Improved Ant Colony Optimization Based Low-Contrast Edge Detection

An algorithm for edge detection with low contrast image using improved ant colony optimization is proposed in this paper. In edge detection, some parts of the edge are disappeared or disconnected because of low contrast in their intensities. This paper concentrates to this low-contrast problem. To detect these pixels, the direction tendency of edge lines is computed and ant colony optimization is used to connect these lines. Canny edge detector is utilized to initialize the edge lines and the endpoints of the edge lines are used as starting points for ant colony optimization. Each ant in optimization process starts at the endpoints with the starting direction and moves on low contrast edge pixels. The results show that the low contrast edges are detected. Moreover, the proposed method takes less computational time than the conventional ant colony optimization-based method.