Abstract
A method is proposed herein PALSAR (Phased Array-type L-band Synthetic Aperture Radar) data to improve the detection of sea ice. This method utilizes scattering entropy calculated from SAR data to discriminate sea ice from open water. In this study, we used sea ice reference derived from MODIS (Moderate Resolution Imaging Spectro-radiometer) visible and near-infrared data. The proposed method is applied to fully polarimetric PALSAR data acquired in 2009 and 2010 that covers the Sea of Okhotsk. We applied linear discriminant analysis to samples of sea ice and open water in order to compare our method with the conventional method based on backscattering coefficients. We found that sea ice and open water were best differentiated using scattering entropy compared to any combination of backscattering coefficients.
