Geometric constraints for building reconstruction from InSAR data of urban areas

The automatic 3-D reconstruction of buildings for the generation of city models is of great interest for different tasks. Besides the classical photogrammetric approaches to determine object heights indirectly, range sensors like laser (LIDAR) and interferometric SAR (InSAR) play an important role since recent years. From both sensors images are produced with a ground resolution better than a meter. In this paper we focus on the analysis of InSAR data for the reconstruction of buildings in dense built-up areas. The data acquisition and the special properties of the InSAR principle are briefly described. A segmentation approach for building reconstruction is proposed. For the analysis of InSAR special effects are taken into account, which are caused by the side-looking illumination. The estimation of the signal noise is considered in the segmentation process. Besides the elevation, the intensity information is exploited to detect building areas. The results show that building reconstruction is possible from InSAR, but the achievable level of detail cannot be compared with LIDAR. Inherent geometric constraints which limit the reconstruction of buildings from SAR data of dense urban areas are discussed. A simulation of geometric effects in SAR images based on LIDAR data is carried out. This may be used for assessing different situations of data acquisition and mission planning.