一种结合图像增强的高斯噪声组合滤波算法

摘要

In order to filter the Gaussian noise effectively in digital image,combing with image filtering algorithm and enhancement technique,a new hybrid filtering algorithm based on image enhancement is put forward.Firstly,according to the defect of classical median filter cannot deal with the Gaussian noise,the filter template of the median filter are improved to remove the Gaussian noise with multi-angle,multistage tandem filtering,so as to eliminate the influence of part of the Gaussian noise.Secondly,the average filter is introduced to the wavelet domain,the residual noise in the image is removed step by step with the wavelet domain average filter.Then,a new wavelet domain enhancement function is raised,setting the threshold to divide the filtering image into several areas,se as to do enhancement processing with different degrees with these areas.Finally,quantitative analysis is done with the different intensity Gaussian noise between the factor values of the enhancement function,the function expression between them is given.The experiment results show that the filtering algorithm in this paper has the better filtering effect of the high strength Gaussian noise,and has some certain degree of adaptability.%为了有效滤除图像高斯噪声,将滤波算法与增强技术有机结合,提出了一种具有增强效果的组合滤波算法.该算法首先针对经典中值滤波无法有效滤除高斯噪声的缺陷,从滤波模板的角度对其加以改进,对高斯噪声进行多角度、多尺度、多级串联滤波处理,以滤除一部分噪声;然后将均值滤波引入到小波域中,对图像残余噪声实现小波域逐级均值滤波,实现对噪声的基本滤除;最后设计出一种新型小波域增强函数模型,通过设定阈值,将滤波后图像分为若干个区域,分别进行不同程度的增强处理,通过结合具体实验对不同强度下噪声与增强函数系数的取值进行定量分析,给出两者之间的函数关系式.实验证明,该滤波算法对于高强度的高斯噪声有较好的抑制效果,并且具有一定的自适应性.