Research On Improved Ant Colony Algorithm In Image Edge Detection

Image border is one of the essential features of digital image and it is key step in the course of knowing,analyzing and processing the image. Overthe past several decades, a great number of methods were put forward on edge detection. These methods have their different characters but detection effect is not ideal duo to great differences on image itself, so there commonly existed the problems of versatility of these methods.Consequently, it still remains hot and difficult problem in the fields of digital image processing and computer vision and it is very important to study new edge detection method with new techniques and methods.Colony algorithm is a kind of intelligent bionic optimization algorithm; the algorithm has robustness, parallelism,distribution,global optimization,and can easily combine with other optimization algorithms. It can find the optimal solution of the problem in a relatively short period of time. Therefore,colony algorithm is applied to image edge detection as well as the experiments show. Besides colony algorithm can complete all kinds of image edge detection. However,the effect of image edge detection is not good and there-exist a number of problems.Basic ant colony algorithm has a strong commonalityapplied toedge detection, but it has characteristics of slow convergence and easily results in stagnation behavior. So the detection effects are inaccuracy and unsatisfactory and even needs multiple loop iteration and repetitive computation. As a consequence, one type of ant colony algorithm of image edge detection was proposed with the perceptive function and variable step size ability. In the algorithm, firstly an edge point is wanted to random detect, then the point was taken as the starting point to perceive other edge points with functions of perception,finally integrated and continuousedge will be found.During the course ofhunting foredge points, variable step length strategy was proposed in order to improve search efficiency and which can perceivenearbynodes according to the actual conditions, that is, small step can be employed if there are edge points or large step is employed in the algorithm onthe contrary. The ant search will be more flexible and accelerate the process for edge-searching. The results of simulation experiments showed that the algorithm can effectively cut down irrelevant mobile computing, enhance the effects and shorten the time when the ants conducted image edge search.In order to solve the edge detection of ant colony image,a large number of ants will carry on many independent mobile in the non edge area of the image. In this paper, an ant colony algorithm for image edge detection is proposed, which has the ability of detecting and sensing. The first stage of the algorithm is the reconnaissance ants to carry out reconnaissance,to remove a large number of non edge region, and to release the information element. The second stage is edge ants search process, according to the pheromone concentration, a number of edges have been ruled out at the same time, search efficiency of the edge ants is improved obviously, simulation results show that the speed of the algorithm is faster and the efficiency is better.