Fundamental Study on Semi-blind Scattering Component Decomposition for POLSAR data using ICA

Abstract:
Image decomposition of POLSAR data is an important technique for many applications. The model component decompositions have a preferable feature because they decompose the image into several fundamental components corresponding to their physical characteristics. However, negative-power problem often occur when model mismatch happens. Definition of the model for the volume scattering is of importance. However, general volume model available for urban area is still a problem. To prevent the problem, we investigate a decomposition technique by using Independent Component Analysis. Concept of minimum redundancy array is also introduced to enhance the degree-of-freedom of the observed data sets.