A Method Based on Statistical Characteristics for Image Denoising

摘要

The median filter will cause fuzzy problems on edge details of the image when it is used to remove noise of an image. In this paper we discuss an improved median filtering method. Firstly according to the arithmetic mean we preliminarily detect signal points and suspicious noise points in the image, and further distinguish between signal points and noise points by standard deviation. The determined noise points will be processed by median filter and signal points will be retained the original pixel values. This method is used in numerical experiments. The results show that the clearer image may be obtained by the proposed method for image denoising that can prevent fuzzy image. The denoising image will retain edge details of the image and improve PSNR.