A new method of removing salt-and-pepper noise basing on grey system model in images

Abstract

Focusing on the problem of image details losing while denoising image using Median filter method, the article proposes a salt and pepper noise filtering algorithm basing on grey system theory. The algorithm employs the idea of sorting out current pixel firstly and then filtering, where all the pixels are classified into two groups, according to the noise characteristics, one is suspicious noise and the other is signal pixel. For suspicious noises, the grey system model is adopted to filter noises, and there are nothing to be done for signal pixels in order to preserve more image details. The results of experimental show that the proposed method not only achieved better noise reduction properties, but also has strong ability to preserve details compared with the commonly used median filter algorithm.