An Improved Edge Detection Algorithm Based on Area Morphology and Maximum Entropy

A novel method based on area morphology and maximum entropy theory for edge detection is presented. Area morphological filters can enable the elimination of connected components within image level sets by selecting different area scales. Using area morphology approach, thin, closed contours without boundary distortion can be provided, which are suitable for use in image segmentation. By the introduction of maximum entropy, the improved method not only preserves the advantages of the area morphology, but also can automatically highlight desired objects against background without user input or parameter specification. The experimental results show the improved method can obtain good results by selecting optimal image scales automatically.