Based On Ant Colony Algorithm For Image Segmentation Method

Image segmentation is the primary task for image analysis and understanding. In recent years, some bionic optimization algorithms, such as neural network, genetic algorithm and ant colony algorithm, have merged with other algorithm and applied into image segmentation. Practice has shown that ant colony algorithm could be used to solve some complicated and large data. While image segmentation can be seen as clustering of large data, the application of ant colony algorithm in image segmentation has an important practical value.In this paper, the main contributions are described as follows:The thesis background, image segmentation and some edge detection methods are introduced briefly.The principle of ant colony algorithm is studied for further step, ACA and CPACA are studied in order to solve TSP problem.Color image segmentation are introduced in detail, the color space and clustering algorithm are given a detail description.Based on the above study and the solved specific problems of edge detection and color image segmentation, new application model of ant colony algorithm is proposed and the effectiveness of the algorithm is tested by experiments.The new idea lies in (1) To solve the problem of edge detection, a new definition of pheromone based on the gradient operator for image edge extraction makes it more accurately to extract the effective edge; (2) For the problem of color image segmentation, ACA is used to intelligently optimize the initial cluster centers. By analyzing the influence of new parameters on algorithm performance, the right parameters are chosen to obtain the more accurate initial classification of pixels, and then FCM Algorithms will be used to get better segmentation results.