针对三维块匹配(BM3D)算法在合成孔径雷达(SAR)图像相干斑抑制时对图像块的相似性比较造成计算量过大的问题,提出一种基于特征提取的SAR-BM3D相干斑抑制算法.首先通过梯度奇异值分解对SAR图像的平坦区和边缘纹理特征区进行描述,从而获得SAR图像的结构信息,然后生成特征区域的相似集,最后根据得到的相似集来缩小同质区的搜索区域和扩大边缘纹理区的搜索区域,达到提高SAR-BM3D相似块匹配效率的目的.通过对比实验证明,算法能够有效加快SAR-BM3D运算时间,同时获得相干斑抑制效果较为良好的SAR图像.

由于SAR是相干成像系统,在获取SAR图像时会含有相干斑噪声,使得SAR图像的解译,理解变得困难,因此有必要对SAR图像的相干斑进行抑制;首先介绍了相干斑的形成机理,之后对多视处理技术,空域滤波技术,变换域滤波技术和非局部均值滤波技术进行总结,在简要介绍各种方法的来源,理论基础和发展的基础上对比了各自的优缺点,最后展望了SAR图像相干斑抑制研究领域的进一步发展方向.更多还原

Aimed at the inaccurate problem of block similarity metric for synthetic aperture radar (SAR)image despeck-ling using Block-matching and 3D filtering (BM3D)algorithm,a new SAR-BM3D despeckling algorithm is proposed based on generalized likelihood ratio criterion (GLR).According to the good results of BM3D algorithm in optical image denoising,based on the transform domain method and nonlocal mind,considering SAR image speckle character,proba-bility similarity measure method is used to block matching thus to achieve better results.The experiment results show that the new algorithm has advantages in the SAR image despeckling performance with higher PSNR and better visual results.

Because SAR is a coherent imaging system containing coherent noise and it makes SAR image difficult to understand,it is necessary to despeckleling SAR image.In this paper,we firstly analyzes the forming of speckle,and then summarizes SAR image despeckling algorithm,such as multi-look filter technology,spatial domain filter technology,transform domain filter technology and non-local means filter technology.On the basis of the brief introduction of each algorithm' s origin,theoretical foundation and development,their advantages and disadvantages are compared.At last,the paper summarizes the prospects of the developing direction of SAR image despeckling algorithm.