Polarimetric scattering indexes and information entropy of the SAR imagery for surface classification

The Mueller matrix solution and eigen-analysis of the coherency matrix for completely polarimetric scattering have been applied to analysis of the SAR (synthetic aperture radar) imagery. Usually, the polarization index is defined as a parameter to classify the difference between co-polarized scattering signatures from the terrain surfaces. In this paper, the eigen-values of the coherency matrix and information entropy are derived to directly relate with co-polarized and cross-polarized indexes. Thus, it combines the Mueller matrix simulation, the information entropy of the coherence matrix, and two polarization indexes to yield an overall theory for quantitative understanding of the SAR imagery. This theory is applied to the AirSAR images.