基于邻域均值的去椒盐噪声算法

Salt and Pepper Noise Removal Algorithm Based on Neighbourhood Mean

为了改善图像效果,利用图像邻域相关性提出了一种适用于椒盐噪声的图像去噪滤波算法。首先利用最大最小法则检测出被椒盐噪声污染的像素点,然后将被污染像素点邻域中的8个像素点按距离远近分为两类,最后利用近距离邻域中未被污染像素灰度值的均值重构图像灰度值。当近距离邻域像素全部被污染时,以远距离邻域中未被污染像素灰度值的均值代替该点的灰度值。仿真结果表明,该算法具有较大的峰值信噪比,能有效地抑制椒盐噪声并保护图像的细节。

To improve the image effect,a filter algorithm for image noise removal suitable for salt and pepper noise is proposed based on the correlation of the neighbourhood.The pixels contaminated by salt and pepper noise are detected by using the maximum-minimum principle.Eight pixels near the contaminated one are divided into two classes according to their distances.The gray value of contaminated pixel is reconstructed by the mean of those uncontaminated pixels with near distance.The gray value of contaminated pixel is reconstructed by the mean of those uncontaminated pixels with far distance if all the pixels with near distance are contaminated.Computer simulation results show that the proposed algorithm with superior peak signal-to-noise ratio can restrain the noise and preserve the detailed information of images.