摘要
根据图像直方图的特点,针对椒盐噪声,提出了一种基于直方图的加权均值滤波方法。该方法首先通过寻找局部极值确定噪声点,并对图像的所有像素做分类标记。处理过程中只考虑标记为噪声的点,以噪声图像的直方图函数作为滤波器的权值,最后将领域内非噪声点的加权均值作为滤波输出。实验结果表明该方法优于中值滤波方法。
According to the characteristics of histograms of images, A weighted mean filtering algorithm is presented according to the salt and pepper noise. It first determines the noisy point by looking for the local extremum and the classified markers are made for all pixels of the image. In the image processing, only the marked noisy points are taken into account and the histogram function of the corrupted image is used as weight of the filter, in the end, the weighted mean values of non-noisy points in the neighborhood are taken as the output of the filter. Primary experiment suggests that this method is superior to median filtering algorithm.
出处
《微计算机信息》
北大核心
2006年第05S期202-204,共3页
Control & Automation
基金
国家自然科学基金(60275028)资助项目
关键词
椒盐噪声
直方图
加权均值滤波
中值滤波
salt and pepper noise
histogram
weighted mean filtering
median filtering
作者简介
唐彩虹(1980-),女,硕士研究生,主要研究方向:数字图像处理.E—mail:iristang@sina.com.cn或者tangeaihong2002@sohu.com 通讯地址:(510632广州市黄埔大道西601号暨南大学真茹25栋434房)
蔡利栋(1947-),男,教授,主要研究方向:计算机视觉、图像处理、模式识别、网络信息安全。