基于多重检测及双权值的椒盐噪声自适应滤波算法

提出一种针对椒盐噪声的扩大窗口多重检测及使用双权值的自适应滤波算法.采用对可疑噪声扩大窗口进行多次重复检测来提高对噪声的判别,采用基于图像整体相似度和像素相似度的两种权值构造加权滤波函数,采用噪声剪切和自适应窗口实现滤波.实验表明,算法对不同密度的椒盐噪声具有良好的适应性,在高密度噪声下去噪效果有明显改善.

Double weighted self-adaptive filtering algorithm based on multiple detection for salt and pepper noise images

A self-adaptive filtering algorithm for images with salt and pepper noises based on multiple detection and double weighted is presented in this paper. The accurateness of noise detection is improved by self-adaptive expanding window and multiple detection on the dubious noise pixels. Two weight factors of filter function are constructed by holistic similarity of images and similarity of pixels. The filtering process is realized by noise trimming operation and self adaptable windows. Experiments showed that the algorithm has adaptability to images with different density of salt and pepper noises and excellent performances on images with high density noises.