Research Of Image Segmentation Based On Ant Colony Algorithm

With the rapid development of information technology, people are eager to get to knowthis world in which we live. Through pictures showing information, people understand thisworld. Image segmentation technique is a technique, which people are interested in targest ofpictures in order to separate targets. Whether the good effect of image segmentation or not, itplays a key role in the understanding and analyzing of pictures. Due to the complexity ofdifferent images and different pictures which use camera, so far there is no a unified approachto segment image. Studying on the segmentation methods, there are two main ideas: The firstone is by improving its original algorithms to improve the performance of the originalalgorithm; the second one is to new ideas and new methods introduce in image segmentation.Due to the image segmentation can be seen as combinatorial optimization problems, sothis paper adopts the intelligent algorithm of ant colony algorithm. And this paper firstlyintroduces the ant colony algorithm, in1992, Italy researcher M.Doirgo frist proposed the antcolony algorithm where his paper has written. Ant colony algorithm is applied by TSP, inorder to find out the advantages and disadvantages, and for the shortcoming of ant colonyalgorithm, it sets up an iterative area, in order to reduce local pheromone, prevent the localoptimal solution. Experiment shows that the performance of the improved ant colonyalgorithm is better.Secondly, for these advantages, ant colony algorithm combines with fuzzyC mean clustering. For complex image segmentation, although fuzzy C-means clustering isexcellent, it has computational complexity, the initial parameters need to be set manually andother shortcomings, the fuzzy clustering which is applied to image segmentation is notaccurate. So, ant colony algorithm and fuzzy C mean clustering is combined, using improvedant colony algorithm get the initial cluster numbers and centers, they are used as the initialparameters of fuzzy C-means clustering, using fuzzy C-means clustering makes imagesegmentation. The experimental shows that the improved ant colony clustering algorithm inimage segmentation and denoising is better.Finally, aiming at the shortcoming of ant colonyalgorithm, this paper is proposed to combine with ant colony algorithm and quantumalgorithm. Quantum rotation gate and mutation operation of quantum algorithm can increaseants searching space, in order to escape from locally optimal solution and prevent premature.The quantum ant algorithm is applied to image segmentation, the ants move by calculating therotation angle, mutate operation uses the Pauli-Z gate. The experimental results shows that thequantum ant colony algorithm in image segmentation effect is better.