Image Segmentation Algorithm Based on Improved Ant Colony Algorithm

In the process of image segmentation, the basic ant colony algorithm has some disadvantages, such as long searching time, large amounts of calculation, and rough image segmentation results. This paper proposes an improved ant colony algorithm. Applying different transfer rules and pheromone update strategies to different regions of an image, including background, target, edge and noise, we develop a highly adaptive image segmentation method with high edge detection accuracy and high algorithm implementation efficiency. In the initial stage of image segmentation, we apply the idea of fuzzy clustering, which enables ants to gather quickly to the edge in the background and the target area of the image. In the later stage of image segmentation, we introduce an edge search strategy in the edge area. A following experiment shows that this developed image segmentation method can split the target more quickly and accurately.