Comparison of processing results of median filter and mean filter on Gaussian noise

Abstract

In people's daily lives, if the image is polluted with noise, it will become very blurred, so it is pretty necessary to use filtering to denoise the image and get a clearer image to assist people in work and study. The purpose of this article is by comparing the denoising result of the Median filter and Mean filter on Gaussian noise, and the filtering method which is more suitable for reducing Gaussian noise is found. Matlab is a very accurate and reliable scientific calculation standard software tool, which is very common in people's lives. The Median filter is a kind of nonlinear filter. It is so virtual at decreasing impulse noise. The common basic theory of the Median filter is to supersede the gray value of pixels with the median of the gray value in a neighborhood of the pixels, and not use the average proportion. The Mean filter is a sort of plain sliding-window space filter that takes the place of the central value with the mean of all the pixel values in the plain window. This essay takes advantage of Matlab to complete the process of adding Gaussian noise in photos, reducing Gaussian noise, and calculating the PSNR. The processing result of two types of noise reducers on Gaussian noise is compared. And by observing the image clarity, PSNR and MSE evaluation techniques to find which is a better filter to decrease Gaussian noise.