关键词: 自适应中值滤波, 脉冲耦合神经网络, 椒盐噪声

Abstract: An adaptive median filter algorithm is proposed to slove the problem of the simplified Pulse Coupled Neural Network（PCNN） model in image filtering.At first，the simplified model is proved to fail to detect pepper noise using reductio Ad absurdum，then the model is improved by using the method of dividing and rule，finally the adaptive median filter algorithm is acheieved by detecting the pollution level，ascertaining the specific locations of noise points and determining the size of the median filtering window adaptively.Experimental results show that this algorithm improves the accuracy of noise detection and the fidelity of image filtering，and has a better performance on different noise densities.

Key words: adaptive median filter, Pulse Coupled Neural Network（PCNN）, salt and pepper noise