【摘要】 针对pL级微量胶液转移分配过程中，移液针重复行程较多、作业效率低下的问题，通过机器视觉和改进后的遗传-蚁群算法精准快速地规划路径，提高微量胶液转移分配的效率。搭建了包括机器视觉系统、胶液分配系统和驱动控制系统在内的胶液转移实验平台。通过对采集图像的既定点位进行模板匹配得到相应位置信息，基于位置信息进行路径规划实验并将此算法与常见算法进行比较，观察其能否精确、快速地规划微量胶液转移分配的路径。实验结果表明：通过机器视觉和改进后的遗传-蚁群算法，移液针转移后形成的胶滴体积达到pL级，胶滴平均直径为68.6μm,且准确落在既定点位上，其工作行程缩短了67.9%,作业时间大幅度降低，可有效提高pL级微量胶液转移分配的工作效率。更多还原

【Abstract】 In the process of pL micro-volume adhesive distribution, the itinerary of the transfer needle is repetitive, and the work is inefficient.To solve this problem, machine vision and improved GA-ACO algorithm were used to plan the path quickly and accurately, and improve the efficiency.The experimental platform of micro-volume adhesive distribution was built which consisted machine vision system, adhesive distribution system and drive control system.The location information of the fixed point on the acquired image was gotten by template matching.The path planning of micro-volume adhesive distribution was completed.Finally, the performance of the designed algorithm was compared with common algorithms.The results show that based on machine vision and improved GA-ACO algorithm, the volume of transfer adhesive reaches the pL level; the average diameter of the adhesive is 68.6 μm, and the adhesive falls exactly at the preset point; the itinerary of the transfer needle is cut by 67.9%,the working time is reduced drastically.It can improve the efficiency of pL micro-volume adhesive distribution significantly.更多还原