A Modified Nonlinear Filtering Technique for Removal of High Density Salt and Pepper Noise

In this paper, a very effective nonlinear filtering technique is presented for denoising the image with severe salt & pepper noise. The filter reconstruct a corrupted pixel in three steps: first, it will be replaced by the median of the uncorrupted pixels of its neighbors or average value of them if all of them were noisy. Second, the impulse noises that made in the first step will be found and replaced by salt & pepper noise, third we replace the remaining corrupted pixel by the median or its neighbor’s pixel value. The algorithm takes no action on healthy pixels. The proposed algorithm is more effective in removing the salt & pepper noise and also keeps the image features. This algorithm will be test and evaluated by Lena, Boat and Lotus. The results of the simulation show that the proposed method can eliminate salt and pepper noise of densities up to 80% while keeping the edges and fine details sufficiently. Our algorithm has a better performance on all the experiments and shows much more robustness than other algorithms