一种改进全变分的图像去噪算法模型  

An improved total variation algorithm model for image denoising

作　　者：赵鑫春 李碧原 张军[1] ZHAO Xinchun;LI Biyuan;ZHANG Jun(College of Electronic Engineering,Tianjin University of Technology and Education,Tianjin 300222,China)

机构地区：[1]天津职业技术师范大学电子工程学院,天津300222

出　　处：《计算机辅助工程》2022年第3期42-48,54,共8页Computer Aided Engineering

基　　金：国家自然科学基金青年科学基金(11601372);天津市教委自然科学科研计划项目(2020KJ124);天津职业技术师范大学科研启动基金(41401-KRKC012007)。

摘　　要：为减少光学相干层析成像(optical coherence tomography, OCT)散斑噪声的干扰,利用变分图像分解原理提出一种新的全变分(total variation, TV)算法模型TV-Hilbert-Shearlet,将OCT图像分为背景、结构和噪声3个部分,将背景图与结构图相结合形成结果图,去噪效果良好。2幅有病变的OCT人体视网膜图像实验表明,TV-Hilbert-Shearlet算法能够去除图像散斑噪声,同时保持图像较完好的层状结构和边缘结构。To reduce the interference of speckle noise in optical coherence tomography(OCT), a new total variation(TV) algorithm model TV-Hilbert-Shearlet is proposed based on the principle of variation image decomposition. The OCT image is divided into three parts, that are background, structure and noise. The background image are combined with the structure image to form the result image, and its denoising effect is satisfied. Two OCT images of human retina with lesions are selected for the experiment, and the results show that the TV-Hilbert-Shearlet algorithm can remove the speckle noise of the image while maintaining the perfect layered structure and edge structure of the image.

关 键 词：图像去噪 散斑噪声 变分图像分解 全变分算法 光学相干层析成像 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP391.99[自动化与计算机技术—计算机科学与技术] TB115.1[理学—数学]