Abstract: To overcome the drawbacks of the standard median filter, the adaptive median filter is proposed in this paper. In the algorithm, noise detection is utilized to determine impulse noise and the degree of noise interference in sub-images. Based on the degree of noise interference, window size could be selected, and the noise found in sub-images is removed by means of improved median filter. Results of computer simulation experiments on the adaptive median filter show that the method can suppress noise very effectively while preserving fine details very well and has better filtering performance than the standard median filter. It also shows a great advantage over the standard one in filtering speed.