An Improved SAR Image Speckle Reduction Algorithm Based on 3D Block Matching

For any Synthetic Aperture Radar (SAR) system, obtaining high-quality images is always a prerequisite for improving the performance of SAR applications. However, since the coherent imaging mechanism is affected by speckle noise, speckle suppression is required for SAR images. Aiming at the problem of block matching (BM3D) algorithm in SAR image denoising due to the problem of block matching grouping errors, this paper uses the structural similarity index to improve the block similarity measure, and when similar blocks are selected, according to the standard normal The distribution principle is used to select the threshold value of similar blocks. Compared with several other non-local mean filtering algorithms, the experimental results show that the proposed algorithm can not only effectively suppress the speckle noise in SAR images, but also maintain certain edge information details and obtain a good denoising effect.