Updating the 1:50.000 topographic maps using ASTER and SRTM
DEM. The case of Athens, Greece.
Konstantinos G. Nikolakopoulos*a & Nektarios Chrysoulakisb
Remote Sensing Laboratory, Department of Geology & Geoenvironment, University of Athens,
a
: 1 Iroon Polytehniou Str. 15127 Melissia, Athens, Greece,
b
Foundation for Research and Technology - Hellas, Institute of Applied and Computational
Mathematics
b
:Vassilika Vouton, P.O. Box 1527, GR-71110, Heraklion, Crete, Greece,
a
ABSTRACT
The 1:50.000 topographic maps present a nominal horizontal accuracy of 20 meters and a nominal vertical accuracy of
10 meters with 90% confidence. The data were in most cases extracted with photogrammetric techniques from aerial
stereo-photographs during the 80’s. The usual update rate for these maps ranges from ten to twenty years.
The Advanced Spaceborn Thermal Emission and Reflection Radiometer (ASTER) offers along-track stereoscopic
viewing capability. Its viewing geometry is suitable for DEM generation even without the use of ground control points.
Recent studies have proved that in this case the vertical accuracy of DEM is about 20m with 95% confidence. The
horizontal geolocation accuracy appears to be limited by the spacecraft position accuracy which is considered to be
better than 50 m. Other studies have shown that the use of GCP’s resulted in a plannimetric accuracy of 15 m and in a
near pixel size vertical accuracy.
The Shuttle Radar Topography Mission (SRTM), used an Interferometric Synthetic Aperture Radar (IFSAR) instrument
to produce a near-global digital elevation map of the earth's land surface with 16 m absolute vertical height accuracy at
30 meter postings. An SRTM 3-arc-second product (90m resolution) is available for the entire world.
In this paper we examine the possibility of updating the 1:50.000 topographic maps using ASTER and SRTM DEMs.
The area of study is the broader area of Athens, Greece.
Presupposing, that the horizontal and vertical accuracy of the ASTER and SRTM DEM is similar to the relative
accuracies of the DEM from digitized contours, optical comparison of the DEMs and statistical analysis (difference,
correlation) can immediately prove if there is any need for update to the topographic maps.
A DEM from digitized contours from the 1:50.000 topographic maps was created and compared with ASTER and SRTM
derived DEMs. Almost three hundreds points of known elevation have been used to estimate the accuracy of these three
DEMs.
The resulted accuracy of the SRTM and ASTER derived DEMs was satisfactory, therefore they are considered as
suitable for updating 1:50.000 topographic maps.
Keywords: SRTM DEM, ASTER DEM, digitized contours, trigonometric points, filters, accuracy.
1. INTRODUCTION
The Digital Elevation Model (DEM) is a very important precondition for many applications as map generation, threedimensional GIS, environmental monitoring, geo-spatial analysis etc1,2. As a result the scientific efforts focalized to
create high accuracy DEM and to cover the entire planet. The DEM from satellite stereo-pair images seem to fulfill both
these two conditions.
Despite the continuous improvement of the spatial resolution of the satellite images and the possibility of global covering
of the entire planet that gave the necessary impulsion for the development of many new algorithms for the automatic
DEM extraction, the performance of which have been assessed and reported in the bibliography3,4,5,6, the 1:50.000
topographic maps remain the main source of elevation data.
The 1:50.000 topographic maps of the Hellenic Army Geographical Service present a nominal horizontal accuracy of 20
meters and a nominal vertical accuracy of 10 meters with 90% confidence. The data were in most cases extracted with
Remote Sensing for Environmental Monitoring, GIS Applications, and Geology VI,
edited by Manfred Ehlers, Ulrich Michel, Proc. of SPIE Vol. 6366,
636606, (2006) · 0277-786X/06/$15 · doi: 10.1117/12.689016
Proc. of SPIE Vol. 6366 636606-1
photogrammetric techniques from aerial stereo-photographs during the 80’s. The usual update rate for these maps ranges
from ten to twenty years. Thus, in many cases there is a need for updated and more recent elevation data.
In this paper we examine the possibility of updating the 1:50.000 topographic maps using ASTER and SRTM DEMs.
We compare the accuracies of three DEM’s constructed from different sources of data: A DEM created from digitized
contours from 1/50.000 scale topographic maps, an ASTER DEM from satellite stereo images and a SRTM DEM
produced from interferometric SAR images. The broader area of Athens, capital of Greece, was selected for the
comparison of the three DEM’s. The topography is quite complex. At the east the Hymettus Mountain, at the west the
Aigaleo Mountain and at the North the Pendeli Mountain determine the boundaries of the Athens basin and of the urban
centre (Figure 1). The elevation of the broader area ranges from 0 to more than 1050 meters.
2. ASTER DEM
The Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) is an advanced multispectral imager
that was launched on board NASA’s Terra spacecraft in December, 1999. Its viewing geometry is suitable for DEM
generation with horizontal spatial resolution of 15 meters and a near pixel size vertical accuracy. ASTER consists of
three separate instruments subsystems, each operating in a different spectral region, using separate optical system. The
Visible - Near Infrared system, which is used in DEM production, consists of two telescopes – one nadir looking with a
three band detector and the other backward looking (27.7° off-nadir) with a single band detector. The most important
specifications of the ASTER stereo subsystem that govern the DEM generation capabilities include: stereo geometry;
platform altitude of 705 km and base to height ratio of 0.67. The ASTER method gives a strong advantage in terms of
radiometric variations versus the multi-date stereo-data acquisition with across-track stereo, which can then compensate
for the weaker stereo geometry8. The viability of stereo correlation for parallax difference from digital stereoscopic data
has been described and evaluated in previous studies9,10. ASTER DEM data provided the first high-resolution near global
elevation source. Although this remarkable data set is extremely useful due to its relatively high resolution, it suffers
from some drawbacks, such as the lack of coverage in several areas due to the weather conditions during the stereoimagery acquisition.
According to the pre-launch specifications10 ASTER DEM could be more accurate from the existing DTM like
GTOPO30 NIMA DTED-1 and their accuracy will be ranged, proportionally to the accuracy of the GCPs, from 7 to 30
meters.
As the data became available several studies have been published concerning the accuracy of ASTER DEM. In one of
these studies11 the vertical accuracy of DEM data was precisely evaluated using high-accuracy GCPs. The results
consistently show that the standard deviation is about 10 m, and therefore the vertical accuracy of DEM is about 20m
with 95% confidence (2σ) without the use of GCPs. The horizontal geolocation accuracy appears to be limited by the
spacecraft position accuracy, which is considered to be better than 50 m. In addition, a slight increase in the accuracy
was observed for the absolute DEM data generated with referencing GCPs11. Other studies12,13 have proved that the
planimetric and vertical accuracies may be better. As a part of REALDEMS project, high spatial resolution ASTER
stereo imagery was analyzed to produce DEM for Heraklion and Sitia areas in Crete Island. Differentially corrected GPS
measurements were performed to provide GCPs for DEM correction and geo-location. The planimetric and elevation
accuracy of the produced DEMs was 15.0 m and 12.4 m, respectively12.13. It was also proved14 ASTER data have very
good accuracy that depends on the accuracy of the ground control points. In the case of Milos Island Greece, the
accuracy ranged between 4.3 and 32.7 meters14 and in the case of Lefkas Island the accuracy ranged between 10 and 24
meters15.
Another study has proved that the ASTER data should also prove suitable for topographic mapping of high relief areas at
scales of 1:50,000 to 1:100,000 with contour intervals of 40m or larger. As necessary, ASTER DEMs also can be used to
correct for relief displacements in images produced by other satellites and to provide the basis for orthoimage
development from both aircraft and satellite image data16.
An ASTER image acquired on October 10, 2001 was used for a DEM creation with a 15m-pixel size. The ASTER DEM
is presented in Figure 2.
3. SRTM DEM
The Shuttle Radar Topography Mission (SRTM) mission17,18 has been the first mission using space-borne single-pass
interferometric SAR. This mission was a partnership between NASA and the Department of Defense’s National Imagery
and Mapping Agency (NIMA). In addition, the German and Italian space agencies were contributing an experimental
Proc. of SPIE Vol. 6366 636606-2
high-resolution imaging radar system. Flown aboard the NASA Space Shuttle Endeavour February 11-22, 2000, the
SRTM successfully collected data over 80 percent of the Earth’s land surface, for most of the area between 60º N. and
56º S. latitude. The SRTM system included the Spaceborne Imaging Radar-C (SIR-C) and X-band Synthetic Aperture
Radar (X-SAR) systems that had flown twice previously on other space shuttle missions. The American C-band system
SIR-C operated with a wavelength of λ = 5,6 cm, the wavelength of the German / Italian X-band system was λ = 3,1 cm.
In order to obtain a global coverage between 60 degrees north and 58 degrees south the Shuttle was flown at an altitude
of 233 km with an inclination of 57 degrees.
The SRTM DEM products are being distributed mainly under two forms; these are the SRTM 1” and SRTM 3”, with
spatial resolution of 1 arc second and 3 arc seconds, respectively. The first is available only for the United States, while
the latter is freely distributed for the rest of the globe. Both datasets can be downloaded from various data gateways; one
is the USGS seamless data distribution website19. According to mission specifications20, SRTM was expected to generate
DEMs with a vertical RMSE of 16 m. The equivalent vertical accuracy requirements for topographic data at scale
1:250.000 that meet the USA map accuracy standards10,21 indicate that RMSE should be around 15,3 m.
As the data became available several studies have been published concerning the accuracy of SRTM DEM. It was
shown22 that in the case of Kos Island, Grece, the SRTM DEM is very similar to the DEM from the digitized contours
from the 1/50.000 topographic maps. The two DEMs give almost identical 3D representation of the island and they
present quite similar statistics. The digitized drainage network fit very well to the relief of the two DEMs. There is also a
strong positive correlation between the two DEMs22. Comparison between the SRTM and ASTER DEM was made for
two regions in Crete Greece8. The RMSE was used to evaluate the vertical accuracy of SRTM DEM. The RMSE was
calculated at the value of 41.6 m for the case of North Heraklion, whereas for the case of Sitia it was calculated at the
value of 46.4 m. Both values were considered quite satisfactory for SRTM derived DEM8.
SRTM elevation data was obtained from the University of Maryland (Global Land Cover Facility). The SRTM elevation
data used in this study belong to the p183r034_utm.tif tile. The minimum value (representing the voids) was –32,768 and
the maximum value was 2365. The original SRTM elevation data (Figure 3) were reprojected to the HGRS 87 using the
nearest neighborhood resampling method and preserving the 3-arc pixel size. A first spatial filter was applied to the
SRTM DEM. This filter detects all the negative values and changes them to the value – 5, resulting a DEM with less
noise and homogenous voids. Then, a second spatial filter was applied to the SRTM DEM to interpolate the missing
values. The filter initially detected the areas with the specified value of -5. If the majority of the pixels around this area
had to be replaced, then a low pass filter 5 x 5 was applied; otherwise a low pass filter 3 x 3 was applied. As a result an
interpolated DEM without any voids is created (Figure 4).Finally, in order to be comparable with the other two DEMs
the pixel size of the SRTM DEM was resampled to 15m.
4. DEM FROM TOPOGRAPHIC MAPS
The most popular data sources for the creation of DTM are the digitized contours of the topographic maps. In this study
we used the 1/50.000 topographic maps of the Hellenic Army Geographical Service that covers the broader area of
Athens. We digitized the elevation contours. The contour interval is 20m. From the digitized contours we created a new
DTM with a 15m-pixel size (Figure 5).
5. NEED FOR UPDATING THE 1:50.000 TOPOGRAPHIC MAPS
As already mentioned the usual update rate for the topographic maps ranges from ten to twenty years. As a result in
many cases a lot of changes have been done during this period and there is a need for updating. Also in some other cases
there are mistakes in the topographic maps that have to be located and corrected. From a simple optical comparison of
the three DEMs used in this study both problems have been detected.
There is an area in the urban centre of Athens where the DEM from the topographic maps present a quite big error to the
elevation data. As it can be observed in Figure 7 (upper left part) in the DEM from the topographic maps there is a small
hill (position of the cursor). This hill was not observed in the other two DEMs (Figure 7, upper right and lower left). An
orthorectified Ikonos image with one meter spatial resolution was used to check the specific area. It can be easily
observed that there is no hill in that area and thus the DEM from the digitized contours has to be corrected.
The new airport of Athens was built at about 40 km the urban centre in the area of Spata. A new highway “ATTIKI
ODOS” was constructed to connect the new airport and the capital. This highway can be located in the ASTER DEM
(Figure 8) but it cannot be detected in the DEM of the topographic maps. Thus there is a need for updating the elevation
data along the new highway.
Proc. of SPIE Vol. 6366 636606-3
6. DEM COMPARISON AND ACCURACY CONTROL
6.1 Statistical Comparison
The statistical parameters of the three DEMs were examined. As it can be observed in Table 1 the three DEMs give a
minimum value of 0 (elevation of the sea surface) as expected. The DEM from the digitized contours gives a maximum
value of 1022. The respective higher values for the ASTER DEM and the SRTM DEM are 1016 and 1021.
The mean value of the three DEMs is almost the same. The difference between the ASTER DEM and the DEM from the
digitized contours is 1.48. The mean value difference between the SRTM DEM and the DEM from the digitized contours
rises to 2,53. The standard deviation values are also almost the same. The standard deviation value of the SRTM DEM is
1.18 higher than the value of the DEM from the digitized contours.
The elevation difference between the three DEM was also calculated. From the DEM from the digitized contours the
other two DEMs were subtracted. The mean difference between the DEM from the digitized contours and the ASTER
DEM was less than 1 meter (0.962). The mean difference between the DEM from the digitized contours and the SRTM
DEM was 2.135m.
Table 1. Statistical parameters of the two DTM’s.
Topo_50000 DEM
ASTER DEM
SRTM DEM
Minimum
0
0
0
Maximum
1022
1016
1021
Mean
45,80
46,94
48,33
Standard Deviation
101,60
102,67
102,78
6.2 Accuracy Control
Two hundreds eighty points of known elevation have been used to estimate the accuracy of these three DEMs. The
spatial distribution of the points is presented in Figures 9 and 11 for the ASTER and the SRTM DEM respectively. A
buffer zone of fifteen metes was created around those points in order to correspond with the spatial resolution of the
three DEMs. Then the elevation difference between the points of known elevation and the relative points (pixels) of the
three DEMs was calculated. The elevation difference (∆z) between ASTER DEM and the known elevation points is
presented in Figure 10. The relative difference (∆z) for the SRTM DEM is presented in Figure 12. It was calculated that
the RMSE for the ASTER DEM is 10,05 meters and the RMSE for the SRTM DEM is 33,21. The RMSE of the ASTER
DEM is acceptable for updating the 1:50.000 topographic maps as the nominal vertical accuracy of these maps is 10m.
The RMSE of the SRTM DEM is quite big but in any case the SRTM DEM can be used for locating areas where the
topographic maps need updating.
7. CONCLUSIONS
An ASTER and a SRTM DEM have been compared to a DEM from digitized contours from 1:50.000 topographic maps.
The optical comparison of the DEMs detected errors in the elevation data of the DEM from the digitized contours.
Significant changes have been also detected and thus there is a need for updating the topographic maps.
The vertical accuracy of the elevation data of the ASTER and the SRTM DEM has been also checked using two hundred
eighty points of known elevation. It was calculated that the RMSE for the ASTER DEM is 10,05 meters and the RMSE
for the SRTM DEM is 33,21. The accuracy of the ASTER DEM is considered as suitable for updating 1:50.000
topographic maps and the SRTM DEM is suitable for the detection of areas that need updating.
REFERENCES
1.
2.
3.
Kim, S., and Kang, S., 2001. Automatic Generation of a SPOT DEM: Towards Coastal Disaster Monitoring.
Korean Journal of Remote Sensing, 17(2), pp. 121-129
Vadon, H., 2003. 3D Navigation over Merged Panchromatic –Multispectral High Resolution SPOT5 Images. In:
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol.
XXXVI, 5/W10
Lee, H-Y., Kim, T., Park, W., and Lee, H-K., 2003. Extraction of digital elevation models from satellite stereo
images through stereo matching based on epipolarity and scene geometry. Image and Vision Computing,
ELSEVIER, 21, pp.789-796.
Proc. of SPIE Vol. 6366 636606-4
4.
5.
6
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
Toutin, T., 1998. Evaluation RADARSAT Stereoscopic Pairs For DEM Generation, IEEE.
Toutin, T., Chénier, R., and Carbonneau, Y., 2001. 3D Geometric Modelling of Ikonos Geo Images. In: CDROM, ISPRS Joint Workshop “High Resolution from Space”, Hannover, Germany.
Zhen, X., Huang, X., and Kwoh, L., K., 2001. Extracting DEM from SPOT Stereo Images. 20th Asian Conference
on Remote Sensing, Singapore.
ABRAMS, M., 2000, ASTER: data products for the high spatial resolution imager on NASA’s EOS-AM1
platform. International Journal of Remote Sensing, 21, 847-861.
Konstantinos G. Nikolakopoulos, Emmanuel K. Kamaratakis & Nektarios Chrysoulakis, 2006. “SRTM vs
ASTER elevation products. Comparison for two regions in Crete, Greece.”, International Journal of Remote
Sensing”. In Press.
Ai-Rousan, N. And Petrie, G., 1998, System calibration, geometric accuracy testing and validation of DEM and
orthoimage data extracted from SPOT stereopairs using commercially available image processing systems.
International Archives of Photogrammetry and Remote Sensing, 32, 8 - 15.
Lang H & R Welch, 1999. ATBD-AST-08 Algorithm Theoretical Basis Document for ASTER Digital Elevation
Models. Standard Product AST14 Report, The Jet Propulsion Laboratory, California Institute of Technology, Los
Angeles, USA.
Hiroyuki Fujisada, G. Bryan Bailey, Glenn G. Kelly, Seiichi Hara, and Michael J. Abrams, 2005. “ASTER DEM
Performance”, IEEE Transactions On Geoscience And Remote Sensing, Vol. 43, No. 12, p.2707-2714.
Chrysoulakis, N., Diamandakis, M. And Prastacos, P., 2003, GIS Integration of ASTER stereo imagery for the
support of Watershed Management. In: Lekkas, T. D. (Ed), Proceedings of the 8th International Conference on
Environmental Science and Technology, pp. 111 – 118.
Chrysoulakis, N., Abrams, M., Feidas, H. And Velianitis, D., 2004, Analysis of ASTER Multispectral Stereo
Imagery to Produce DEM and Land Cover Databases for Greek Islands: The REALDEMS Project. In: Prastacos,
P., Cortes, U. De Leon, J. L., Murillo, M. (Eds): Proceedings of e-Environment: Progress and Challege, pp. 411 –
424.
Konstantinos G. Nikolakopoulos, Dimitrios A. Vaiopoulos, Georgios Aim. Skianis, 2004, “Comparing a DTM
created with ASTER data to GTOPO 30 and to one created from 1/50.000 topographic maps”, Proc. of SPIE, Vol.
5574, p. 43-51.
Konstantinos G. Nikolakopoulos, Dimitrios A. Vaiopoulos, Georgios Aim. Skianis, 2004, “ASTER DTM vs
GTOPO30 and 1/250.000 topographic maps. The case of Lefkas Island, Greece”, ΙΕΕΕ 2004 International
Geoscience and Remote Sensing Symposium, Anchorage Alaska 20-24 September 2004,Vol. III p. 2237-2240,
ISBN: 0-7803-8742-2.
Akira Hiranoa, Roy Welch, Harold Lang, 2003. “Mapping from ASTER stereo image data: DEM validation and
accuracy assessment” ISPRS Journal of Photogrammetry & Remote Sensing, 57 p. 356– 370.
Werner, 2001. Status of the SRTM data processing: when will the world-wide 30m DTM data be available? GeoInformationssysteme 12/2001. Herbert Wichmann Verlag, Hüthig GmbH & Co. KG. Heidelberg, pp. 6-10.
Rosen, P., Eineder, M., Rabus, B., Gurrola, E., Hensley, S., Knöpfle, W., Breit, H., Roth, A. and Werner, M.,
2001. SRTM Mission – Cross Comparison of X and C Band Data Properties. Proceedings of IGARSS, Sydney,
Australia, unpaginated CD-ROM.
USGS, 2005. SRTM DTED. US Geological Survey, URL: http://edc.usgs.gov/products/elevation/ srtmdted.html.
Hensley, S., Munjy, R., Rosen, P., 2001. In: Maune, D. (Ed.), Interferometric Synthetic Aperture Radar (IFSAR),
Digital Elevation Model Technologies and Applications: The DEM Users Manual. American Society for
Photogrammetry and Remote Sensing, Bethesda, Maryland, pp. 143–206.
Welch, R., Marko, W., 1981. Cartographic potential of spacecraft line-array camera system: stereosat,
Photogrammetric Eng. Remote Sensing, 47 (8): 1173–1185.
Konstantinos G. Nikolakopoulos, Dimitrios A. Vaiopoulos, Georgios Aim. Skianis, 2005, “SRTM DTM vs one
created from 1/50.000 topographic maps. The case of Kos Isalnd”, Proc. of SPIE, Vol. 5980, p. 122-131.
Proc. of SPIE Vol. 6366 636606-5
Fig. 1. The study area. At the east the Hymettus Mountain, at the west the Aigaleo Mountain and at the North the Pendeli
Mountain determine the boundaries of the Athens basin and of the urban centre.
A
n
:460667.00, 4192113.16 (179140974969709007 (9410960)
Fig. 2. The original ASTER DEM.
Proc. of SPIE Vol. 6366 636606-6
—
+ ''
I
I
444
Fig. 3. The original SRTM DEM and its statistics. There are many voids.
—
"
4J t''
_1_
• UUNI& I mi
SPill
I
Fig. 4. The SRTM DEM after the application of the two filters. There are no voids and the missing data were replaced by
interpolation.
Proc. of SPIE Vol. 6366 636606-7
:
Fig. 5. The SRTM DEM of the study area.
0
Fig. 6. The DEM from the digitized contours from the 1/50.000 topographic maps.
Proc. of SPIE Vol. 6366 636606-8
r19
__
.
JLJ
1415402.42, 4006053.37 (119140914969109001 1905 00601
Fig. 7. Locating areas in Athens where the elevation data need immediate update. From Upper Left to Lower Right, the
DEM from the digitized contours, the SRTM DEM, the ASTER DEM and an Ikonos orthophoto with one meter
resolution. The small hill that appears in the DEM from the digitized contours doesn’t exist.
FiI Utility VII VOIRtt HUp
JJ2j
FIV Utility Vylt VOl RIttly HOp
469095.65, 4190143.06 (T,ooooo,446430o0031946 0960)
Fig. 8. Locating areas in Athens where the elevation data need immediate update. At the left the DEM from the digitized
contours. At the right the ASTER DEM. The new highway “ATTIKI ODOS” that connect the new airport “Elefterios
Venizelos” with Athens is very well mapped from the ASTER DEM. It doesn’t appear to the DEM from the digitized
contours.
Proc. of SPIE Vol. 6366 636606-9
'c- r
In-
Fig. 9. The spatial distribution of the two hundreds eighty points of known elevation on the ASTER DEM.
-
IS
5
147
Iliilliiilii
I—I
I—I--I
1i11i
i
I!
ID 13 lB 19 2226 28 31 3437 40 4348 49 62 66 68 81 8487 70 73 78 7982 8688 91 94 97100
Az ISM - ASTER
Fig. 10. Histogram of the elevation difference ∆z between the ASTER DEM and 280 points of known elevation.
Proc. of SPIE Vol. 6366 636606-10
Fig. 11. The spatial distribution of the two hundreds eighty points of known elevation on the SRTM DEM.
14!
I
—
12
ID
E8
U.
IIIIHJIM__
81118212831334148616881887178818891 8E1U11U8111118121126131138141148
Az ISM - SRTMI}
Fig. 12. Histogram of the elevation difference ∆z between the SRTM DEM and 280 points of known elevation.
Proc. of SPIE Vol. 6366 636606-11