基于相似性验证与子块排序的NSST域SAR图像去噪

为了改进传统的非局部变换域合成孔径雷达（synthetic aperture radar，SAR）图像去噪算法不考虑子块关系的缺点，结合相似性验证与子块排序提出一种新的非下采样剪切波（non-subsampled shearlet transform，NSST）域SAR图像去噪算法.构造NSST域SAR图像相似块之间距离的密度分布；利用子块之间的相似性，去除相似性较低的子块；结合子块排序和最优一维滤波对SAR图像进行去噪.实验结果表明，与其他经典去噪算法相比，等效视数平均提升6.92，边缘保持指数更接近1，无参考质量评价指数平均降低2.51，能更好地保持图像边缘和纹理信息，改善图像的视觉效果. 

SAR Image Denoising Based on Similarity Validation and Patch Ordering in NSST Domain

In order to overcome the shortcoming of traditional synthetic aperture radar(SAR)image denoising algorithm in non-local transform domain without considering the patch relationship,a new SAR image denoising algorithm in non-subsampled shearlet transform(NSST)domain was proposed based on similarity validation and patch ordering.Firstly,the density distribution of distances between similar patches of SAR image in the NSST domain was constructed.Then,the patches with lower similarity were removed according to the similarity between the patches.Finally,SAR image was denoised by combining the patch ordering and the optimal one-dimensional filter.The experimental results show that,compared with other transform domain algorithms,the equivalent numbers of looks in this algorithm can increase by 6.92 on average,and the edge preservation index is close to 1.And the unassisted measure of quality can reduce by 2.51 on average.The algorithm can better maintain the image edge and texture information,and improve the visual effect of the image.