A method for wide density salt and pepper noise removal

A new method for removal of wide density salt and pepper noise in image is proposed. Every pixel in a noise corrupted image is classified to which it is a noise pixel or noise-free pixel. A noise pixel should be modified by a filter. On the contrary, noise-free pixel is kept unchanged. If noise-free pixel is detected as noise pixel wrongly, the restored image will be blurred. The difference between the central pixel and its neighbors aligned with four directions in local window is used to detect noise. Then the noise pixel is replaced by a histogram weighted mean filtering value. Experimental results show that the proposed method can not only efficiently remove 10%~90% wide density salt-and-pepper noise, but also can well preserve the detail information of an image.