基于机器视觉的苹果自动采摘分拣系统研究

Research on the Automatic Picking and Sorting System of Apples Based on Machine Vision

为提升苹果自动化采摘、分拣效率,设计了一种基于机器视觉的苹果自动采摘分拣系统。为设计该系统,首先基于对采摘机器人机械手和末端执行器结构、机械手避障的分析,提出采摘机器人的苹果采摘流程;然后,搭建了基于实时数据处理的苹果自动分拣平台,该平台由两类传送带、可编程逻辑控制器、个人计算机和带有机器视觉、称重传感器和控制面板单元的封闭舱等组成。为验证所构建苹果自动分拣系统的有效性,对来自3个不同苹果品种的183个苹果进行了试验。试验结果表明,所构建的苹果自动采摘分拣系统对苹果按照颜色、尺寸分拣的准确率分别达到96.17%和92.77%,按重量分拣时对苹果重量的估算误差只有5.44克,同时还可对结痂、污点、腐烂等缺陷区域进行100%准确检测。所设计的苹果自动采摘分拣系统具备较高采摘分拣效率,从而对促进苹果产业的发展具有一定实用价值。

To improve the efficiency of automatic apple picking and sorting,an automatic apple picking and sorting system based on machine vision is designed.Firstly,the apple picking process of the picking robot is proposed based on the analysis of the structure of the picking robot manipulator,the end-effector,and the obstacle avoidance method of the manipulator.Then,an apple automatic sorting system platform based on the real-time data processing is constructed.This platform consists of two types of transporter conveyors,a programmable logic controller,a personal computer and an enclosed pods with sensors of machine vision and weighing,and control panel units.To verify the effectiveness of the designed automatic apple picking and sorting system,183 apples from 3 different apple varieties are tested.The experimental results show that,for the designed automatic picking and sorting system,the accuracy rates of apple sorting by color and size can reach 96.17%and 92.77%respectively.In addition,the error in estimating the weight of apples is only 5.44 grams for the automatic picking and sorting system.What’s more,the automatic sorting system can detect scab,stain and rot regions on apples in 100%accuracy.The designed automatic picking and sorting system has high picking and sorting efficiency,which has certain practical value to promote the development of apple industry.