Classification using polarimetric and interferometric SAR-data

Abstract

The investigation presented in this paper demonstrate a first order approach to an automatic classification and extraction of cartographic relevant features from SAR data. The authors propose a fusion of polarimetric and interferometric classification techniques that is able to solve several classification ambiguities which are not resolvable with one method alone and is also able to improve significantly the accuracy of the classification results. The complimentarity of the polarimetric and interferometric coherence based classification approaches and the improvements resulting from their combination are demonstrated using data from the space-shuttle-borne SIR-C/X-SAR radar system.