Optimal classification of polarimetric SAR images using segmentation

The paper presents an optimised polarimetric segmentation technique for synthetic aperture radar (SAR) images, based on a generalised maximum likelihood approach. A full theoretical derivation is presented, together with a closed form analytical performance evaluation. The technique is compared to other known polarimetric segmentation schemes by application to a polarimetric SAR image of agricultural areas. A complete characterisation of the technique is provided in terms of polarimetric sensitivity and memory requirements.