一种抑制高斯噪声的加权滤波算法

A weighted filtering algorithm for suppression gaussian noise

提出了一种加权的滤波算法,以3×3滤波窗口为中心,先将滤波窗口划分左上、右上、左下、右下四个2×2大小的子窗口,求出这些窗口的中值像素点,然后求出每个中值像素点与它们的均值差的绝对值,利用这些绝对值的平均值采用归一化方法计算出权值.最后将这些中值与它们对应的权值进行加权运算,结果作为中心点的滤波输出.仿真实验结果证明,本文算法对高斯噪声具有较强的去噪能力,较好地保护了图像的边缘等细节,滤波性能高于传统中值滤波算法和传统均值滤波算法.

A weighted filtering algorithm has been proposcu, ,1 vided the window into four 2 × 2 sub - windows of upper left, upper right, lower left and lower right, found out the median pixel of every sub - window, then calculated the absolutions of difference between these median pix- els and these pixel average, calculated the average of these absolutions, used the average and normalization method to obtain the weights. At last, these weights and median has been weighted, the result as the filtering output of center pixel in the filtering window. Experiments indicate that the proposed algorithm has good filtering performance for Gauss noise images, and protects the image edge detail well, the filtering ability are superior to the traditional median filtering algorithm and mean filtering algorithm.