Wavelet Fusion on Ratio Images for Change Detection in SAR Images

This letter presents a novel method based on wavelet fusion for change detection in synthetic aperture radar (SAR) images. The proposed approach is applied to generate the difference image (DI) by using complementary information from mean-ratio and log-ratio images. To restrain the background (unchanged areas) information and enhance the information of changed regions in the fused DI, fusion rules based on weight averaging and minimum standard deviation are chosen to fuse the wavelet coefficients for low- and high-frequency bands, respectively. Experiments on real SAR images confirm that the proposed approach does better than the mean-ratio, log-ratio, and Rayleigh-distribution-ratio operators.