Abstract: This paper presents a new kind of image enhancement and restoration based on Fuzzy ART network and data fusion in order to process the images effectually.This approach which uses grayscale histogram threshold can not only automatically determine the number of classes of an image to analyze clusters, but also process the image which comes from the same scene with noises. Simulation experiment demonstrates that FART network can realize the image enhancement and restorat ion very well.

Key words: Fuzzy ART network, data fusion, image enhancement and restoration, cluster analysis