基于蚁群算法的边缘检测技术组合优化  

The combinational optimization for edge detection technology based on ant colony algorithm

作　　者：詹宝容[1] 骆金维 黄炜杰 李杏清[1] ZHAN Bao-rong;LUO Jin-wei;HUANG Wei-jie;LI Xing-qing(Department of Information Engineering,Guangdong Innovative Technical College,Dongguan 523960,China;Department of Electronics Engineering,Jinan University,Guangzhou 510632,China)

机构地区：[1]广东创新科技职业学院信息工程学院,广东东莞523960 [2]暨南大学信息科学技术学院,广东广州510632

出　　处：《电子设计工程》2019年第23期27-31,38,共6页Electronic Design Engineering

基　　金：广东省自然科学基金项目(2017GKTSCX112)

摘　　要：针对图像边缘检测过程中,现有算法存在阈值设定缺乏自适应性、收敛速度较慢、容易陷入局部最优解等问题,通过改进蚁群算法的子块梯度比的求取方式优化算法初始设定,并与扰动因子结合优化蚁群转移规则,再根据蚁群动态情况调整信息素阈值等操作进行组合优化。实验结果表明与现有文献的算法相比,改进后的算法在运算速度上提升2.81%,检测效果上提升12.24%。In order to solve the problems of lacking of adaptability in threshold setting slowness in convergence speed and falling into local optimization solution in existing algorithms of image edge detection,an improved combinational Optimization is proposed.The new algorithm optimizes the initial setting of the ant colony algorithm by improving the sub-block gradient ratio acquisition method,and then introduces it into the disturbance factor to optimize the ant colony transfer rules,finally adjusts pheromone threshold based on ant colony dynamics.The experimental results show that compared with the algorithm in the existing reference,the improved algorithm improves the processing speed by 2.81%and detection effect by 12.24%.

关 键 词：蚁群优化算法 CANNY算法 边缘检测 扰动因子 动态阈值 

分 类 号：TN911.73[电子电信—通信与信息系统]