东南大学学报（自然科学版）2016，Vol.46Issue(6)：1132-1136,5.DOI:10.3969/j.issn.1001-0505.2016.06.004
基于先验信息的全变分图像复原算法
Total variation image restoration algorithm based on prior information
摘要
Abstract
In order to improve the performance of image restoration of the total variation (TV)mod-el,an improved TV image restoration algorithm based on prior information is proposed.First,the nonlocal means (NLM)filtering algorithm,which can effectively protect the structural information of the filtered image,is employed to reduce the noise within the image to restore.Thus,the filtered prior image information is obtained.Then,an improved total variation restoration model based on the obtained prior information is established.The proposed model can not only maintain the TV model' advantage of protecting the boundary information of restorated image,but also maintain the NLM model' advantage of protecting the structure information.Finally,the proposed model is opti-mized by the split Bregman alternating direction multiplier iteration algorithm and the restored image is obtained.The experimental results show that compared with other algorithms,the proposed algo-rithm achieves better restoration effect in terms of the subjective visual effect and the objective quan-titative indices such as peak signal to noise ratio (PSNR)and structural similarity (SSIM).关键词
图像复原/先验信息/全变分/非局部均值/分裂BregmanKey words
image restoration/prior information/total variation/nonlocal mean/split Bregman分类
信息技术与安全科学引用本文复制引用
张俊峰,罗立民,舒华忠,伍家松..基于先验信息的全变分图像复原算法[J].东南大学学报（自然科学版）,2016,46(6):1132-1136,5.基金项目
国家重点基础研究发展计划（973计划）资助项目（2010CB732503）、国家自然科学基金资助项目（61201344）． （）
