Abstract

Preprocessing an image is an important element of image processing. When images are corrupted with fixed-value impulse noise or random-value impulse noise, it is essential to restore the quality of the degraded image using linear or non-linear filters, in order to make them suitable for subsequent processing. It has been proved that non-linear filters are effective in suppressing or eliminating fixed-value impulse noise [1]. Moreover, non-linear filters preserve the details and edges of an image during the process of denoising [2]. In this paper we propose, two non-linear filters Adaptive Iterative Median Filter with threshold (AIMF) and Adaptive Iterative Rank-ordered Filter with threshold (AIRF) to restore the images corrupted with fixed-value impulse noise. These filters are based on the principle of order statistics. Simulated results show that the proposed filters perform much better than many other existing median-based filters and are found to give comparable results as that of JM filter based on Jarque-Bera test [20]. Moreover, the proposed filters are computationally simpler.

 

Keywords: Impulse noise, fixed-value impulse noise, salt and pepper noise, noise filters, highly corrupted image, noise removal, median filter, rank-ordered mean (ROM) filter..

 

(P1150849520, 877 KB)