Analysis of SAR polarimetric signatures distortions

Abstract

The purpose of this paper was to analyze SAR polarimetric signatures distortions according to known scattering mechanisms. Using canonical signatures we could compare them with real generated from SAR data. The main point is that in reality there are always distortions and sometimes it’s hard to simply determine scattering type. Analyzing SAR polarimetric signatures distortions we can affirm that many factures has an influence on signatures shape. The pixels with one dominant scattering mechanism have a polarimetric signature most similar to the corresponding canonical object's signature. As the number of scattering mechanisms grows, the polarimetric signature becomes more difficult to determine. Also rough surfaces like buildings or trees cause greater multiple scattering as compared to the smoother surfaces like water body. SAR images are affected by the multiplicative noise which also affects on signatures shape. Examining scattering mechanisms characteristics can useful in identifying objects.