Image Edge Detection Based On Improved Antcolony Optimization

Digital images have a great application in the production of all kind of life in our today’s digital life. Image edge information is the basic information in a digital image. This information can be used for the difference between the image’s information, background, noise and so on,provide the most direct image content required for people. This makes edge detection to be an important issue in the field of image processing, image analysis, visual inspection,pattern recognition, etc.And its detection methods originated in the initial of the first two order differential operator template, has developed greatly after continuous improvement, fusion and innovation.Ant colony algorithm is a new kind of search optimization algorithm which according to ants foraging behavior of group collaboration to imitate an algorithm of parallelism, positive feedback and strong robustness etc. The algorithm applied to the TSP problem at the first time and achieved good results, later it had used in the edge information detection and search the global optimal solution of the problem. But in the application of edge detection, we found that the algorithm simulation results can have operation time is too long, computing accuracy is not high, poor resistance to noise faults.In order to solve the problem of the long running time, this paper mainly introduces the Sensory Ant Colony Optimization Algorithm. Sensory Ant Colony Optimization Algorithm is mainly aimed at initial distribution of random distribution of useless search and the long running time in the traditional ant colony algorithm. First step,the Inspected image deal with the grayscale images,select a certain threshold for selective initial distribution According to the characteristics of the edge position variance is larger.It selected a initial point as the starting point according to the continuity of the edge curve,then judge the surrounding neighborhood points,finally we get the The edge of the complete information.In order to solve the problem of the poor noise resistance,the system which easy to fall into locally optimal solution., I proposed an improved ant colony algorithm----edge extraction ant colony algorithm based on gradient. It’s main idea is to first in the initialization of the algorithm of image gray gradient, choose the initial points according to the gray value. With the system startup,we choose the ants moving direction according to the pheromone and heuristic information.Then with the use of maximum minimum ant colony system in mobile, it can be good at deal with system easy to fall into local optimal problems. Finally based on the simulation results of several algorithms,it is analyzed, compared, and their advantages and disadvantages of each algorithm are evaluated.