Building Recognition Based on Airborne and Space Borne Insar Data

Because of its independence of time of day and its all weather capability, synthetic aperture radar (SAR) has become a key remote sensing technique in the last decades. In the past, spatial resolution of space borne SAR systems was rather coarse, so the exploitation of building signature and even the reconstruction of buildings in dense urban areas based on InSAR data was restricted to commercial and modern experimental airborne SAR systems. These are capable of providing single pass InSAR data with spatial resolution well below half a meter and much better. Now, the new generation of space borne SAR satellites (e.g. TerraSAR-X, RADARSAT-2, or COSMO-SkyMed) allows a more detailed analysis at the object level even for urban areas. This gives rise to the question: Is it possible to transfer basic algorithms of building reconstruction based on airborne sensors to such data? Because of the high availability and short repeat time of the space borne sensors, they are particularly useful e.g. in case of disaster management. In this paper a known algorithm of gable-roofed building reconstruction is applied to airborne and space borne InSAR data of the same test site. The algorithm exploits the appearance of buildings in magnitude images, which is dominated by effects of the inherent oblique scene illumination, such as layover, radar shadow and salient lines of bright scattering caused by direct reflection or multipath signal propagation. Especially, in urban residential districts often salient pairs of parallel lines of bright magnitude are caused by gable-roofed buildings. By exploitation of this magnitude signature, building hypotheses are assembled. The ambiguity in the group of building hypotheses can be solved by considering additional information and by investigation in the interferometric phase signature. Therefore, interferometric phases are simulated based on the intermediate 3D building hypotheses and finally compared with the real interferometric phases. The hypothesis showing higher correlation with the real phases is chosen as final reconstruction result. The quality assessment of the reconstruction results is supported by high-resolution LIDAR data and cadastral data. The study is carried out on airborne AeS-1 and space borne TerraSAR-X data. * Corresponding author. This is useful to know for communication with the appropriate person in cases with more than one author.