Optimal multi-look polarimetric speckle reduction and its effect on terrain classification

This paper presents an multi-look polarimetric whitening filtering (MPWF) method to reduce the image speckle using complete multi-look polarimetric SAR data. The amount of speckle reduction achievable by using the MPWF is quantified particularly for nonhomogeneous scenes. Experimental results with NASA/JPL L-band 4-look polarimetric SAR data are given for illustration. With the class separability index, we show how the speckle reduction can improve the terrain classification performance, and the classification results with the speckle-reduced data are also given