An illumination compensation algorithm based on anisotropic pulse coup led neural network (anisotropic-PCNN) model is proposed for low and non-uniform illumination images in order to solve the problems of ashing phenomenon and shortage illumination compensation of shadow area in the traditional puls e coupled neural network (PCNN) model.Firstly,the influence of connection w eight matrices W and M′s values on automatic wave array and wave propagation direction is discussed based on analysis of the stat istical properties of images and firing characteristics of PCNN model;Secondly,connection weight matrices W and M are re assigned based on Weickert′s diffusion rate function,and the anisotropic-PCNN model is designed;At last,a simplified non linear mapping function from anisotropic-PCNN model′s outputs to illumination compensated images is put forward.The simulati on results show that the overall and partial dynamic ranges of images are adjusted,the detail information of shadow region is fully shown,and the ashing phenomenon caused by traditional PCNN model is eliminated.It is a universally ap plicable illumination compensation method for low and non-uniform illumination images to some extent.