Abstract: In response to the flaw that the median filtering algorithm has a poor handling capacity to high-density and fine texture noise, noise detection-based adaptive median filter algorithmis proposed. According to the associated level between noise-point information and the surrounding, the new algorithm adjust the Noise point filter value, which can get a better deal with the details of the images. The adaptive strategies strengthen the performance of the filtering de-noising algorithm, which is good at dealing the high-density noise. Through simulation analysis, the new algorithm is better than other algorithms.