基于YOLOv5目标检测的机械臂精准授粉系统设计与试验OA北大核心CSTPCD

Aiming at the problems of heavy labor intensity,low efficiency,large amount of pollen and high cost of mechanical pollination,this paper proposed a scheme combining YOLO deep target detection algorithm and robotic arm execution system to achieve precise pollination of pear flowers.Firstly,the whole system is divided into depth camera sensing module,target pollination decision module and robotic arm execution module,and the fusion technology route of each module of robotic arm pollination is proposed.Then,through the fusion of depth camera and YOLOv5 target detection algorithm,the three-dimensional coordinate position of the output target of the system is detected.Further,the coordinates in different coordinate systems are unified to the base coordinate of the robot arm through coordinate conversion,and the pixel coordinate value and depth value of the pear flower are converted to the three-dimensional coordinate of the pear flower in the base coordinate system.Finally,the experiment completes the pollination execution of the robot arm and explores the accuracy of the pollination system of the robot arm.The results show the average absolute error of pollination is 3.83 mm.