一种图像脉冲噪声小波域广义滤波算法

提出了一种脉冲噪声滤波算法.首先对噪声图像进行二维小波分解,得到高频和低频子图像；其次对高频子图像序列采用改进自适应加权中值滤波进行处理,以排除水平、垂直、对角方向的噪声;然后对于低频子图像引入基于修正系数的维纳滤波进行处理,并进行小波系数重构;最后设计出一种小波域图像增强模型,通过设置调节系数,将图像分为不同区域分别进行相应比例的对比度拉伸处理,结合实验定量讨论了噪声强度与模型系数的函数关系.实验表明,该滤波算法不仅优于几类单一滤波算法,相对于某些组合滤波算法而言,也具有一定的优势.

A generalized filtering algorithm based on wavelet domain of image impulse noise

In order to filter the digital image impulse noise effectively and improve the visual effect of digital image,combing with image filtering and enhance treatment method,a algorithm of the impulse noise is put forward from the angle of generalized.Firstly,the noise image is conducted two-dimensional wavelet decomposition,obtaining high-frequency and low-frequency sub image.Secondly the high frequency sub image sequence is processed by the improved adaptive weighted median filter so as to eliminate the noise of horizontal,vertical and diagonal direction.Then the low frequency sub image is conducted by the modified coefficient of wiener filtering,and the filtering wavelet coefficient is reconstructed at last.Finally,according to the situation of the distortion of filtering image in a certain degree,a image enhancement model of wavelet domain is designed.The image is divided into different areas,by setting the related adjustment coefficients to stretch the image contrast in different proportion.The function relationship of the model coefficient value and noise intensity is quantitative discussed combined with the specific experiment,the certain general function model is given.The experimental results show that,the performance of the filtering algorithm in this paper is not only better than the median filtering,weighted median filtering and the other single filtering method,but also superior to several combination filtering algorithms.