自适应非局部极化SAR影像去噪算法

An adaptive non-local PolSAR image denoising method

针对极化合成孔径雷达(PolSAR)影像存在较严重的相干斑噪声问题,该文提出一种自适应非局部去噪算法。定义和计算PolSAR影像像素的粗糙度,据此自适应确定搜索窗口的尺度;在搜索窗口内结合子块像素协方差矩阵的统计特征进行相似块匹配,筛选同质像素;计算像素的统计概率值,来衡量像素是代表性样本的可能性,在此基础上确定滤波权重;利用线性滤波器对极化协方差矩阵进行加权平均实现影像去噪。利用高分三号PolSAR影像数据验证提出算法的有效性,并将本文算法与3种对比算法的降噪结果进行定性和定量比较分析,结果表明,该文提出算法在较好抑制相干斑噪声的同时,也能更好地保持地物目标的结构信息和极化特征。

Aiming at the serious speckle noise problem in polarimetric synthetic aperture radar(PolSAR)images,an adaptive non-local denoising method was proposed.Firstly,the roughness of the pixel was defined and calculated on the PolSAR image,and the scale of search window was determined adaptively;Secondly,within the search window,the similar patches matching was performed by combining the statistical characteristics of the pixel covariance matrix in the sub-block image,so as to select the homogeneous pixels;Then,the statistical probability of pixels was calculated,which was utilized to measure the possibility that pixels were representative samples.On this basis,the filtering weight was computed;Finally,the weighted average of polarization covariance matrix was realized by the linear filter to achieve image denoising.The effectiveness of the proposed algorithm was verified by using two GF-3 PolSAR images that contain different kinds of objects.And the filtered results of the proposed algorithm were compared with those of three comparison algorithms qualitatively and quantitatively.The results showed that the proposed method can suppress speckles noises to a certain extent while performing better in preserving the structure information and polarization characteristics of objects.