Image Edge Detection Method Based On Interval Type-2 Fuzzy Sets

Edge detection is one of the classic technical problems in the image processing.It is the basic problem in the early visual theory and algorithm, and an important factor in the success or failure of middle and late visual processing. Edge of image which is the basic feature carries most of the image information. It has a wide field of applications in image recognition, image segmentation, image enhancement and image compression. Edge detection directly affects the quality of subsequent image processing effects, and even determines its success or failure. Therefore, edge detection is crucial.Image is essentially fuzzy and has three reasons following:First, three-dimensional image is projected on the two-dimensional plane, so some information of the image is lost. Second, the definitions of the image border, regional and texture exist ambiguity. Third,the results of image underlying interpretation have ambiguity.But all uncertainties are not random, so only using the probability theory is not appropriate. So using the fuzzy theory to deal with image processing problems has a certain rationality and necessity.Interval type-2 fuzzy set, whose second membership value equals one, is a special type-2 fuzzy set. Compared to the general type-2 fuzzy set, it is relatively simple. The construction of traditional interval type-2 fuzzy set is using a fuzzy factor which changes the membership function of type-1 fuzzy set to the upper and lower membership function. HM method is introducted in this article, which is that the image corresponding to the fuzzy matrix operation directly comes into the upper and lower membership function.Fuzzy entropy describes the fuzziness degree of a fuzzy set. Based on fuzzy entropy, the effect of edge detection method is better than the traditional one, so it becomes research focus of image analysis. Currently, Burillo Bustince and Zeng proposed interval type-2 fuzzy entropy have some defects, for this reason the current interval type-2 fuzzy entropy are improved.The aim of this article is to deal with image edge detection, by HM the interval type-2 fuzzy sets and improved interval type-2 fuzzy entropy. In the end, using MATLAB do simulation experiments which compared with the previous conclusions, and obtaining good results.