A Filter Algorithm for Suppressing Mixed Noise Based on Grey Relevance

摘要

In order to suppress the mixed noise composed of salt & pepper noise and Gauss noise in a digital image, a filter algorithm for suppressing mixed noise based on grey relevance is proposed. The algorithm classifies the noise point in the center of the filtering window, calculates the correlation coefficient of non-pepper and salt noise points in window by using the grey relational degree, and performs the weighted operation to replace the salt & pepper noise point. For Gaussian noise point, the algorithm uses the mean of non-pepper and salt noise points in the filtering window to replace. The experiments results show that the algorithm can suppresses the mixed noise in images, and improves the sharpness of the image, which is better than traditional filtering algorithms.