一种带椒盐噪声的图像恢复快速方法  

Fast image restoration method for salt-and-pepper noises removal

作　　者：谭家敏 张建军[1] 

机构地区：[1]上海大学理学院,上海200444

出　　处：《应用数学与计算数学学报》2016年第1期104-112,共9页Communication on Applied Mathematics and Computation

基　　金：上海市自然科学基金资助项目(15ZR1416300)

摘　　要：研究被椒盐噪声干扰的模糊图像恢复问题.提出了一种利用两步法来恢复图像的快速算法.第一步,用自适应中值滤波(adaptive median filter,AMF)识别被噪声干扰的图像的像素.第二步,基于无噪声的像素,对图像进行恢复.利用交替方向极小化,提出了一种带椒盐噪声的图像恢复快速方法.实验结果显示,该方法较其他现有的两步法恢复图像的效果更好.The problem of restoring blurred images corrupted by salt-and-pepper noises is studied. A two-phase approach is proposed. In the first phase, the adap- tive median filter （AMF） is used for identifying image pixels corrupted by noise. In the second phase, images are deblurred and denoised simultaneously using essen- tially the noise-free data. By using the alternating direction minimization, a fast and effective method for image restoration corrupted by salt-and-pepper noises is proposed. Experimental results show that the quality of restored images by this method is competitive with those restored by the other existing two-phase image restoration methods.

关 键 词：图像去模糊 交替方向法 椒盐噪声 全变分 正则化 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]