Sar Simulatio N Based Change Detectio N With High-Reso Lutio N Sar Images In Urban Enviro Nments

C mbined pr cessing using different sens r types, i .e. f r applicati ns like change detecti n, require s a g d ge-referencing. Furtherm re the individual sens r pr perties have t b e taken int acc unt. SAR systems are side-l king and run-time systems. They suffer fr m cclusi ns and ambiguities especially in urban areas. Additi nally lay ver and shad w effe cts disturb the ge-referencing f SAR images in urban areas, which is a prerequisite f r a successful change detecti n. An impr ved g e-referencing can be achieved by simulating 3D-city m dels r street dat asets using a SAR simulat r and c mparing the simul ated image t the real image. C rresp ndences between simulated and real image can be used f r ge-referencing the image accrding t the c rdinates f the 3D-city m del r street dataset. The ge-refere nced dataset can afterwards be used f r change dete cti n analysis. SAR images represent a side-view f the three dimensi nal w rl d. An aut mated change detecti n using SAR images sh uld take this fact int c nsiderati n and theref re sh uld use 3D-m dels as reference f r the change-detecti n. These m dels a re simulated and the simulated image is c mpared t the ge-referenced i mage, revealing changes between the simulated m del and the real image. 1. INTRO DUCTIO The urban envir nment is f the utm st imp rtance f r human s ciety. I n 2001, ar und 50% f the human p pulati n lived in cities and these numbers are still rising, especially in less develped c untries (UNCHS, 2001). The dense placement f buildings in cities requires a g d res luti n f the re m te sensing systems, t distinguish between the neighb uring bu ildings. M dern high-res luti n airb rne SAR systems reach very high res luti ns up t 10cm (Ender & Brenner, 2003) and are theref re usea..