A Median Filter Based on Judging Impulse Noise by Statistic and Adaptive Threshold

In order to improve the performances of median filter, a median filter based on judging impulse noise by statistic and adaptive threshold is proposed. This method includes two steps of judging impulse noise pixels and restoring it. First judging impulse noise pixels and signal pixels of edge and smooth area in the corrupted image, then median filter is used for the pixels corrupted by impulse noise and signal pixels of edge is preserved to avoid making the image fuzzy. The experimental results show that the algorithm can denoise impulse noise effectively and preserve image detail very well. Through comparing image fuzzy degree, the peak signal-to-noise ratio and the mean absolute error, the algorithm has better restoration performances in getting rid of impulse noise.