Salt & Pepper Noise Removal from Digital Color Images Based on Mathematical Morphology and Fuzzy Decision

  • Muhammad Sharif COMSATS Institute of Information Technology Pakistan
  • Isma Irum
  • Mussarat Yasmin
  • Mudassar Raza

Abstract

A robust method to eliminate the noisy pixels from digital color images contaminated often during the acquisition and transmission process has been proposed in this study. A set of fuzzy decision and mathematical morphological dual operators which are well known erosion and dilation is used in the method. With invariant structuring elements noise is propagated throughout the image with morphological dual operations. Sometimes it results in the loss of image details and features. Firstly a set of fuzzy decisions is designed to generate variant structuring elements and morphological operation to stop the propagation of noisy pixels. Secondly a very simple alternate sequence of erosion and dilation with variant structuring elements and median filter is implied. The quantification measure PSNR has been used to demonstrate the proposed technique strength as well as existing methods has also been compared with this proposed technique to prove that the proposed method outperforms by giving splendid results up to 90% of noise densities.
Published
2018-03-24