Denoising Method for Removal of Impulse Noise Present in Images

Abstract

Images  are  often  corrupted  by  impulse  noise  in  the  procedures of image  acquisition  and  transmission.  The  efficient  impulse  noise  detector  is used  to  detect  the  noising  pixels. Once  the  noisy  pixel  is identified, its value  is  replaced  by  a  linear  combination  of  its  original  value  and  median of  its local  window.  Thus, a  filter  is  used  according  to  the signal  and  noise  characteristics  while  preserving  edges in  images. In this project, an extension of a bilateral filter that can additionally remove impulse  noise  and  an  edge  preserving  filter  to  reconstruct  the intensity  values of noisy pixel. The  algorithm  involves  mainly  detection  of  the  noise  and classifying   it  based  on  an  efficient  detector.  Based on the detection, the proposed filter is applied for noise removal. Especially, the proposed method can preserve edges very well while removing impulse noise. Since  our  algorithm  is  algorithmically  simple,  it  is very  suitable  to  be  applied  to  many  real-time  applications. To achieve the goal of low cost, MATLAB tool   is Proposed.