Infrared image edge detection applied research based on improved ant colony algorithm

Abstract

The paper based on the analysis of ant colony algorithm and any improved algorithm, We change the initial position of the ant placed from randomly into placed near the edge points and define bounded domain node gradient values as the heuristic information,and also adaptive computing method is used to reach the final threshold.Meanwhile histogram equalization method for image preprocessing is introduced.Simulation experiment results show that this algorithm can better continuously detected edges and test results are efficient and accurate. This algorithm provides a theoretical basis for the more effective detection in infrared image edge character. Introduction Ant Colony Optimization (ACO) is a kind of distributed intelligence simulation algorithm, the basic idea is to imitate ants which rely on pheromones to communicate and show the social behavior[1].That is to use biological information as the basis of ants to select subsequent behavior and through the cooperation and interaction between ants to complete global optimization search process. Nezamabadi pour is the first successful person who realized image edge detection based on ant colony algorithm in 2005[2].Then many scholars put forward many improved algorithm, such as the improvement based on stimulating factor, combine with other swarm intelligence algorithm or merge with classic image edge detection technology.Literature [3] proposed the use of gradient image gray value change as the heuristic information of edge detection and join the movement factor change quantity, and change quantity inspire ant colony move to the edge .Literature [4] proposed to integrate the traditional Canny edge detection operator with ant colony algorithm, different pixels proportion and image angular points which Canny operator get as a priori knowledge of edge points,through the calculation of ant colony algorithm to realize image edge extraction.Literature [5] put forward the idea of using genetic algorithm to improve the ant colony algorithm, and variation factors which can be adjusted with algorithms process were introduced in the algorithm , and so on. In this paper, the histogram equalization method is applied to preprocess images to enhance the contrast of images, then a kind of improved ant colony algorithm is proposed according to the characteristics of infrared image and applied it to the infrared image edge detection. ACO algorithm and its improvements Improved ACO (Improve Ant Colony Optimization, IACO) Initialize improvements: ant’s initial position is amended and place it from random into near edge, the specific approach is to take the image gray gradient threshold to optimize this process. Inspired pheromone improvements: Neighborhood node gradient values will be defined as heuristic information, then we can get the calculating formula of heuristic information as