Impulse Noise Detection and Filtering Algorithm Based on Leap Points of Rank-ordered Pixels
Wang Wei · Computer Engineering and Applications Journal · 2006
This paper proposes a novel algorithm for detecting noise pixels in an image corrupted by impulse noise.Firstly,the pixels in the filter-window are rank-ordered according to their magnitudes.Then the two most significant leap points are found.If they are close to the minimum or maximum value and exceed the extent concluded by the author,we can judge the center pixel as noise pixel if its magnitude is outside of the range confined by the leap points.At last,we restore the noise pixels on the basis of the signal pixels.The experimental results indicate that our algorithm can detect pixels contaminated by impulse noise accurately,and achieves better de-noising effects.