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 amountsof calculation, and roughimage segmentation results. This paper proposes an improvedant 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 methodwith high edge detection accuracy and high algorithm implementation efficiency. In the initialstage ofimage segmentation, we apply the idea of fuzzy clustering, which enablesants to gatherquicklyto the edge in the background and the target area of the image. Inthe later stage of image segmentation, we introduce an edge search strategy in the edge area. A following experiment showsthat this developedimage segmentation method can split the target more quickly and accurately. © 2014 SERSC.