摘要
针对菠萝智能机械化采摘需求,研究一种基于机器视觉的多机械臂菠萝采摘机器人。根据菠萝茎秆较脆的特点,设计一种由V形悬臂、切刀及挡板组成的推剪式末端执行器,实现菠萝果实与茎秆的分离;设计由多个两自由度机械臂组成的机器人,每个机械臂可带动末端执行器在菠萝垄体的高度、长度和宽度方向上移动;为实施智能采摘,应用深度神经网络YOLOv5检测菠萝果实,通过三维点云分析获得菠萝采摘点位置。结果表明:菠萝果实检测精度为97%,检测召回率为95.75%;在实验室环境下菠萝采摘成功率为90%,平均成功采摘时间为5.4 s,验证了该机器人结构的合理性。
A pineapple picking robot with several robot arms based on machine vision is studied with the intention of meeting the demand for intelligent mechanization picking pineapple.A push-shear end-effector made of a V-shaped cantilever,a cutter,and a baffle is created to separate pineapple fruit from stem based on the brittle nature of the pineapple stem;a robot is built with multiple two degrees of freedom arms,each arm has the ability to move the end-effector in the height,length,and width directions of the ridger body;deep neural network YOLOv5 is used to detect pineapple fruit in order to execute intelligent picking,and three-dimensional point cloud analysis is used to find the pineapple picking sites.The results show that the detection accuracy of pineapple fruit is 97%,and the detection recall rate is 95.75%;the success rate of pineapple picking in laboratory environment is 90%,and the average successful picking time is 5.4 s,which proves the rationality of the robot structure.
作者
林桂潮
吴志铭
严茂森
梁仁杰
吴遥禺
严富威
吴天骏
邓广坤
姚佳炎
张有柳
Lin Guichao;Wu Zhiming;Yan Maosen;Liang Renjie;Wu Yaoyu;Yan Fuwei;Wu Tianjun;Deng Guangkun;Yao Jiayan;Zhang Youliu(School of Electro-mechanical Engineering,Zhongkai University of Agriculture and Engineering,Guangzhou 510631,China)
出处
《机电工程技术》
2023年第10期141-144,154,共5页
Mechanical & Electrical Engineering Technology
基金
国家自然科学基金资助项目(32101632)
大学生创新创业训练计划项目(202111347015)。
