基于蚁群算法的SMT自动光学检测路径规划

SMT automatic optical inspection path planning based on ant colony algorithm

对采用表面安装技术的印刷电路板进行自动光学检测时,检测路径规划属于组合优化问题,每个检测窗口又存在一定的可移动范围。针对以上问题文中提出了首先用蚁群算法来求能覆盖检测对象的最少检测窗数。并确定每窗口的可移动范围;然后再结合逐次逼近来确定检测路线和每个检测窗的准确位置。模拟实验结果表明,该方案求出的路径规划检测窗数量最少,检测路线最短。

Inspection path planning in SMT PCBA automatic optical inspection belongs to combinatorial optimization problem. However, because of many inspection windows tolerable displacement, the path planning is more complex than typical travelling salesman problem. This paper proposed to do the path planning in two steps: at first to determine the least number of inspection windows based on ant colony algorithm, then combined with gradually approaching method to seek the shortest inspection path and to fix every inspection window. Simulating experiments show that this approach can effectively get the least number of inspection windows and find the shortest inspection path among these windows.