An improved SAR image speckle filter
Zhuang Li, Kunpeng Wang · 2010
An improved SAR image denoising method based on sigma filter was proposed here. Firstly, we analyzed the traditional sigma filter and showed its weakness in removing complex noise while maintaining sharpness of edges. Secondly, the sigma filter was modified to accommodate to image with complex noise. Then, to maintain the sharpness of weak edges, we put forward a new method named grey center offset sigma filter (GCOSF). In GCOSF, canny detector is utilized to detect regions containing edges. We adopt grey center offset value instead of center value to calculate filtering result. Last, experimental results demonstrate that the GCOSF performs better than traditional sigma filter and several other filters by considering the effect of noise removing and edge keeping simultaneously.