一种改进的基于结构相似性的非局部均值图像去噪算法  

Based on structural similarity improved non-local means image denoising algorithm

作　　者：焦晨光 张小波 JIAO Chenguang;ZHANG Xiaobo(School of Computer Science,Xi'an Shiyou University,Xi'an 710065,China)

机构地区：[1]西安石油大学计算机学院,西安710065

出　　处：《智能计算机与应用》2025年第2期17-23,共7页Intelligent Computer and Applications

基　　金：国家自然科学基金(61401383);陕西省自然科学基础研究计划(2021JM-518)。

摘　　要：非局部均值滤波(NLM)使用像素邻域的加权平均进行图像去噪,但使用加权欧式距离测量像素邻域的相似程度有一定的局限性和不准确性。为了提高非局部均值算法的去噪效果,本文提出了一种改进的基于结构相似性的非局部均值图像去噪算法。该算法将图像质量评价指标结构相似性(SSIM)引入到像素邻域相似性测量中,提高算法保留图像结构信息的能力;并基于预去噪图像进行像素调整,降低噪声和对比度对去噪任务的影响。实验结果表明,相较于传统非局部均值去噪算法和一些基于非局部均值的改进算法,本文提出的算法能够得到更好的图像去噪效果。Non-local means filtering(NLM)uses the weighted average of pixel neighborhoods for image denoising,but using the weighted Euclidean distance to measure the similarity of pixel neighborhoods has some limitations and inaccuracies.In order to improve the denoising effect of the NLM algorithm,this paper proposes an improved NLM image denoising algorithm based on structural similarity.The algorithm introduces the image quality evaluation index structure similarity(SSIM)into the pixel neighborhood similarity measurement to improve the ability of the algorithm to retain the image structure information.The pixel adjustment based on the pre-denoised image can reduce the influence of noise and contrast on the denoising task.The experimental results show that compared with the traditional NLM algorithm and some improved algorithms based on NLM,the proposed algorithm can obtain better image denoising effect.

关 键 词：图像去噪 非局部均值滤波 结构相似性 预去噪 像素调整 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]