基于视觉识别的自动收获分拣机器人设计与实验
Design and experiment of an automated harvesting and sorting robot based on visual recognition
荆旭 1何智宏 1沈行健 1赵智鹏 1丁栩彬 1邢健强1
作者信息
- 1. 江苏大学 京江学院,江苏 镇江 212100
- 折叠
摘要
设计了一种自动化球型模拟果蔬收获机器人,该机器人主要由轮式可移动底盘、STM32 主控系统、超声波避障系统、陀螺仪姿态调整系统、视觉识别系统和收获分拣机构等组成.采用C语言编写自动控制模块,Python结合OpenMV编写视觉识别控制模块.样机测试结果显示:该机能完成指定路线运动,规避果园垄等障碍,正确识别应收果实,按分拣条件进行正确分拣,可自主完成果实采摘和收集功能.混合光照条件下果实收获成功率为 85.4%,果实分拣成功率为 88%,收获总时间为 1′30″,通过 5 条垄,收获并分拣果实 18 个,验证了本系统的实用性和可靠性.
Abstract
An automatic fruit harvesting and sorting robot was designed for simulated spherical fruits and vegetables.The robot was mainly composed of wheeled mobile chassis,STM32 main control system,ultrasonic obstacle avoidance system,gyroscope attitude adjustment system,visual recognition system and harvest sorting equipment.The auto control section was written in C and the visual recognition section in Python and OpenMV.The test results showed that the robot could complete the designated route movement,avoid obstacles such as orchard ridges,correctly identify the receivable fruits,sort them correctly according to color sorting conditions,and independently complete fruit picking and collection.Under mixed lighting conditions,the success rate of fruit harvesting was 85.4%,and the success rate of fruit sorting was 88%,the total harvesting time should not exceed 1'30",passing through five ridges,simultaneously harvested and sorted 18 fruits,which verified the practicality and reliability of this system.
关键词
果蔬收获/机器人/视觉识别/收获分拣机构/自动化控制Key words
fruit and vegetable harvest/robot/visual recognition/harvest sorting agency/automated control引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据