An empirical optimisation strategy for Model-based Polarimetric Target Decomposition

Inhalt:
Polarimetric target decomposition methods generally have more physical parameters than equations, are thus underdetermined and hence have no unique solution. The common approach to avoid this is to make certain assumptions, thus fixing some parameters, allowing the other parameters to be solved freely. We recently showed how to obtain additional information from fourth-order statistics to find unique solutions to model-based polarimetric decompositions. The current work extends our previous analyses to include additional moments, and we implement this methodology for the 3-component Freeman-Durden decomposition. Preliminary results are demonstrated for a well known real polarimetric SAR scene.