Abstract:

In polarimetric synthetic aperture radar (PolSAR) images with two different phases, the same target may appear different scattering characters due to the changed data acquisition conditions, which will have an influence on the correctness of change detection. In order to overcome this problem, a new change detection algorithm based on polarization state extracting is proposed. First, invariant targets are used to extract the optimal polarization state of the image, then received power ratio is calculated as the feature of change detection under this state. Finally, the dual-threshold segmentation method is used to achieve change detection. Fully polarimetric SAR data collected by U.S. UAVSAR system is adopted to test and verify the new approach. Experimental results show that the novel method can obtain a good change detection result with low false alarm rate and clear contour.