基于Shearlet变换的SAR图像自适应去噪算法

摘要

In view of the characteristics of relevant speckle noise of SAR images, this paper proposes a self-a- daptive shrinkage method for denoising based on Shearlet transform. First, the image is processed with Shearlet transform. Then, the Shearlet coefficient is contracted by the self-adaptive shrinkage method considering the noise measurement and the relevance of multi-scale spaces. Finally, one image is reconstructed through the inverse Shear- let transform using the amended coefficient. The experiments indicate that the algorithm can wipe off the relevant speckle noise of SAR images and has strong capability of maintaining the edge.%针对SAR图像相干宽噪声的特点，提出了一种基于Shearlet变换的自适应去噪算法。首先对图像进行Shearlet变换，在考虑噪声度量和各尺度空间相关性的基础上，对Shearlet系数进行自适应收缩，将修正后的系数通过Shearlet逆变换重构图像。实验结果表明，文中算法在SAR图像处理相干斑得到有效抑制的同时，具有较强的边缘保持能力。