Adaptive weighted filtering algorithm for salt and pepper noise based on image amalgamation

摘要

Aim at the problem that median filtering algorithm loses the detail information and border information, an adaptive weighted filtering algorithm based on image amalgamation is proposed in this paper. The algorithm firstly detects the impulsive noisy pixels, then the eight pixels around the noisy pixel are divided into four parts. Based on the amalgamation of the four parts, the weights are obtained self-adaptively. At last the noisy pixels are disposed according to the weights. The proposed algorithm are compared with the median filtering algorithm and other improved algorithm, experimental results show that the proposed algorithm has better ability of filtering and protecting the detail information, especially when the image is highly corrupted.