A new SAR image denoising algorithm of fusing Kuan filters and edge extraction

摘要

Due to the advantage of all-weather, multi-angle data acquisition, Synthetic Aperture Radar has been widely applied in many areas. However, the speckle noise affects its application seriously. Therefore, suppressing speckle noise effectively is significant for its analysis and application. Against the shortcoming that the Kuan filter can't both suppress speckle effectively and maintain the edge details suitably, we propose a new algorithm, which fuses the Kuan filter and the edge extraction technology. In this algorithm, first we use the Kuan filter with 5×5 window to process the image, we can get the filtered image which suppress the speckle effectively, but the edge details and texture information loss seriously. Then we use the edge extraction technology to get the image's edge and texture information. At last the pixel values of the edge and texture area of the filtered image are replaced by the result of the edge extraction. Experimental result shows that the improved filtering method not only suppresses the speckle effectively, but also improves the capability of edge details and texture information maintaining. Compared with the traditional filters, the proposed improved filter is effective.