Research on the method of suppressing speckle noise in SAR image

Abstract:
Synthetic Aperture Radar (SAR) is a high resolution active imaging sensor. Because of its coherent superposition imaging mechanism, a large number of bright or dark speckle noise is scattered in the SAR image, which seriously affects the texture features and edge contour of the image and increases the difficulty of segmentation algorithm. To solve this problem, this paper proposes a filtering algorithm based on the combination of guided filtering and morphology. Firstly, the appropriate structural elements are selected for morphological processing of SAR ship image slices, and then the processed image is guided filtering to filter out the speckle noise and retain the target edge. The experimental results show that for complex multi-target SAR images, the equivalent number of views and edge preserving index of the image after filtering by the proposed algorithm can reach 59.824495 and 0.537019, respectively, and the speckle noise index is reduced to 0.129289, indicating that the proposed algorithm can more effectively suppress the speckle noise and retain the edge information of the target.