Fuzzy Object Segmentation Based on Maximal Fuzzy Exponential Entropy

Because of the fuzzy infrared object,over-segmentation or less-segmentation is shown.Combining the theo- ries of fuzzy sets and entropy,a new segmentation method of maximal fuzzy exponential entropy is proposed.The orig- inal image is segmented into dark and bright fuzzy sets based on pixel gray value.And based on the rule of fuzzy expo- nential entropy,the fuzzy scope is established.To look for the best segmentation threshold,genetic algorithm is used to search optimal combination.Contrast experiments show that segmentation results are better than OSTU’S to fuzzy object.While rand noise is added,over segmentation is found in two methods.