基于蚁群算法的图像边缘检测研究  

Research of Image Edge Detection based on Ant Colony Algorithm

作　　者：张景虎[1] 边振兴[2] 

机构地区：[1]曲阜师范大学物理工程学院,山东曲阜273165 [2]山东理工职业学院,山东济宁272000

出　　处：《火力与指挥控制》2010年第2期115-118,共4页Fire Control & Command Control

基　　金：陕西省自然科学基金资助项目(2005A12)

摘　　要：将蚁群算法应用于图像分割领域,提出了一种新的基于蚁群算法的图像边缘检测方法。详细阐述了蚁群算法与该方法的基本原理和具体实现过程。为了提高算法效率,进行两处改进,第一将蚂蚁初始位置由随机放置修改为放置在图像边缘附近,可取一图像灰度梯度阈值来实现,第二将信息激素强度和启发式引导函数值均定义为像素点灰度梯度值的函数。大量实验结果证明了该算法能有效地检测出图像边缘,而且具有适应性强、效率高等特点。Ant colony algorithm is applied in image segmentation and a new method of image edge detection based on ant colony algorithm is proposed. The basic principle of ant colony algorithm and the method and concrete realization process are elaborated. In order to improve the efficiency of algorithm, two aspects are modified. First, according to a threshold of gray gradient value the ants are initialized nearby the edge of image instead of random position. Second, both the intensity of pheromone and the visibility of node are defined by the function of gray gradient value on the pixel. Many experiment results prove the algorithm can detect the edge of image effectively with high efficiency and adaptability.

关 键 词：蚁群算法 边缘检测 图像分割 模糊聚类 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]