The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B4. Beijing 2008
not enough information about the existence and the position of
the bridge (blue lines). Automatic image tools could detect the
position of the bridge more robustly together with the 3D
information from the 3K DSM.
Figure 11 Difference between the 3K DSM from 17 th June
2007 and the LIDAR DEM (bottom) and the
corresponding orthophoto (top).
Figure 12 DSM of bridge with road data base segments
(bottom) and Orthophoto (top)
5. DISCUSSION AND OUTLOOK
The 3K camera system was proved to be a robust and accurate
tool for the automatic generation of DSMs. The generation of
DSMs was performed with a hierarchical based matching
followed by region growing algorithm. In particular for disaster
monitoring, the results respective to the reached accuracy, the
outlier detection, and the point density were quite satisfying as
demonstrated in four applications. The main bottleneck of the
DSM generation is the processing time in particular of the
matching, which proved to be too long for this kind of
applications based on our implementation.
A lot of progress has been made in the last years to speed up the
matching algorithms. By using programmable 3D hardware, e.g.
graphic processing units, the processing time for matching
reduces drastically, e.g. for 512x512 pixels stereo epipolar
patches the processing time is 32ms (Zach 2004). Different
techniques for fast matching using the OpenGL interface were
developed, e.g. space sweeping (Bauer 2006) or mesh-based
stereo algorithms (Yang 2003). Currently, tools for non epipolar
images and color matching are under development.
Future work will be the integration of ideas to speed up the
generation of DSMs for disaster monitoring applications.
ACKNOWLEDGEMENT
The authors would like to thank Rolf Stätter from the German
Aerospace Center (DLR) and Alexandra Kollmeier from the
University Munich for their support in the validation of digital
elevation models.
REFERENCES
Bauer, J., Zach, C., Kamer, K., 2006. Efficient Sparse 3D
Reconstruction by Space Sweeping. Technical Report VR- VIS,
Graz, Austria.
Heipke, C., 1996. The Evaluation of MEOSS Airborne Three-
Line Scanner Imagery: Processing Chain and Results.
Photogrammetric Engineering & Remote Sensing, Vol. 62, No
3, pp 293-299.
Kurz, F., Müller, R., Stephani, M., Reinartz, P., Schroeder, M.
2007. Calibration of a wide-angle digital camera system for
near real time scenarios. In: ISPRS Hannover Workshop 2007,
High Resolution Earth Imaging for Geospatial Information,
Hannover, 2007-05-29 - 2007-06-01, ISSN 1682-1777
Lehner, M., Gill, R.S. 1992. Semi-automated derivation of
digital elevation models from stereoscopic 3-line scanner data.
Proceedings of Satellite Symposia 1&2 from the International
Space Year Munich, Germany, 30 March - 4 April 1992.
Otto, G.P., Chau, T.K.W., 1989. ‘Region growing’ algorithm
for matching of terrain images. Image and Vision Computing,
7(2): 83-92.
Yang, R., Polleieys, M. 2003. Multi-resolution real-time stereo
on commodity graphics hardware. In Conference on Computer
Vision and Pattern Recognition.
Zach, C., Kamer, K., and Bischof, H. 2004. Hierarchical
disparity estimation with programmable 3D Hardware,
International Conference in Central Europe on Computer
Graphics, Visualization and Computer Vision
