Wetland Classfication Based on Plarimetric SAR Decompositions

Abstract

Polarimetric SAR observations are very useful to characterrize and monitor the Earth's surface because they provide more information than single channel SAR data. We developed a new decomposition, which accounted for double bounce scattering in cross polarization signal. We used our four-component decomposition to characterize the wetland vegetation in the Everglades using three quad-polarimetric data sets: TerraSAR-X, Radarsat-2, and ALOS. We used the decomposition results to generate vegetation classification maps using Trimble eCognition software.