Research on digital image edge detection with local entropy and fuzzy entropy algorithms

A local entropy algorithm and a fuzzy entropy algorithm for image edge detection are proposed respectively in this paper. The precise quantitative relationship between the entropy and the amount of information is utilized in the local entropy algorithm. Because the local window has the filtering characteristics and information extraction characteristics, thus it can be used to process the image with the additive noise without pre-filtering. The fuzzy entropy algorithm gives full consideration to the direction characteristics and the structural characteristics of the edge pixels. Because the gray distribution of neighborhood is orderly and directional, and the gray mutation is structural. Some features are constructed based on fuzzy entropy and then used to detect the image edge. The results of the fuzzy entropy algorithm are compared with that of the local entropy algorithm and other traditional algorithm by adding different noises. The experimental results to various types of images verify the effectiveness of the proposed algorithms and their wide applications.