In: Wagner W„ Székely, B. (eds.): ISPRS TC VII Symposium - 100 Years ISPRS, Vienna, Austria, July 5-7, 2010, IAPRS, Vol. XXXVIII, Part 7B 
TOPOGRAPHIC ESTIMATION BY TERRASAR-X 
U. G. Sefercik a ’ *, U. Soergel b 
a Zonguldak Karaelmas University, Engineering Faculty, Department of Geodesy and Photogrammetry Engineering, 
67100 Zonguldak, Turkey - ugsefercik@hotmail.com 
b Leibniz University Hannover, Institute of Photogrammetry and Geoinformation, 30167, Hannover, Germany - 
soergel@ipi. uni-hanno ver. de 
Commission VII, WG VTI/2 
KEY WORDS: SAR, Estimation, DEM, Generation, Comparison, Accuracy, Visualization, Analysis 
ABSTRACT: 
SAR Interferometry (InSAR) is a technique to derive Digital Elevation Model (DEM) from at least two complex SAR images. The 
data are either taken simultaneously (single-pass mode) or sequentially (repeat-pass mode) by airborne or space-bom sensors 
(carriers: plane, satellite, shuttle etc.). To date, one of the most important single-pass interferometry measurement campaigns is the 
Shuttle Radar Topography Mission (SRTM) completing its mission successfully after 11 days of operation between 11 th and 22 nd of 
February 2000. On the other hand, repeat-pass InSAR has been used by several satellite systems: ENVISAT, ERS 1-2, RADARSAT 
1-2, ALOS, JERS-1 etc. One of the most advanced systems is the German TerraSAR-X (TSX) satellite launched on June 15 th , 2007. 
TSX offers high resolution (~lm by Spotlight mode) imagery which could not been achieved from radar technologies up to this time 
similar to high resolution optical imagery. In contrast to optical sensors, TSX can be operated under all weather conditions without 
being influenced by clouds. The data sets provided by TSX newly obtained by scientific community and evaluations are currently 
being performed. As mentioned above, utilizing the advantages of SAR technology, indeed the planimetric locations of target ground 
objects, elevations of them can be determined using interferometry. Through the interferometric data, interferograms (fringe maps) 
can be generated and applying interferometric processing steps height models can be created for large coverage interest areas. The 
main targets of this investigation can be summarized as; generation of height models derived from TSX InSAR image-pairs and 
evaluation by comparison with more accurate reference height models as well as height models based on high resolution optical 
satellite images. Absolute and relative accuracy, stability, homogeneity and dependency upon various parameters are determined. 
The approach will be demonstrated using TSX data covering Istanbul area, Turkey. 
1. INTRODUCTION 
As it is known, Radar remote sensing has a significant role in 
remote sensing technologies and develops rapidly. Once in a 
few years new SAR (synthetic aperture radar) satellites are 
launched to space including various types of operation modes 
which offer different resolutions and advantages to each other. 
The most actual SAR satellite is German TerraSAR-X was 
launched to its orbit on 15 th June 2007. This satellite is a 
revolution for the SAR technologies at the resolution side and 
offers lm high resolution in Spotlight mode. To assess the 
quality of DSM which is derived from this high resolution data 
of this current satellite a comparison has been made between the 
DSM of one of the most advanced optical systems IKONOS. 
For this aim, the DSMs have been generated using both 
satellites’ data in Istanbul, Turkey and compared with a more 
accurate DEM of same area, produced by photogrammetry. 
2. TEST FIELD AND DATA SETS 
2.1 Test Field 
Istanbul is located in north-west area of Turkey. The Greater 
Municipality Area has a coast line to the Black Sea and the 
Marmara Sea connected by the Bosporus. Istanbul is one of the 
biggest cities in the world. About 14 million people leave in the 
city and most of settlements are at the surrounding of Bosporus 
and coast line of Marmara Sea. The city is a suitable test area 
for the accuracy analysis because its topography contains 
various characteristics and this enables to understand the quality 
of evaluated DEMs in different types of terrain formations. In 
patches, terrain is open-flat, hilly-steep and woody. The test 
area is a part of Istanbul and covers 10kmx8km. It includes the 
historical peninsula and near surroundings. Historical Peninsula 
(Old City) is one of the most important regions in Istanbul, 
located on the European side, neighbored to the Bosporus and 
Marmara Sea. This part named as Historical Peninsula because 
of its historic heritage. Figure 1 shows the high resolution 
satellite image of the test field with the frequency distribution of 
terrain inclination. This area has smoother topography in 
relation to the rest of Istanbul. The elevation reaches from sea 
level up to 130m. 
45 
40 
„ 35 
t 30 
725 
c 20 
5 
0 
0 0.05 0.1 0.15 0.2 025 0,3 0.35 0.4 
Tan(slope) 
Figure 1. Test area and slope distribution of terrain inclination 
Umut Giines Sefercik, ugsefercik@hotmail.com