An Outline of Various Non-Linear Median Filters for Impulse Noise Removal in Images

Images are often contaminated by noise and degradations of quality of digital data due to flawed exchange mean of communication or noisy channels or flawed camera in the transmission of images over channels. Removal of noise is one of the most vitally necessary and exigent tasks in signal and image processing. Digital images are degraded by impulse noise which has two models namely, random valued impulse noise and salt & pepper noise. In imagery, high frequency components carry precious information such as edges and fine points of imagery. Linear filters are not able to preserve edges in images corrupted by non-Gaussian noise. Non linear filters are thus used for the removal of non-Gaussian noise with edge and fine detail preservation. The paper focuses on testing the functioning of various non-linear median filters for the removal of impulse noise that focused around past studies & compared using Peak Signal to Noise Ratio (PSNR). The non-linear filters were able to identify corrupted pixels from noisy image and then filter those noisy pixels. The outcomes of filters is tested on gray scale images that are degraded with variable percentage of salt and pepper noise and random valued impulse noise.