一种混合噪声图像去噪的平滑算法

提出了一种新的混合滤波算法,对混有高斯和椒盐噪声的图像进行去噪处理.该算法首先对受椒盐噪声污染的像素点,采用自适应中值滤波算法进行去除;然后利用高阶统计量针对高斯噪声的不敏感特性,对受高斯噪声污染的像素点,采用其周围梯度和最小的几个点的灰度平均值来代替其灰度值去除噪声.实验结果表明,该算法能够在去除高斯和椒盐噪声的同时,保留更多的图像细节信息,特别是对感染较大噪声的图像有更好的去噪效果.

A method of mixed filter for smoothing of image

In this paper,a new mixture filter method is proposed.The proposed method first classifies the pixels into two classes,one is the pixels which are corrupted by Gaussian noise and the other is the pixels which are corrupted by impulse noise.Using the property that high order statistics are not sensitive to Gaussian noise to processing gaussian noise,and can mostly save the image details which greatly dwindling noise.In the experiment,this mixed method can deal with the images which have Gaussian noise and salt-pepper noise effectively,and it saves more details,it has better effect especially for the images have more noise.