Classical median filtering and mean filtering are often used to filter out impulse noise and Gaussian noise, but if both impulse noise and Gaussian noise exist in the image, effect of the two filtering algorithms turn out to be dissatisfactory. In order to filter out two different kinds of noise simultaneously, a new filtering algorithm for mixed noise is proposed. According to the characteristics of impulse noise and the local energy information of pixel, firstly, the impulse noise is separated, and to be removed by median filtering algorithm. Then the image containing Gaussian noise is denoised by mean filtering algorithm. Test results show that the proposed algorithm can filter out mixed noise effectively and preserve image details very well; it provides an effective way for removing mixed noise in images.%经典的中值滤波和均值滤波常常被分别用来滤除脉冲噪声和高斯噪声,但是当图像同时存在脉冲噪声和高斯噪声时,这两种滤波算法都不能达到最好的滤波效果.为了能同时滤除两种不同性质的噪声,提出了一种新的混合噪声滤波算法.该算法首先根据脉冲噪声的特点和像素的局部能量信息,分离出脉冲噪声并采用中值滤波算法加以去除,然后对含有高斯噪声的图像采用均值滤波算法进行去噪.试验结果表明,该算法在有效滤除混合噪声的同时,能很好地保护图像的细节,从而为去除图像中的混合噪声提供了一种有效的途径.
展开▼
