一种新型的图像去噪滤波器

图像滤噪时利用的先验知识越多滤波效果越好，但是一般情况下只能得到一幅被污染的图像，因而无法获得这些信息。针对这点，提出了一种去除图像中椒盐噪声的新型滤波器。该滇波器首先根据椒盐噪声的特点，对滤波窗口中的象素进行一种排除最大和最小的操作，波除椒盐噪声点。其次，根据图像区域特性信息对滤波窗口再进行一次排除次最大和最小的操作。实验表明该滤波器在性能和速度两方面均强于现有的滤波器，特别是当噪声概率超过0．3时，这种优势尤为明显。

A new filter for removing noise of image

The more prescise knowledge used in removing image noise, the better will the effect of filter be. But we can only gain a piece of noise-corrupted image in general situation,so we can not get this information. In the light of this we propose a novel filter for removing salt & pepper-noise in this paper. Firstly, in the light of the character of the salt-noise,the filter operates excluding maximum and minimujm, to pixel in filter window and to remove salt & pepper-noise dot. Secondly,according to region-character information of image, an operation can reduce next-maximum and next-minimum filter window. Experimental results show that the filter gives superiority, compared with now available filter on both performance and speed, especially, the superiority will be more obvious when noise probability exceeds 0 3.