SAR图像去噪的非局部均值算法改进  

Improvement of Non-Local Means Algorithm for SAR Image Denoising

作　　者：陈前 刘本永[1] CHEN Qian;LIU Benyong(Guizhou University,Guiyang Guizhou 550025,China)

机构地区：[1]贵州大学,贵州贵阳550025

出　　处：《通信技术》2022年第3期299-304,共6页Communications Technology

基　　金：国家自然科学基金项目(60862003)。

摘　　要：针对非局部均值算法在合成孔径雷达(Synthetic Aperture Radar,SAR)图像的边缘和纹理区域去噪效果较模糊的问题,探讨改进了非局部均值算法的权重部分,把高斯加权的空间距离与非局部均值算法的权重系数相结合,构造新的权重系数,提出了一种基于改进非局部均值算法的SAR图像去噪算法。首先,使用对数变换处理图像;其次,进行高斯滤波;再次,利用了所提出的算法将高斯滤波结果去噪;最后,采用指数变换方法处理去噪结果。去噪结果表明,所提算法在抑制散斑噪声和保持图像结构信息方面有较好的性能。Aiming at the problem that the denoising effect of non-local means algorithm is fuzzy in the edge and texture area of the SAR(Synthetic Aperture Radar) image, the weight part of non-local means algorithm is improved, and a new weight coefficient is constructed by combining the Gaussian weighted spatial distance with the weight coefficient of non-local means algorithm, a SAR image denoising algorithm based on improved non-local means algorithm is proposed. Firstly, the image is processed through logarithmic transformation;secondly, the Gaussian filtering is performed;thirdly, the proposed algorithm is used to denoise the Gaussian filtering results;finally, the denoising results are processed by exponential transformation method. The denoising results indicate that the proposed algorithm has better performance in suppressing speckle noise and maintaining image structure information.

关 键 词：合成孔径雷达图像 图像去噪 空间距离 非局部均值 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]