The Research and Application of Improved Ant Colony Algorithm with Multi-thresholds in Edge Detection

The traditional ant colony algorithm has been applied to image edge detection, but the result is not sufficiently ideal, and this paper puts forward an improved ant image edge detection algorithm. The improved algorithm optimizes the selection of the initial pheromone and ant's starting point, pheromone update strategy and ant's search termination condition; considering the edge curve's continuity, the inertia matrix is introduced to the heuristic information, and we set up multiple thresholds to make the edges more continuous, closer to the real edge and effective noise restraining. Compared with the traditional ant colony algorithm and Canny operator, the improved algorithm shows that it is more effective in detecting image edges.