基于改进ACS算法的SMT自动光学检测路径规划  

SMT AUTOMATED OPTICAL INSPECTION PATH PLANNING BASED ON IMPROVED ANT COLONY SYSTEM

作　　者：程良伦[1] 赖宇锋[1] 

机构地区：[1]广东工业大学自动化学院,广东广州510006

出　　处：《计算机应用与软件》2010年第9期128-130,183,共4页Computer Applications and Software

基　　金：广东省教育部产学研联合基金项目(2007A090302071)

摘　　要：对采用表面安装技术的印刷电路板进行自动光学检测时,需要对检测路径进行合理规划,以在最短时间完成对PCB上所有检测窗的图像采集。针对该问题提出用ACS算法进行路径规划,并引入遗传机制,采用在每一次迭代前使蚂蚁智能继承上一代优秀巡回路径的策略,有效地实现算法求解效率和运行效率的平衡。实验结果表明,该算法有较快的收敛速度,并且解的质量优于其它路径规划算法,用于解决AOI中的路径规划问题是行之有效的。In order to collect images from all inspection windows on printed circuit boards （ PCB ） in shortest time when executing automated optical inspection （AOI） on PCB with surface-mount technology （SMT）, it is necessary to design a reasonable planning for inspecting paths. This paper proposed to do the path planning with ACS to solve the above issue. Besides, genetic mechanism was introduced to the adoption of such a strategy which enabled the ants to inherit partial paths smartly from the excellent itinerant routes of former iterations before starting current iteration each time. This strategy is helpful to making the balance between solution efficiency and operation efficiency of the algorithm. Experimental results show that the proposed ant colony system has quicker convergence speed and higher solution quality than other algorithms in path planning, and is effective in resolving the path planning issues of AOI.

关 键 词：ACS 遗传 路径规划 AOI SMT 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]