Wheat identifying in North China with RADARSAT-2 SAR data

Compared with conventional single-polarization synthetic aperture radar (SAR), quad-polarimetric SAR observations provide more information and show potential for land cover classification. This paper evaluates the wheat identifying potential of quad-polarimetric SAR data. The fully polarimetric Radarsat-2 image for the Hebei province, China, is selected to evaluate the identification accuracy. Experimental results indicate that HH with HH-HV is the optimal linear polarization combination, with an accuracy of up to 82.3%. The Freeman-Durden decomposition can discriminate wheat more effectively, with an accuracy of about 86.8%. The combination of the Freeman-Durden decomposition and the parameter H is the best method for wheat identification, with an accuracy of up to 90.5%.