixels
Pixel size in microns
Focal length in millimeters
Principal point coordinates in millimeters
Note that the imagery is assumed to be already corrected for distortion.
2. Digital Sensors with Distortion Table
The second calibration type is meant to specifically support camera systems using a
distortion table. The difference in the calibration file is that radial lens distortion is taken
into account when processing imagery. The parameter values are thus defined as a look-
up table (LUT) . These values specify radial distortion as a function of the pixel distance
from the image center.
3. Scanned Films
The third calibration type, designed for scanned films, contains additional fields (compared
to the digital sensor type) consisting of the coordinates in millimeters of the following:
Bottom-left fiducial
Top-right fiducial
Top-left fiducial
Bottom-right fiducial
Positions of fiducials and the principal point should be referenced to the principal point of
autocollimation as the origin.
4. Polynomial
The fourth calibration type specifically supports camera systems with radial and
decentering lens distortion. The parametric distortion model is defined by the following
equations, where X and Y are measurements in mm referring to the principal point and
r=√(X2 + Y2):
dX = dXradial + dXdecentering,

dY = dYradial + dYdecentering,
where,
dXradial = X (K0 + K1*r2 + K2*r4 + K3*r6 + ... + K7*r14),

dYradial = Y (K0 + K1*r2 + K2*r4 + K3*r6 + ... + K7*r14),
dXdecentering = (1+P3*r2+P4*r4)*(P1*(r2+2*X2)+2*P2*X*Y),
dYdecentering = (1+P3*r2+P4*r4)*(P2*(r2+2*Y2)+2*P1*X*Y),
and,
K0 to K7 are parameters associated with the radial distortion,

P1 to P4 are parameters associated with the decentering distortion.
13.2 Tile Definition File

A tile definition file can be inputted in Correlator3D™ Mosaic Creation module to define
the name, the coordinates and the size of each mosaic tile. A tile is defined by two corners
located on the diagonal of its area. Each tile is specified on a separate line containing the
following five (5) fields:
Tile filename
X coordinate of the first corner defining the tile area
Y coordinate of the first corner defining the tile area
X coordinate of the second corner defining the tile area
Y coordinate of the second corner defining the tile area
Corners can be specified in any order, as long as they are located on a diagonal.
Sample:
***********************************************************************
SimActive Tile Definition File
Copyright(c) 2009 SimActive Inc.
All rights reserved.
***********************************************************************
# Tile Name First Corner (x,y) Second Corner (x,y)
TileName_01 446762.0 5032005.0 447170.0 5031653.0
TileName_02 447170.0 5032205.0 447643.0 5031415.0
TileName_03 446862.0 5031653.0 447170.0 5031415.0
TileName_04 447061.0 5031415.0 447351.0 5031120.0
TileName_05 446762.0 5030920.0 447351.0 5031120.0

13.3 Script Format Description

This section describes the operations and parameters available in a Correlator3D™ script.
Script syntax is the following:
<OperationName>
{
ParameterName ParameterValue
...
}
The script file is case insensitive. This means that “DTMExtraction” and “dtmextraction” are
both valid.
The name of the operations must appear between ‘<’ and ‘>’ marks. And the parameter lists
is delimited by a pair of braces (‘{‘ and ‘}’).
There is no limitation on the number of white spaces (spaces, tabs and new lines) as long
as the ParameterName and its ParameterValue are on the same line.
Comments can be added in the file. Each line of comment must be preceded with ‘#’.
OperationName can be one of the following:

 AerialTriangulation
 DSMGeneration
 DTMExtraction
 Orthorectification
 MosaicCreation
 ExportMosaic
 ChangeDetection
 DEMMerging
 DEMTiling
 DEMInspection
 ExportGeoidDEM
 PointCloudColorization
The list of parameters and acceptable values varies according to the operation. For a given
operation, the order of the parameters has no importance. The description of each
available parameter follows.
Parameters for AerialTriangulation

Parameter name Type Description
TiePointExtraction ‘On’, ‘Off’ Indicates if tie point extraction must be performed.
CameraCalibration ‘None’, ‘Constrained’, Indicates the type of camera calibration.
‘Unconstrained’

EOAdjustment ‘None’, Indicates the type of EO adjustment.

‘BoresightCalibration’,
‘ConstrainedAT’,
‘UnconstrainedAT’,
‘DirectGeo‘
PerformBoresight ‘On’, ‘Off’ Bound boresight within total uncertainty. Only required if
EOAdjustment is 'ConstrainedAT'
ExtractionType ‘Standard’, ‘Exhaustive’ Type of tie point extraction to perform.
ImageFiltering Float, 'None', Indicates the type of image filtering.
'Automatic'
InputEO String Path to the input interior/exterior orientation file.
GCPFile String Optional. Path to the input ground control point file.
InputDEM String Optional. Path to the reference DEM. Required for
satellite projects.
ReferenceORIList String Optional. Path to the reference ORI file.
AdjustFocalLength ‘On’, ‘Off’ Optional. Indicates if parameter will be adjusted during
bundle adjustment.
AdjustXmm ‘On’, ‘Off’ Optional. Indicates if parameter will be adjusted during
bundle adjustment.
AdjustYmm ‘On’, ‘Off’ Optional. Indicates if parameter will be adjusted during
bundle adjustment.
AdjustK0 ‘On’, ‘Off’ Optional. Indicates if parameter will be adjusted during
bundle adjustment for odd/even order polynomial
distortion model.
distortion model.
distortion model.
distortion model.
distortion model.
distortion model.
distortion model.
distortion model.
AdjustP1 ‘On’, ‘Off’ Optional. Indicates if parameter will be adjusted during
distortion model.
distortion model.
distortion model.
distortion model.
AdjustLUT ‘On’, ‘Off’ Optional. Indicates if parameters will be adjusted during
bundle adjustment for LUT distortion model.
AdjustX Float Optional. Expected uncertainty in parameter to be
adjusted during bundle adjustment.
AdjustY Float Optional. Expected uncertainty in parameter to be
AdjustZ Float Optional. Expected uncertainty in parameter to be

AdjustOmega Float Optional. Expected uncertainty in parameter to be

AdjustPhi Float Optional. Expected uncertainty in parameter to be
AdjustKappa Float Optional. Expected uncertainty in parameter to be
AdjustGCPXY Float Optional. Expected uncertainty in parameter to be
AdjustGCPZ Float Optional. Expected uncertainty in parameter to be
AdjustGCPObs Float Optional. Expected uncertainty in parameter to be
ATFolder String Path to the AT folder.
OutputDEM String Path to the output coarse DEM file.
Parameters for DSMGeneration

Projection String Projection of the DSM in the PROJ.4 format. Required
for sensors other than frame-based. Invalid for frame-
based sensors since the projection is read from the input
EO file.
Resolution Float, ‘Coarse’, Required resolution of the resulting DSM.
‘Optimal’,
‘Maximal’
VertAccuracy Float, ‘Optimal’, ‘Fast’, Required vertical accuracy of the resulting DSM.
‘Precise’
GeneratePointCloud ‘On’, ‘Off’ Optional. Indicates if a point cloud must be generated.
OutputDEM String Path to the output DSM file.
3DConstraintFile String Optional. Path to the 3D constraint file.
Parameters for DTMExtraction

InputDEM String Path to the input DEM file.
3DConstraintFile String Optional. Path to the 3D constraint file.
ConstraintType 'Exclusion', Optional. Type of constraint to be used.
'Elevation'
DSMFeatures String Optional. Path to the output DEM.
OutputDEM String Path to the output DTM file.
Parameters for Orthorectification

InputDEM String Path to the DEM file.
DSMBased ‘On’, ‘Off’ Parameter used to generate true orthos from a DSM.
3DConstraintFile String Optional. Parameter used to generate true orthos using a 3D
constraint file.
OutputORIFolder String Path to the output folder.
Resolution Float, ‘Optimal’, Required resolution of the orthorectified images.
‘Coarse’
Overlap Integer, Required overlap between orthorectified images. Must be
‘Optimal’, commented out if CropWidth and CropHeight are specified.
‘Maximal’
CropWidth Float Image crop width. Must be commented out if Overlap is
specified.
CropHeight Float Image crop height. Must be commented out if Overlap is
specified.

OutputFormat ‘GeoTIFF’, ‘TFW’ Output format of the orthorectified images.

TiledTiff ‘Tiled’, ‘Untiled’ Indicates if the orthorectified images must be tiled or untiled.
Overviews ‘On’, ‘Off’ Indicates if the orthorectified images must be generated with
overviews.
Parameters for MosaicCreation

InputORIList String The default file name is the name of the IEO file used to
orthorectify the images.
MosaicArea String Optional. Path to the tile definition or vector file.
OutputFolder String Path to the output folder.
MosaicName String Name of the mosaic.
Feathering Integer Optional parameter. Default value is 37.
MinNumBlocks Integer Optional parameter. Default value is 1.
ColorBalancing ‘On’, ‘Off’ Indicates if color balancing must be applied.
OptimizeViewAngle ‘On’, ‘Off’ Indicates if nadir optimization must be applied.
ExclusionZone String Optional. Path to the vector file. The exclusion zone file is only
considered when color balancing is on.
Parameters for 3DModelGeneration

InputDEM String Path to the DEM file. Must be commented out if Input3DModel is
specified.
Input3DModel String Path to the 3D model file. Must be commented out if InputDEM is
specified.
InputEO String Optional. Path to the input interior/exterior orientation file. Only
required if GenerateTexture is 'On'.
Output3DModel String Path to the output 3D model file.
GenerateTexture ‘On’, ‘Off’ Indicates if texture must be generated.
ModelSize Integer, Required size of the 3D model. Numerical values must be specified
‘Coarse’, in megabytes (MB).
‘Optimal’,
‘Maximal’
Parameters for ExportMosaic

BlockDefFile String Path to the mosaic definition file.
MosaicType ‘Standard’, Type of the mosaic to be created.
‘NDVI’
OutputFolder String Path to the output folder.
TileDefFile String Optional*. Path to the tile definition file.
ExportArea String Optional*. Path to the area file.
MosaicName String Optional*. Name of the output mosaic.
TileWidth Integer Optional*. Tile width in pixels. Must be commented out if
TileDefFile is specified or SingleTile is 'On'.
TileHeight Integer Optional*. Tile height in pixels. Must be commented out if
TileDefFile is specified or SingleTile is 'On'.
SingleTile ‘On’, ‘Off’ Optional*. Indicates if the mosaic must be exported to a single tile.
Must be commented out if TileDefFile is specified.
OutputFormat ‘GeoTIFF’, Output format of the mosaic tiles.
‘TFW’
TiledTiff ‘Tiled’, ‘Untiled’ Indicates if the orthorectified images must be tiled or untiled.
Overviews ‘On’, ‘Off’ Indicates if the orthorectified images must be generated with
overviews.
ColorMap ‘Beige to Dark Optional**. The color map to be used.
Green’, ‘Dark
Blue to Dark
Green’, ‘DEM

Color Style’,
‘Grayscale’,
‘Red to Blue’
InvertColorMap ‘On’, ‘Off’ Optional**. Invert the color map.
MinimumThreshold Float, ‘Optimal’ Optional**. Minimum bound of the NDVI index.
MaximumThreshold Float, ‘Optimal’ Optional**. Maximum bound of the NDVI index.
ClassCount Integer Optional**. Number of classes for the index. If omitted, the
mapping is continuous.
*Either TileDefFile or MosaicName, TileWidth and TileHeight can be specified.
**Must be defined when mosaic type is set to NDVI.
Parameters for ChangeDetection

ReferenceDEM String Path to the DEM file before the changes.
ComparisonDEM String Path to the DEM file after the changes.
MinimumThreshold Float Minimum detection threshold (in model units).
MaximumThreshold Float Maximum detection threshold (in model units).
OutputDEM String Path to the output DEM file.
BlockDefFile String Optional. Path to the block definition file.
OutputFolder String Optional. Path to the mosaic output folder.
Parameters for DEMMerging

FillGaps ‘On’, ‘Off’ Indicates if the gaps in the resulting DEM must be filled.
Parameters for DEMTiling

OutputFolder String Path to the mosaic output folder.
TileDefFile String Optional*. Path to the tile definition file.
DEMName String Optional*. Name of the output mosaic.
TileWidth Integer Optional*. Tile width in pixels
TileHeight Integer Optional*. Tile height in pixels
OutputFormat 'GeoTIFF', 'AsciiGrid', 'SMFFormat' Format of the output DEM tiles.
'Shapefile', 'LIDARLAS'
*Either TileDefFile or DEMName, TileWidth and TileHeight can be specified.
Parameters for DEMInspection

ReferenceData String Path to the reference data file.
OutputFile String Path to the output statistics file.
Parameters for ExportGeoidDEM

GeoidFile String Path to the geoid file.

Parameters for PointCloudColorization

InputPointCloud String Path to the input point cloud file.
BlockDefFile String Path to the mosaic file.
OutputPointCloud String Path to the output point cloud file.
Parameters for DEMContourExtraction

Interval Float Interval between contours (in model units).
Offset Float Offset from zero relative to which to interpret intervals (in model
units).
ElevationAttribute String Optional*. Name of the attribute containing the elevation.
OutputFile String Path to the output contour file.
Parameters for DEMVolumeCalculation

InputArea String Path to the area file.
ReferenceType 'InterpolateDEM', The type of reference to be used to calculate volume.
'InterpolatePoly',
'FixedElev', 'Dig',
'Fill', 'RefDEM'
Offset Float Optional. Only used with 'Dig' and 'Fill' options.
Elevation Float Optional. Only used with 'FixedElev' option.
ReferenceDEM String Optional. Only used with 'RefDEM' option.
OutputFile String Path to the output statistics file.

Correlator3D User Manual

Uploaded by

Copyright:

Available Formats

Correlator3D User Manual

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Correlator3D User Manual

Uploaded by

Copyright:

Available Formats

Correlator3DTM User Manual

3. System Requirements .............................................................................................................................7

4. Data Requirements .................................................................................................................................8

© 2019 SimActive Inc. All rights reserved. 2

6.4.2 Project Parameters ..............................................................................................................28

9. Feature Extraction ................................................................................................................................57

© 2019 SimActive Inc. All rights reserved. 3

10.8.3 Export with Geoid ................................................................................................................73

© 2019 SimActive Inc. All rights reserved. 4

This symbol indicates an important notice. Failing to follow the notice

This symbol indicates a tip or helpful information. Following this

This manual uses the following abbreviations as summarized in Table 1:

DEM Digital Elevation Model

© 2019 SimActive Inc. All rights reserved. 5

Correlator3D™ consists of the following modules:

- Aerial Triangulation: exterior orientation correction and camera calibration

The software also provides other tools including:

­ Automated Workflow: automatic execution of Correlator3D™ workflow

DSM DTM Mosaic 3D Model

Figure 1: Sample geospatial data.

© 2019 SimActive Inc. All rights reserved. 6

It is therefore recommended to use Correlator3D™ on a PC that meets or exceeds the

© 2019 SimActive Inc. All rights reserved. 7

Data Type Supported Formats File Extensions

4.1 Sensor Types

© 2019 SimActive Inc. All rights reserved. 8

UAV/Medium/Large Format Satellite ADS VMA3

Project Creation Project Creation Project Creation Project Creation

Figure 2: Correlator3D™ workflow.

© 2019 SimActive Inc. All rights reserved. 9

4.2 Input Image Formats

Image Type Data Format

4.3 DEM Format

4.4 Orthophoto Format

­ 8, 16 or 32 bits per band

© 2019 SimActive Inc. All rights reserved. 10

Figure 3: Vector elevation dialog.

4.6 Coordinate Systems

© 2019 SimActive Inc. All rights reserved. 11

Figure 4: Projection dialog.

Proj.4 values representing a geographic coordinate system (latitude/longitude) are not

© 2019 SimActive Inc. All rights reserved. 12

Figure 5: Correlator3D™ main window.

5.1 Multilingual User Interface

5.2 Project Tree

© 2019 SimActive Inc. All rights reserved. 13

© 2019 SimActive Inc. All rights reserved. 14

Figure 6: Image window.

Please refer to section 7.1 for details.

© 2019 SimActive Inc. All rights reserved. 15

DEMs can be saved by clicking the contextual button in the tree.

5.2.6 Point Clouds

© 2019 SimActive Inc. All rights reserved. 16

© 2019 SimActive Inc. All rights reserved. 17

5.2.11 Reference Orthos

5.2.13 Online Data

5.4 Zooming and panning

© 2019 SimActive Inc. All rights reserved. 18

5.5 Image Band Setup

Automated Workflow: automatic execution of Correlator3D™ workflow

8, 16 or 32 bits per band