基于改进TV修复模型的椒盐噪声去除算法  

Salt and Pepper Noise Removal Algorithm Based on the Improved TV Inpainting Model

作　　者：杨陈东[1] 侯海娜[2] 

机构地区：[1]西安航空学院理学院,西安710077 [2]西安理工大学理学院,西安710054

出　　处：《计算机与数字工程》2016年第11期2118-2123,共6页Computer & Digital Engineering

摘　　要：椒盐噪声去除是图像处理领域中一项重要任务。将网函数插值与TV模型相结合提出Net-TV修复算法,通过噪声检测将椒盐噪声点看作缺失信息,利用Net-TV修复算法进行修复。针对三幅包含有不同水平椒盐噪声的图像进行数值实验,并将结果与多种算法进行对比。实验结果表明,Net-TV图像去噪算法有效滤除噪声的同时,能够良好保护图像边缘信息,且算法效率高,对不同密度下的含噪图像均可达到较好效果。The removal of salt and pepper noise is an important task in the image processing.Combining network function interpolation and TV inpainting algorithm,this paper presents a Net-TV inpainting algorithm.Noisy pixels are detected as missing data by noise detection and restored by Net-TV inpaining algorithm Numerical experiments are carried out for three images with defferent levels of salt and pepper noise,its results compared with a variety of algorithms.The experimental results demonstrate that the Net-TV inpainting algorithm not noly remove noise effectively but also preserve the information at the edges of image.This algorithm is more efficient than others.The algorithm can achieve good effect with noisy images at different densities.

关 键 词：椒盐噪声 图像修复 TV修复算法 网函数插值 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]