A method of color image edge extraction based on Manhattan distance map

Abstract

Edge extraction is an important topic of image processing, and it is also the basis for subsequent processing. For this reason, this paper proposed a method of color image edge extraction based on Manhattan distance map. Firstly, Manhattan distance map is constructed by adjusting the fuzzy radius and the standard deviation of Gauss nuclear. Edge of the image will be enhanced through the Manhattan distance map, and the gray value of background will be reduced. Secondly, Edge of the area is extracted by analysis the gray distribution features of distance map through threshold, and edge region is constructed by the area of edge as foundation. Because the edge region has continuous characteristic, it can be used as the initial distance matrix for GAC model. Finally, edges will be extracted by iterative GAC model in the original image. The results show the method has a better accuracy and practicability.