See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/323178096

PRE-PROCESSING AND POST-PRECESSING LIDAR DATA

Article · May 2015

CITATIONS READS

0 553

2 authors:

Adrian Ghimbasan Iosif Vorovencii

Universitatea Transilvania Brasov Universitatea Transilvania Brasov
4 PUBLICATIONS   1 CITATION    50 PUBLICATIONS   205 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Situation of cadastre in Romania View project

All content following this page was uploaded by Adrian Ghimbasan on 14 February 2018.

The user has requested enhancement of the downloaded file.

STUDIA UNIVERSITATIS “VASILE GOLDIŞ” ARAD May 2015 Vol.10 Iss.1 pp.7-13
SERIA ŞTIINŢE INGINEREŞTI ŞI AGRO-TURISM http://studia.facultateadeinginerie.ro
“Vasile Goldiş” University Press
PRE-PROCESSING AND POST-PRECESSING LIDAR DATA
Adrian Cosmin Ghimbasan, Cornel Cristian Teresneu, Iosif Vorovencii

“Vasile Goldiş” Western University of Arad

________________________________________________________________________________

PRE-PROCESSING AND POST-PRECESSING LIDAR DATA

Adrian Cosmin Ghimbaşan1

Cornel Cristian Tereşneu1
Iosif Vorovencii1
1
Forest Management Planning and Terrestrial Measurements Department,
Faculty of Silviculture and Forest Engineering, Transilvania University, Braşov
no.1 Şirul Beethoven St., 500123, Braşov, Romania

Abstract
The present paper is aimed at establishing whether and to what extent LiDAR technology proves useful and
what precision can be obtained with this type of data. The methodology was applied on a series of LiDAR
surveys from the area of United States of America, State of California, area of Los Osos, that were closely
analyzed in the pre-processing stage from the viewpoint of their accuracy by means of suggestive statistical
indicators.
In the post-processing stage, it was used two software packages –QTReader 800 and Global Mapper -
capable of processing this kind of LiDAR surveys in order to obtain a highly precise digital terrain model
(DTM) and other products.

Keywords: LiDAR technology, accuracy, software, DTM.

The base concept is to determine the distance from

INTRODUCTION the airborne platform to the monitored Earth’s surface,
the off-nadir scan angle, the aircraft’s position in its
The surveying of all types of land use actually reference system and, subsequent to the pre-processing
allows us to get to know our territorial fund, this being stage, the coordinates of the points reflected off the
achieved in the past through field works that assumed ground or by other objects that the laser pulses
time-consuming and tiresome determinations, travels and encounter, the surveyed points being known as point
descriptions, especially in steep land conditions. cloud (Figure 1), [4].
Lately, was attempted and mostly achieved to use
modern means to fulfill this objective, means that are
reflected in modern technologies, such as: GPS-based
positioning or deriving qualitative geo-topographic data
through airborne and satellite remote sensing.
Nowadays, these techniques and technologies are
covered by the geomatics, to which it can be attached
several definitions and whose objective is to get to know
a part of the Earth’s surface from two viewpoints: its
position in a reference system and its
mapping/translation into plans [2].
The laser scanning LiDAR technology (Light
Detection and Ranging) falls within this science and
consists of a laser scanning technique combined with a
highly accurate GPS-IMU (Inertial Measurement Unit) Fig. 1. Operating Mode in Airborne Laser Scanning
(Birjaru, 2011)
system mounted on an airborne platform [1].

* Author to whom all correspondence should de addressed: e-mail: bojanicu@yahoo.com

STUDIA UNIVERSITATIS “VASILE GOLDIŞ” ARAD May 2015 Vol.10 Iss.1 pp.7-13
SERIA ŞTIINŢE INGINEREŞTI ŞI AGRO-TURISM http://studia.facultateadeinginerie.ro
“Vasile Goldiş” University Press
PRE-PROCESSING AND POST-PRECESSING LIDAR DATA
Adrian Cosmin Ghimbasan, Cornel Cristian Teresneu, Iosif Vorovencii

“Vasile Goldiş” Western University of Arad

________________________________________________________________________________
obtained on the occasion of the flight and data
This paper’s objective is based/centred on the processing. We equally made use of indirect
question whether and to what extent we should use observations derived from additional work stages
LiDAR technology, and the goals to be fulfilled are: dependent on collateral factors.
 getting acquainted with the modern 3D laser The methods used during the present study are: the
scanning LiDAR technology both theoretically, direct comparison method and the statistical method.
as well as from the viewpoint of its general As a working methodology, terrestrial and aerial
uses; surveys were performed in Real Time Kinematic (RTK)
 establish what precision can be obtained with mode at times when the Point Dilution Of Precision (the
this type of data; PDOP) was lower or equal to three and the number of
 obtaining a highly accurate digital elevation active satellites on the Earth’s sky defining the satellite
model (DEM) and other specific topographic constellation at the moment of data collection was of
products; minimum six. The terrestrial RTK surveys are used to
 mentioning the fields that could integrate assess the accuracy of the collected LiDAR data, while
LiDAR technology or where this technology the aerial RTK surveys serve to determine the aircraft’s
could serve some indirect purposes. position during the release and receipt of the laser pulses
(Figure 3).
MATERIALS AND METHODS Static surveys were also performed close to the
flight area in order to link the terrestrial and aerial
The data used in the present study were received segments and be able to determine the coordinates of the
from the company Watershed Sciences Inc (WSI) that surveyed point cloud in the pre-processing stage.
collected them for „PACIFIC GAZ AND ELECTRIC
COMPANY”, on request by the latter. The data were
collected in the United States of America, State of
California, area of Los Osos, because such data being
difficult to obtain in Romania.
The LiDAR data were collected between March 17th
and May 31st,2011 , and the area of interest was of 675
km2 (166,696 acres), subsequently extended by including
a buffer area of 100m to ensure full coverage and
adequate point density close to the one required (8
points/sq.m.). The final area was of 688 km2 (169,968
acres) (Figure 2).

Fig. 3. Layout of the Points surveyed with the Static

Method and the RTK Method

RESULTS AND DISCUSSIONS

ACCURACY ASSESSMENT

The relative accuracy refers to the internal

consistency of the dataset, more precisely to the ability to
place a laser point in the same location over multiple
flight lines, GPS conditions or aircraft altitude. This
indicator of LiDAR surveys can be understood as the
Fig. 2. Location of the Study Area divergence between points from different flight lines
over the overlapping area [7]. The system is deemed well
The methods that were applied to perform the calibrated when the divergence between the ground
analysis and present the results are those commonly used levels of the same point from the overlapping area is less
and especially materialized in the methodology used for than 10 cm (Figure 4).
the majority of the works. At the core of this
methodology lies the direct observations performed and
8
STUDIA UNIVERSITATIS “VASILE GOLDIŞ” ARAD May 2015 Vol.10 Iss.1 pp.7-13
SERIA ŞTIINŢE INGINEREŞTI ŞI AGRO-TURISM http://studia.facultateadeinginerie.ro
“Vasile Goldiş” University Press
PRE-PROCESSING AND POST-PRECESSING LIDAR DATA
Adrian Cosmin Ghimbasan, Cornel Cristian Teresneu, Iosif Vorovencii

“Vasile Goldiş” Western University of Arad

________________________________________________________________________________
Table 1 Absolute Accuracy - Divergence between Laser
Points and RTK Surveyed Points
Number of RTK points = 2482
Root Mean Square Minimum z =
Error (RMSE) = 0,028
m -0,115 m

Standard deviation Maximum z =

0,074 m
1 sigma ( ) = 0,028 m
Average z =
1,96 sigma ( ) = 0,055
-0,005 m
m
Fig. 4. Distribution of Relative Accuracies in the Study
Area of Los Osos per Flight Line, non Slope-adjusted
The initial dataset with a point density higher than 8
points per square meter was then filtered to remove false
Another indicator of the LiDAR surveys is absolute
or inaccurate points. Moreover, some types of surface
accuracy defined as a function dependent on two other
area (for instance: dense vegetation, bare land, water,
indicators - laser noise and relative accuracy- [7].
steep slopes) may return fewer pulses (with a lower
Prior to assessing absolute accuracy, we used a
density) than those issued initially (with an implicitly
number of noise filters and calibration procedures to
higher density).
minimize the effects of these contributions (laser noise
The land cover classification was achieved both
and relative accuracy). 2482 RTK points were collected
through automatic and manual modelling, while the
in the study area of Los Osos on hard-packed surfaces
supervised classification revealed that automatic
(that meet location requirements) distributed over
modelling was not entirely accurate since the derived
multiple flight lines. To assess absolute accuracy, the
model was not compliant with the ground truth at the
coordinates of the 2482 RTK points were compared to
time of the data collection.
those of the nearest laser points on the ground (Figure 5).
Following the processing of the collected data, we
The vertical accuracy of the LiDAR data is defined as
concluded that the purpose of this study was achieved
standard deviation ( resulting from the divergence
with a resolution of 8.28 points/m2 and a vertical
between the coordinates of the LiDAR points and those accuracy of 2.8 cm (Figure 6).
of the RTK surveyed points (Table 1).
Another indicator was used to draw statistical
models of the vertical accuracy, namely the Root Mean
Square Error – RMSE [9].

Fig. 5. Absolute Accuracy Histogram

This RMSE can be computed only for areas void of

vegetation, more precisely for clear, potentially topsoil Fig. 6. Distribution Map for Unclassified Points (up) and
stripped areas where points were surveyed with the RTK Distribution Map for Ground Classified Points (down)
method.
9
STUDIA UNIVERSITATIS “VASILE GOLDIŞ” ARAD May 2015 Vol.10 Iss.1 pp.7-13
SERIA ŞTIINŢE INGINEREŞTI ŞI AGRO-TURISM http://studia.facultateadeinginerie.ro
“Vasile Goldiş” University Press
PRE-PROCESSING AND POST-PRECESSING LIDAR DATA
Adrian Cosmin Ghimbasan, Cornel Cristian Teresneu, Iosif Vorovencii

“Vasile Goldiş” Western University of Arad

________________________________________________________________________________
DATA POST-PROCESSING

The software packages capable of processing this

type of data can be classified according to the
possibilities they offer into two categories [5]:
1. software for the visualization of LiDAR
datasets (viewers);
2. software that allows users both to visualize
LiDAR data and to derive specific topographic
products (digital elevation models, profiles and
other topo-cadastral pieces).
We should mention that for the purpose of the
present study we did not consider the entire area which
served for the collection of the LiDAR data since the
number of points surveyed on this area is of
approximately 5.4 billion points, and current software
can only process approximately 100 million points. For
this reason, we chose an area that would serve the
targeted results, namely to obtain a highly accurate Fig. 7. Inserted Model
digital elevation model based on a number of 98,437,917
points surveyed on a 7.939 km2 area.

QTREADER SOTWARE Fig. 8. Coordinate Display in QTReader 800

After selecting the file that we want to visualize, the The software also allows drawing a grid for the
software loads the model that can be visualized in a few inserted model if needed by accessing the sequence
seconds or in several dozen of minutes, according to its Analysis-Generate Grid Lines and selecting the desired
size. The time needed to load the model is directly settings in the prompt window. The grid complying with
dependent on the number of points that make up above the selected settings shall be displayed after clicking the
mentioned model. In our case, the time needed to open Ok button (Figure 9).
the model was of 14.5 min for a number of 98,437,917
points (the opened file is 2.8 GB).
This software allows users to rotate the model, zoom
it to a particular constituent point, or change colours
according to point ground level, as well as other
operations for the detailed visualization of the model
(Figure 7).
Mention must be made on the fact that this software
allows its users to read the coordinates of targeted points
in the used projection system (Projection system NAD83
-North American Datum 1983/UTM zone 10N-, defined
by NAD83 coordinate system and the GRS 80 ellipsoid),
(Figure 8).

Fig. 9. Grid Overlay on Top of the Existing Model

10
STUDIA UNIVERSITATIS “VASILE GOLDIŞ” ARAD May 2015 Vol.10 Iss.1 pp.7-13
SERIA ŞTIINŢE INGINEREŞTI ŞI AGRO-TURISM http://studia.facultateadeinginerie.ro
“Vasile Goldiş” University Press
PRE-PROCESSING AND POST-PRECESSING LIDAR DATA
Adrian Cosmin Ghimbasan, Cornel Cristian Teresneu, Iosif Vorovencii

“Vasile Goldiş” Western University of Arad

________________________________________________________________________________
GLOBAL MAPPER VERSION 13 SOFTWARE

Global Mapper is more than a viewer that allows

visualizing tri-dimensional models. It is a software which
allows users to perform several operations, such as:
calculation of the surface area, river basin delineation,
spectral analyses, generation of DEMs, contour lines or
profiles for the study areas, as well as a series of other
operations that can serve different purposes.
The file containing the data for the generation of the
highly accurate digital elevation model can be opened
with the sequence File-Open Data File(s) or by left-
clicking the Open Your Own Data Files button which
prompts when the software is opened. After selecting the
file and clicking the Open button, a new window
prompts (LiDAR Load Options) with two options:
generating a point cloud or directly creating a digital
elevation model. Of these two options, we chose the first
one. The software offers the possibility to select the
points in relation to their characteristics that are closely
connected to the pre-processing stage when the points
were classified according to their sensor return time. For
the purpose of the present study, we inserted the points
that were classified as soil by selecting this option
(Figure 10, Figure11). Fig. 10. Lidar Load Options Window
The digital elevation model to be derived based on
this cloud point was obtained by clicking the Open
Control Center button. This opens the window Overlay
Control Center which contains the inserted point layer
and allows the user to read their number according to the
selected inserting options (Figure 12).
It is important to keep in mind that the model
maintains the accuracy of the input data for which reason
one can state that the resulting digital model is highly
accurate being based on highly accurate input data both
altimetrically and planimetrically [8].

Fig. 11. Tri-dimensional View of the Point Cloud

Fig. 12. Visualization of the Digital Elevation Model

11
STUDIA UNIVERSITATIS “VASILE GOLDIŞ” ARAD May 2015 Vol.10 Iss.1 pp.7-13
SERIA ŞTIINŢE INGINEREŞTI ŞI AGRO-TURISM http://studia.facultateadeinginerie.ro
“Vasile Goldiş” University Press
PRE-PROCESSING AND POST-PRECESSING LIDAR DATA
Adrian Cosmin Ghimbasan, Cornel Cristian Teresneu, Iosif Vorovencii

“Vasile Goldiş” Western University of Arad

________________________________________________________________________________
The digital elevation model allowed us to generate
the corresponding contour lines with the sequence File-
Generate Contours (Figure 13).

Fig. 15. Watershed Derived for the Study Area

CONCLUSIONS

Fig. 13. Contour Lines Derived from the Digital Elevation A series of general findings and future concerns can
Model be derived in relation to the performed activities and this
can be resumed as follows:
The software also offers the possibility of deriving 1. The 3D laser scanning LiDAR technology is a
profiles for the study areas based on the digital elevation component of the new science called geomatics
model, and this was achieved by clicking the 3D Path 2. LiDAR technology is an active remote sensing
Profile/Line of Sight Tool button on the upper toolbar system based on the release of a weak scattering
(Figure 14). laser pulse and the receipt by the sensor of the
radiation reflected by the scanned surface area;
the result is a ‘cloud’ consisting of thousands of
3D positioned points.
3. The main advantage of this system is the fact
that the data collection, storage and processing
process is fully automated due to the use of
specialized software which allows users to
obtain coordinates with a horizontal accuracy
(x, y) of less than 10 cm and a vertical accuracy
of 2.8 cm.
4. The ability of the laser pulse to pierce through
forest canopy all the way to the ground is the
essential characteristic of this system.
5. By using software packages capable of
processing this kind of data we managed to
obtain highly accurate digital model of the
Fig. 14. Longitudinal Profile of a Road from the Study ground truth, its high accuracy being due to the
Area input data.
A series of different other products can be obtained
Another useful function of this software is the based on the digital elevation model (contour line maps
possibility to generate the watershed. This function of the study area, profiles, watersheds, etc.), as well as a
meets the needs of other disciplines as well (river basin series of indirect information such as tree height or
management and planning, torrent study) since the building height which allows for predictions on areas
derived information can help identifying areas subject to potentially prone to floods, landslides or torrents.
risks of floods, land slides or torrents (Figure 15).
REFERENCES

Birjaru, C., Cercetări privind utilizarea tehnologiei

LiDAR în lucrările de silvicultură, Rezumat teză
doctorat, Braşov, 2011, 60 p;

12
 STUDIA UNIVERSITATIS “VASILE GOLDIŞ” ARAD May 2015 Vol.10 Iss.1 pp.7-13
SERIA ŞTIINŢE INGINEREŞTI ŞI AGRO-TURISM http://studia.facultateadeinginerie.ro
“Vasile Goldiş” University Press
PRE-PROCESSING AND POST-PRECESSING LIDAR DATA
Adrian Cosmin Ghimbasan, Cornel Cristian Teresneu, Iosif Vorovencii

“Vasile Goldiş” Western University of Arad

________________________________________________________________________________
Boş, N, Geomatica şi realizarea bazei cartografice a ***www.ifp.uni-stuttgard.de/
fondului forestier din România - Revista publications/2004cramer3063-DSS-final.pdf –
pădurilor, 126.6 (2011): pp. 27-36. Cramer M., Performance of Medium Format
Covasnianu, A., et al, Digital Terrain Model by airborne Digital Aerial Sensor Systems, Institute for
LIDAR technique: an essential tool for hydrologic Photogrammetry (ipf), University of Stuttgart;
risks assessment -Journal of Optoelectronics and ***www.isprs.org/congresses/instambul2004/comm2/pa
Advanced Materials 9.11 (2007): 3529; pers/119.pdf- Medvedev E., 2004, Towards the
Vorovencii, I.,– Fotogrammetrie, Editura MATRIX full automation of Laser scanning and aerial
ROM, 2010, 530 p.; photography data processing;
***www.cicade.com- CICADE, Aerial Photography, ***www.nibis.ni.schule.de/trianet/home.htm – DTM-
Photogrammetric Scanning, Aerotriangulation, Digital Terrain Model;
Orthophotos, GPS, Digital Photogrammetry, ***www.zone.ni.com/dzhp/app/ main – Image Filtering
Survey, LiDAR; Overview.

13

View publication stats