Study of Median Filters and Fuzzy Rules for Filtering of Images Corrupted With Mixed Noise

Abstract

Pixels of the digital images has been corrupted by noises came out of faulty communication channel or nonlinear sensors. Such noise models can be either Salt & pepper noise (impulse noise) or Gaussian noise. Sometimes both of the noises corrupt the same image; such combination of two noise model is called as mixed noise. These noises can be removed by using image filtering algorithms. The most popular image filtering algorithm is median filters. Two models of median filters, that are simple median filter and switching median filter has been discussed in this paper. Fuzzy logic controller has been discussed in details so that its combination with median filters can be discussed for future work. Terminologies like fuzzy logic operator, membership functions and fuzzy rules have been discussed. Quality parameter PSNR (Peak Signal to Noise Ratio) has been studied to find the quality of image numerically.