Arun Patel Paper Egyptian
Arun Patel Paper Egyptian
Arun Patel Paper Egyptian
H O S T E D BY
RESEARCH PAPER
KEYWORDS
SRTM;
ASTER;
Cartosat-1;
Interpolation techniques;
DGPS;
GIS
Abstract In Open sources DEMs such as SRTM, ASTER and Cartosat-1, various factors affecting the accuracy of satellite based DEM such as errors during data collection, systematic errors and
unknown errors that are geographically dependent on terrain conditions cannot be avoided. For
these reasons it is very necessary to check and compare the performances and validation of the
above mentioned different satellite based DEMs. Accuracy assessment of these DEM has been done
using DGPS points. For these points proper interpolation of the surface was developed using different interpolation techniques. For the generation of the surface the first step was converting the
satellite based DEMs height into linear interpolation contour maps of 1 m interval. Then came
selecting random sample points on the contour line and generating the interpolated surface using
different interpolation techniques such as IDW, GPI, RBF, OK and UK, LPI, TR and BI, which
are commonly used in geomorphology research. This interpolated surface helps in proper representation of the terrain and was checked under different terrain surfaces. For validation of DGPS
points the height was taken for ground control points and standard statistical tests such as ME
and RMSE were applied. From above investigation, it is reveals that above mention DEMs which
are used for study. Cartosat-1 (30 m) data product is better than SRTM (90 m) and ASTER (30 m)
because it had produced low RMSE of 3.49 m without applying the interpolation method. Investigation also reveals after applying the interpolation techniques on this data error can be reduced. In
the case of Cartosat-1 and SRTM, low RMSE and ME were produced by the BI method, where
Cartosat-1 DEM had an RMSE of 3.36 m with ME of 2.74 m, respectively. But in this case,
RMSE and ME of SRTM is 2.73 m and 0.36 m, respectively. BI is designed for image processing
and can be used for imagery were a maximum height variation in satellite DEM and terrain height is
minimum. But in the case of ASTER DEM, the GPI method with a high polynomial order of 9 had
Abbreviations: SRTM, Shuttle radar topographic mission; ASTER, Advanced Space Borne Thermal Emission and Reflection Radiometer; DGPS,
Differential Global Positioning System; IDW, Inverse Distance weight; GPI, Global Polynomial Interpolation; RBF, Radial Basis Function; OK
and UK, Ordinary and Universal Kriging; LPI, Local Polynomial Interpolation; TR, Topo to raster; BI, Bilinear interpolation; ME, Mean Error;
RMSE, root mean square error; DEM, digital elevation model
* Corresponding author. Tel.: +91 9407144228.
E-mail address: arunpatel123@gmail.com (A. Patel).
Peer review under responsibility of National Authority for Remote Sensing and Space Sciences.
http://dx.doi.org/10.1016/j.ejrs.2015.12.004
1110-9823 2015 National Authority for Remote Sensing and Space Sciences. Production and hosting by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
A. Patel et al.
produced a low RMSE of 4.99 m. The GPI method can be applied where maximum height variation
in satellite DEM and in terrain is more.
2015 National Authority for Remote Sensing and Space Sciences. Production and hosting by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-ncnd/4.0/).
1. Introduction
DEM is one of the most popular data models used for the purpose of terrain modeling. It is a grid based matrix structure,
which records topological relations between data points
implicitly. Since this data structure reflects the storage structure of digital computers (i.e. A grid can be stored as a two
dimensional array of elevations), the handling of elevation
matrices is simple, and many terrain analysis algorithms based
on this structure tend to be relatively straight forward. DEM is
an array representation of squared cells (pixels) with an elevation value associated to each pixel (Manuel, 2004). DEMs had
a wide range of applications in topography, geomorphology,
vegetation cover studies, tsunami assessment, and urban studies. There are various ways of obtaining DEMs either by contour lines, topographic maps, field surveying using auto level,
total station and GPS, Photogrammetry techniques, radar
interferometry, and laser altimetry (Manuel, 2004). Satellite
based DEM such as SRTM, ASTER and Cartosat-1 are freely
available and widely available.
The SRTM provides the most complete, highest resolution
available DEM of the Earth. It is based on the principle of
interferometric SAR or InSAR, which uses phase-difference
measurements derived from two radar images acquired with
a very small base to height ratio (typically 0.0002) to measure
topography (SRTM project). In quantitative terms, the cartographic products derived from the SRTM data are sampled
over a grid of 1 arc-second 1 arc-second (approximately
30 m 30 m). The SRTM global data for the rest of the World
other than the USA is available at 3 arcs second (90 m). The
product consists of seamless raster data, which is provided
according to a user specified area of coverage. The SRTM finished data meet the absolute horizontal and vertical accuracies
of 20 m (circular error at 90% confidence) and 16 m (linear
error at 90% confidence) respectively, as specified for the mission. The vertical accuracy is significantly better than the 16 m.
It is closer to 10 m (Rabus et al., 2002; Sun et al., 2003). Its
application is the concern of various studies which were conducted on topography (Falorni et al., 2005; Koch and
Lohmann, 2000), geomorphology (Guth, 2003; Stock et al.,
2002), vegetation cover studies (Kellndorfer et al., 2004), tsunami assessment (Blumberg et al., 2005), and urban studies
(Gamba et al., 2002). SRTM data verification was performed
using various altimetry data (Helm et al., 2002; Sun et al.,
2003) and digital elevation models (Muller, 2005; Jarvis
et al., 2004;Smith and Sandwell, 2003).
The ASTER DEM product is generated using bands 3N
(nadir-viewing) and 3B (backward-viewing) of an ASTER
Level-I A image acquired by the visible near infrared (VNIR)
sensor. The VNIR subsystem includes two independent telescope assemblies that facilitate the generation of stereoscopic
data. The band-3 stereo pair is acquired in the spectral range
of 0.78 and 0.86 lm with a base-to-height ratio of 0.6 and an
intersection angle of about 27.7. There is a time lag of approx-
9
2. Study area and data resources
The present paper is a case study of Maulana Azad National
Institute of Technology Bhopal (MANITB), in the capital city
of state of Madhya Pradesh, India. The topology of the Bhopal city is highly uneven and it has small hills within its boundaries. The geographic extent of the study area is 231130.44
to 232739.96 N latitude and 772632.86 to 772759 E
longitude with an average elevation of 523 m. Consider
DEM such as SRTM, ASTER and Cartosat-1 and there subset
DEMs are shown in Fig. 1. Horizontal and vertical datums of
all DEM are shown Table 1. where the world geodetic system
1984 (WGS84) is represented by the shape of the ellipsoid and
was calculated based on the hypothetical equipotential gravitational surface of the earth. But the vertical datum is referred to
mean sea level (MSL) as an orthometric height which is determined by the earth gravity model (EGM96) as a geoid model.
A significant difference exists between this mathematical ellipsoid model and the geoid model. The most mathematically
sophisticated geoid can only approximate the real shape of
the earth as shown in Fig. 2. If this ellipsoid vertical datum
is used, height above the ellipsoid will not be the same as
MSL and direct elevation readings for most locations will be
embarrassingly off. The surface of global undulations was calculated based on altimetric observations and very precise (up
to two centimeters) measurements taken from the TOPEX/
POSEIDON satellite. These data are represented in the
EGM96, which is also referred to as the spherical harmonic
model of the earths gravitational potential (Witold, 2003).
But in the case of DGPS default vertical datum is WGS 84
and the height computed relative to this (Kaplan and
Hegarty, 2006). But the elevation of a point on the earths surface computed from MSL can vary from GPS derived elevation because of the WGS84 ellipsoid and EGM96. The
Geoid surface is an equipotential or constant geopotential surface which corresponds to MSL (local Datum). The geoid
height/geoid undulation (N) the difference in height between
geoid (h) and ellipsoid (H) at a point is shown in Fig. 3, and
represented in Eq. (1)
hHN
As per specifications of Magellan Promark-3 single frequency DGPS system uses the Stop-n-go method with a horizontal accuracy of 0.012 m + 2.5 ppm and a vertical accuracy
of 0.015 m + 2.5 ppm and data are processed in GNSS solution after post-processing.
3. Methodology
Commonly used interpolation approaches have been evaluated
from satellite based DEM with reference to the study area and
the adopted methodology is summarized in Fig. 4.
In this SRTM DEM, ASTER DEM and Cartosat-1 DEM
were downloaded from the above mentioned website in
Table 2, subsetting area of interest (AOI) from DEMs as
shown in Fig. 1.
Following two steps involved in drawing the automated
contour on the raster DEM in ArcGIS Spatial analyst tool.
10
A. Patel et al.
Figure 1
Table 1
Study area boundary and DGPS point superimposed on satellite DEM (subset DEMs of SRTM, ASTER and Cartosat-1).
Data
Horizontal data
Vertical data
SRTM DEM
ASTER DEM
Cartosat DEM
WGS84
WGS84
WGS84
EGM96
EGM96
WGS84
Figure 2
11
n
1X
ME
Z Z
n i1
s
n
1X
Z Z2
RMSE
n i1
where
Z* = observed values of the height
Z = modeled values of the height
R
Rassterr to Coontoour linne uusinng
Sppatiial Annalyyst T
Toool inn A
ArcG
GIS
S
100
S
Sateellitte B
Baseed D
DEM
M
S
SRT
TM,,AS
STE
ER & C
Carrtossat1
DG
GPS
S suurvvey andd
preeprroceessiing of GC
CPs
F
Featturee too veerticces point toool uusinng
Dataa M
D
Manaagemeent ttooll inn ArrcG
GIS110
Geenerratiion of thee
Coontoour Maaps off 1m
meteer
inttervval
A
Anaalyzze, Coompparee
aandd
D
Dissplaay thhe R
Ressultts
Geeneeratiion of Raandoom
Poointt sam
mple oon tthe conntouur
linne
DE
EM
M Prrepaartiion usiing IDW,
GP
PI, RB
BF, Kriiginng aandd Toopoo to
Raasteer
M
Maagelllann Prrom
markk-3
ssinggle freequeenccy
D
DG
GPS system
m. andd
G
GN
NSS Sooluttionn for
pprocesssinng
Figure 4
Methodology.
12
A. Patel et al.
Table 2
S.
No
Image used
Resolution
(arc sec)
Satellite
Download
Date of
Procurement
SRTM
DEM
ASTER
DEM
Cartosat-1
DEM
3 (90 m)
Shuttle
Radar
ASTER
GDEM
Cartosat-1
Feb 2005
http://earthexplorer.usgs.gov/
17/10/2011
http://bhuvan-noeda.nrsc.gov.in/download/download/download.
php?c=s&s=C1&p=cdv2
20/08/2011
2
3
1 (30 m)
1 (30 m)
Specifying the parameters such as base contour, Z-factor & output location
Satellite DEM
Figure 5
Figure 6
Satellite DEM with contour lines with 1 m interval and selected vertices to point on the contour line.
Figure 7
Figure 8
13
is shown in Fig. 8. There is a high variation of ASTER as compared to SRTM and Cartosat-1 DEMs but SRTM and
Cartosat-1 approximately have the same variation of height
as shown in Fig. 8. Further comparison of variation of height
is done by considering statistical parameters of satellite DEMs
14
A. Patel et al.
ME
RMSE
SRTM
ASTER
Cartosat-1
0.61
3.72
0.45
6.03
2.71
3.49
SRTM
ASTER
Cartosat-1
Table 4
Table 6
DGPS.
DHmin
DHmean
DStd:dev
2
15
6.14
5.64
14
6.74
3.01
2.02
6.34
0.04
2.19
0.5
Where
DHm ax = maximum height in satellite DEM maximum height in
DGPS.
DHmin = minimum height in satellite DEM minimum height in
DGPS.
DHmean = mean height in satellite DEM mean height in DGPS
DStd:dev = Standard deviation in satellite DEM standard deviation in DGPS
IDW15
GPI
MRBF
LPI
OK
UK
TR
BI
Mean error
SRTM
Aster
Cartosat-1
1.56
0.29
3.42
1.57
0.27
3.45
1.21
0.81
3.63
1.41
0.33
3.41
1.40
0.36
3.35
1.34
0.32
3.33
1.35
0.22
3.36
1.40
0.28
3.40
0.36
0.28
2.74
RMSE
SRTM
Aster
Cartosat-1
3.85
6.01
4.20
3.87
6.01
4.21
4.56
4.99
4.47
3.73
6.17
4.18
3.67
5.93
4.15
3.65
6.14
4.11
3.64
6.11
4.14
3.67
6.13
4.16
2.73
6.12
3.36
Table 5
Height
Maximum
Minimum
Mean
Std. dev
Cartosat-1 DEM
(Ellipsoidal height)
DGPS
Geoided model
EGM96
Ellipsoidal
height
539
517
530.90
4.25
552
503
529.01
6.42
486
461
475.02
3.79
537
511.362
526.99
4.29
479.88
454.262
468.68
References
Aguilar, F.J., Aguera, F., Aguilar, M.A., Carvajal, F., 2005. Effect of
terrain morphology, sampling density and interpolation methods
on grid DEM accuracy. Photogramm. Eng. Remote Sens. 71, 805
816.
Arun, P.V., 2013. A comparative analysis of different DEM
interpolation methods. Egypt. J. Remote Sens. Space Sci. 16,
133139.
Blumberg, D., Bach, D., Weissel, J., Gorokhovich, Y., Small, C.,
Balk, D., 2005. The 2004 Sumatra Tsunami Event: Contribution of
SRTM Data to the Analysis of Devastation. The Shuttle Radar
Topography Mission-Data Validation and Applications, Workshop, Reston, Virginia.
Brus, D.J., Gruijiter, J.J., Marsman, B.A., Visschers, R., Bregt, A.K.,
Breeuwsma, A., 1996. The performance of spatial interpolation
methods and choropleth maps to estimate properties at points: a
soil survey case study. Envirometrics 7, 116.
Burrough, P.A., McDonnell, R.A., 1998. Principles of Geographical
Information Systems. Oxford University Press, New York, pp.
333335.
Carrara, A., Bitelli, G., Carla, R., 1997. Comparison of techniques
for generating digital terrain models from contour lines. International Journal of Geographical Information Science 11, 451473.
Chang, K.T., 2008. Introduction to Geographic Information System,
fourth ed. Tata Mc Graw-Hill.
Chaplot, V., Darboux, F., Bourennane, H., Leguedois, S., Sivera, N.,
Phachomphon, K., 2006. Accuracy of interpolation techniques for
the derivation of digital elevation models in relation to landform
types and data density. Geomorphology 77, 126141, Elsevier.
Creutin, J.D., Obled, C., 1982. Objective analyses and mapping
techniques of rainfall fields: an objective comparison. Water
Resour. Res. 18, 413431.
Declercq, F.A.N., 1996. Interpolation Methods for Scattered Sample
data: accuracy, spatial patterns processing time. Cartograp.
Geograp. Inform. Sys. 23, 128144.
15
Falorni, G., Teles, V., Vivoni, E.R., Bras, R.L., Amaratunga, K., 2005.
Analysis and characterization of the vertical accuracy of digital
elevation models from the shuttle radar topography mission. J.
Geophys. Res. 110, F02005.
Gallichand, J., Marcotte, D., 1993. Mapping clay content for
subsurface drainage in the Nile delta. Geoderma 58, 165179.
Gamba, P., DellAcqua, F., Houshmand, B., 2002. SRTM data
characterization in urban areas. In: International Society for
Photogrammetry and Remote Sensing Commission III, Symposium 2002, September 913, 2002, Graz, Austria 20042008.
Guth, P., 2003. Geomorphology of DEMs: quality assessment and
scale effects. Paper No. 1752. In: Proceedings of GSA, Seattle
Annual Meeting, November 25, 2003.
Helm, A., Braun, A., Eickschen, S., Schune, T., 2002. Calibration of
the shuttle radar topography mission X-SAR instrument using a
synthetic altimetry data model. Can. J. Remote Sens. 28 (4), 573
580.
Holmes, K.W., Chadwick, O.A., Kyriankidis, P.C., 2000. Error in
USGS 30-meter digital elevation model and its impact on terrain
modelling. J. Hydrol. 233, 154173.
Jones, T.A., Hamilton, D.E., Johnson, C.R., 1986. Contouring
Geologic Surface with the Computer (Van Nostrand Reinhold
Catalysis Series), XVI. Springer, p. 314.
Jarvis, A., Rubiano, J., Nelson, A., Farrow, A., Mulligan, M., 2004.
Practical use of SRTM data in the tropicscomparisons with
digital elevation models generated from cartographic data. Working Document, vol. 198, Centro Internacional de Agricultura
Tropical (CIAT), pp. 32.
Johnston, K., Ver-Hoef, J.M., Krivoruchko, K., Lucas, N., 2001.
Using ArcGIS Geostatistical Analyst. ESRI, Redlands, p. 300.
Kaplan, D., Hegarty, C.J., 2006. Understanding GPS: Principles and
Applications. Artech House, Boston, London, p. 32.
Kellndorfer, J., Walker, W., Pierce, L., Dobson, C., Fites, J.A.,
Hunsaker, C., et al, 2004. Vegetation height estimation from shuttle
radar topography mission and national elevation datasets. Remote
Sens. Environ. 93, 339358.
Koch, A., Lohmann, P., 2000. Quality assessment and validation of
digital surface models derived from the shuttle radar topography
mission (SRTM). In: Proceedings, IAPRS, vol. XXXIII, Amsterdam, 2000.
Laslett, G.M., McBratney, A.B., 2002. Further comparison of spatial
methods for predicting soil pH. Soil Sci. Am. J. 54, 15531558.
Manuel, P., 2004. Influence of DEM interpolation methods in
Drainage Analysis. GIS Hydro 04, Texas, USA.
Mitas, L., Mitasova, H., 1999. Spatial interpolation. In: Longley, P.,
Goodchild, K.F., Maguire, D.J., Rhind, D.W. (Eds.), Geographical Information System: Principles, Techniques, Management and
Applications. Wiley, New Work, pp. 481492.
Joshi, S., Mukherjee, P.K., Ghosh, S., Garg, R.D., Anirban, M., 2013.
Evaluation of vertical accuracy of open source digital elevation
model (DEM). Int. J. Appl. Earth Obs. Geoinf. 21, 205217.
Mukherjee, S., Garg, R.D., Mukherjee, S., 2011. Effect of systematic
error and its derived attributes: a case study on Dehradun area
using Cartosat-1 stereo data. Indian J. Landscape Sys. Ecol. Stud.
34 (1), 4558.
Muller, J.P., 2005. Quantitative Assessment of C-band and X-band
SRTM Datasets Over the CEOS-WGCV-TMSG Test Sites and
Intercomparison of Cband DEM with the OS PANORAMA
DTM. The Shuttle Radar Topography Mission-Data Validation
and Applications, Workshop, Reston, Virginia.
Rabus, B., Michael, E., Achim, R., Richard, B., 2002. The shuttle
radar topography mission a new class of digital elevation models
acquired by spaceborne radar. ISPRS 57 (4), 241262.
Robeson, S.M., 1997. Spherical methods for spatial interpolation:
review and evaluation. Cartograp. Geograp. Inform. Sys. 24, 93
123.
16
Rodgriguez, E., Morris, C., Belz, J., 2006. A global assessment of
SRTM performance. Photogramm. Eng. Remote Sens. 72, 249
260.
Smith, B., Sandwell, D., 2003. Accuracy and resolution of shuttle
radar topography mission data. Geophys. Res. Lett. 30 (9), 1467.
Stock, J. D., Bellugi, D., Dietrich, W. E., Allen, D., 2002. Comparison
of SRTM topography to USGS and high resolution laser altimetry
topography in steep landscapes: Case studies from Oregon and
California. In: Eos Trans. AGU, Fall Meet. Suppl., Abstract
H21G09, 2002, vol. 83(47).
Sun, G., Ranson, K.J., Kharuk, V.I., Kovacs, K., 2003. Validation of
surface height from shuttle radar topography mission using shuttle
laser altimeter. Remote Sens. Environ. 88 (4), 401411.
Watson, D., 1992. Contouring: A Guide to the Analysis and Display of
Spatial Data. Pergamon Press, London, pp. 120123.
A. Patel et al.
Weber, D., Englund, E., 1992. Evaluation and comparison of spatial
interpolators. Math. Geol. 24, 381391.
Weber, D., Englund, E., 1994. Evaluation and comparison of spatial
interpolators II. Math. Geol. 26, 589603.
Wilson, J.P., Gallant, J.C., 2000. Digital Terrain Analysis. Principles
and Applications. John Wiley and Sons, New York, NY, 127.
Wise, S., 2011. Cross-validation as a means of Investigating DEM
Interpolation Error. Comput. Geosci. 37, 978991.
Witold F., Mean Sea Level, GPS, and the Geoid, Esri Applications
Prototype Lab, JulySeptember 2003, available at http://www.
esri.com/news/arcuser/0703/geoid1of3.html.
Zimmerman, D., Pavlik, C., Ruggles, A., Armstrong, M., 1999. An
experimental comparison of ordinary and universal kriging and
inverse distance weighting. Math. Geol. 31, 375390.