Restoration of Images for Highly Corrupted Images by Salt and Pepper Noise Based on Multilevel Median
SU Feng-yan · Computer and Information Technology · 2008
A method based on multilevel median filter is presented for noise removal of an highly corrupted images by salt and pepper noise.At first,the method of multilevel median filter is applied,then the method of noise identify is implemented to classify each pixel to be signal pixel and noise pixel,and the noise labeling matrix is built.Based on the matrix,the noise pixel is removed,and the signal pixel is hold on.The experimental results show that the method has better behavior in noise removing and detail-preserving for the corrupted image noise intensity of which is more than 40%,especially for the highly corrupted images by salt and pepper noise.