一种改进权重的非局部均值图像去噪算法  

Improved weighted non-local mean algorithm filter for image denoising

作　　者：赵庆平[1] 陈得宝[1] 姜恩华[1] 方振国[1] 

机构地区：[1]淮北师范大学物理与电子信息学院,淮北235000

出　　处：《电子测量与仪器学报》2014年第3期334-339,共6页Journal of Electronic Measurement and Instrumentation

基　　金：安徽高校省级自然科学研究项目(KJ2013Z228);安徽省自然科学基金(1308085MF82);国家自然科学基金(61304082)项目资助

摘　　要：提出了一种改进权重的非局部均值滤波方法。在高斯加权的欧氏距离基础上,结合相关系数来衡量图像邻域间的相似性,将其应用到图像邻域灰度矩阵间的相似性度量上,更好地利用了图像邻域间的相似性质。通过对添加不同噪声水平的噪声图像进行测试,实验结果表明,与传统的非局部均值滤波算法相比,所提出的算法在去噪性能上尤其是结构信息保持上均有显著提高。An improved weighted non-local mean algorithm for image denoising is proposed in this paper.Traditional non-local mean algorithm utilizes Gaussian weighted Euclidean distance to measure the similarity between patches in one image.In this paper,a novel weight combined the Euclidean distance used in the original NLM algorithm with correlation coefficient is proposed to describe the similarity between the image patches well.The proposed method has been evaluated on testing images with various levels noise.Compared to the traditional non-local means algorithm,the proposed method improves the denoising performance as well as the preservation of structure information.

关 键 词：图像去噪 非局部均值 高斯噪声 加权平均 相关系数 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TN911.73[自动化与计算机技术—计算机科学与技术]