In: Wagner W„ Szekely, B. (eds.): ISPRS TC VII Symposium - 100 Years ISPRS, Vienna, Austria, July 5-7, 2010, IAPRS, Vol. XXXVIII, Part 7B 
274 
But in fact, these flat areas are less correlated, because they are 
covered by uniform vegetation and not many structural elements 
can be found. 
Figure 8 shows the overlay of the correlation image and the am 
plitude image at the north-western mountain subset shown in 
Figure 7. The higher correlation values are painted in darker 
color, which is opposite to the color code used in Figure 6 and 
Figure 7, but this allows for a more meaningful visualization. 
The lakes at the left side and at the top are obviously highly cor 
related and are shown in a very dark green. The picture also in 
dicates that the mountain sides facing the sensor (towards the 
left side of the image) have low correlation values due to the 
layover effect. The mountain sides facing in far-range direction 
(towards the right side of the image) have higher correlation 
values. We can see this also when analyzing the no-data-areas 
of the Pixel Factory™ DSMs in Figure 3 and Figure 4. The 
large no-data-areas are located in near-range of a mountain. 
Figure 8. Overlay of correlation image and amplitude image 
the two-dimensional normalized correlation delivers the best re 
sults for our dataset. 
The RPC based geo-coding improved our processing speed and 
the overall geo-accuracy of our DSM. Still, the DSM created by 
Infoterra’s Pixel Factory is more precise and less noisy. There is 
therefore still a lot of room for improvement. 
ACKNOWLEDGEMENT 
The authors thank the Infoterra GmbH for providing the test da 
ta and the reference data. The work was supported by the Re 
search Fellowship for International Young Scientists of the Na 
tional Natural Science Foundation of China (Grant No 
60950110351). 
REFERENCES 
Balz, T., He, X.Y., Zhang, L., Liao, M.S., 2009. TerraSAR-X 
stereo radargrammetry for precise DEM generation in South- 
East Asia. In: Proc. Asian Conference on Remote Sensing, Bei 
jing. 
Fayard, F., Meric, S., and Pottier, E., 2007. Matching stereos 
copic SAR images for radargrammetric applications. In: Proc. 
Geoscience and Remote Sensing Symposium (IGARSS 2007), 
Barcelona, Spain. 
Fraser, C.S. and Hanley, H.B. 2005. Bias-compensad RPCs for 
Sensor Orientation of High-resolution Satellite Imagery. Photo- 
grammetric Engineering & Remote Sensing, 71, pp. 909-915. 
Infoterra, 2009. Pixel Factory™, the Next Generation Solution 
for Industrial Geo-Production. Available online: 
http://www.infoterra.if/Infoterra/Pixel_Factory.php [last ac 
cessed on May 26, 2010] 
Infoterra, 2010. Free TerraSAR-X data. Available online: 
http://www.infoterra.de/tsx/freedata/start.php [last accessed on 
June 3,2010] 
5. CONCLUSION 
Figure 9. 3D model of the DSM 
Stereo radargrammetry with TerraSAR-X can provide precise 
DSMs even in mountainous and strongly vegetated areas as 
shown above and in Figure 9. We achieved a low mean error in 
the DSM, but the standard deviation of the error is comparably 
high. Compared with our previous work (Balz et al. 2009), we 
improved the accuracy and strongly reduced the number of out 
liers in the data by using the SRTM data and by improving the 
processing. We tested different matching criteria but found that 
Leberl, F., 1990. Radargrammetric Image Processing. Artech 
House, Boston, MA. 
Raggam, H., Gutjahr, K., Perko, R., and Schardt, M., 2010. As 
sessment of the stereo-radargrammetric mapping potential of 
TerraSAR-X multibeam spotlight data, IEEE Transactions on 
Geoscience and Remote Sensing, 48, pp. 971-977. 
Tao, C.V. and Hu, Y., 2001. A comprehensive study of the ra 
tional function model for photogrammetric processing. Photo- 
grammetric Engineering & Remote Sensing, 67, pp. 1347-1357. 
Toutin, T., 2000. Evaluation of radargrammetric DEM from 
RADARSAT images in high relief areas. IEEE Transactions on 
Geoscience and Remote Sensing, 38, pp. 782-789. 
Tupin, F. and Nicolas, J.M., 2002. Matching criteria for radar 
grammetry. In: Proc. IGARSS 2002, Toronto. 
Volpe, F., 2004. RPC based processing of QuickBird high reso 
lution satellite data. In: Proc. Polish Remote sensing and GIS 
Conference 2004, Bialobrzegi, Poland. 
Wei, X.H., He, X.Y., Zhang, L„ Balz, T., Liao, M.S. 2010. RPC 
modelling for spacebome SAR and its application in radar im 
age geocoding. In: Proc. 2010 Dragon-2 Mid Term Results 
Symposium, Guilin, China.