Ant colony search for edge extraction in noise image

Traditional edge extracting methods are sensitive to image noise, and discontinuities often occur in extracted edges. This paper presents an edge leading ant colony algorithm to suppress the noise for edge extraction in noise image. Firstly, it detects the possible edge points which include the real edge points and the noise points. Then, the information of possible edge points is used as heuristic measure to guide iteratively searches of ants to get local edge points. In each cycle, pheromones on the traversed route of each ant are updated proportional to the length of the route, and the searching routes converge on real edges progressively based on the pheromone updating rule. Finally, real edges can be extracted according to the intensity of pheromones. Compared with traditional ant colony algorithms, the proposed method uses leading information to guide the searching process of the ants, which enhances the intention of the search, and improves the efficiency of the algorithm. Experimental results on noise images show that the method can extract real edges effectively, which keeps the edge details and suppresses the noise at the same time.