一种高斯噪声组合滤波方法

为了有效滤除高斯噪声,提出了一种组合滤波方法.该方法首先通过定义新的维纳滤波模板进行预处理,以滤除一部分噪声;然后将图像进行二维小波分解,保留低频成分,对高频成分根据其噪声分布特征设计出新的形态学滤波模板分别进行滤波,并进行小波系数重构;最后通过设计一种新的小波增强函数,以提高图像的清晰度,最大限度保留图像细节信息.实验证明该方法滤波效果优于维纳滤波和形态学滤波,是一种较为实用的高斯噪声滤除方法.

A Mixed Filter Method on Gaussian Noise

In order to filter Gaussian noise effectively,a kind of combination filter method was put forward.Firstly,new wiener filter templates were defined by the filter method so as to filter the part of the Gaussian noise;Then the image was conducted two-dimensional wavelet decomposition,keeping the low frequency components,according to the characteristics of the noise distribution.New morphological filter templates were designed,and the wavelets coefficients were reconstructed;Finally,a new wavelet enhancement function was designed to improve the clarity of the image,and reserve the detail information of image.Experiment results show that the filtering method is better than that of the wiener filtering and morphological filtering.It is a more practical Gaussian noise denoising method.