Extreme Values Filtering Algorithm of Secondary Salt-and-Pepper Noise Detection

In order to accurately detect and filter salt-and-pepper noise in an image,an algorithm of secondary noise detection and filtering based on extreme values is proposed.The algorithm sets the range of the noise point according to the salt-and-pepper noise characteristics,and separates the quasi-noise and signal points.To identify the signal points misjudged as quasi-noise,noise detection based on extreme values is applied in eight texture directions.From the frequency of the center pixel being judged as an extreme,it can be classified as signal or noise,and the noise pixel is substituted by the median of the signal in the filter window.Experimental results show that the algorithm can effectively suppress salt-and-pepper noise.