Algorithm for de-noising of color images based on median filter

Noise degrades the image quality at a large extent in all kinds of images. Impulse noise is the most common one. Correlation is used for analysis of features of the filtering mask. In this work a new algorithm for de-noising, based on median filter is proposed. It has been implemented for the denoising of grayscale and color images which are corrupted by impulse noise. This median filter based algorithm removes the noisy pixels by median value or with the help of collective study of the processed neighboring pixel values. The problem of increasing window size is also removed. To check the efficiency and the performance of the proposed algorithm it has been compared with various filters and corresponding algorithms. The experiments are performed over a range of noise density, which ranges from 5% to 80% for Lena and Peppers images. Results are compared in terms of PSNR and MSE.