摘要
以西门子S7-1200PLC为控制核心,结合深度学习目标检测算法,构建了一套基于机器视觉的工业零件自动分拣系统。传统工业分拣系统在进行分类与位置信息确定时需要分步进行,采用端到端目标检测算法在分类的同时获得物体的位置坐标。视觉模块将数据位置与类别信息输送至机器人控制系统中为后续机械臂抓取做准备,同时视觉模块判断输送带是否存在零件,将指令输入至PLC控制器中,进而控制输送带以及非标设备的工作状态。实验结果表明,利用目标检测算法最终的分拣准确度为91.1%,抓取速度可以达到1.3 s/个。
This article combines the Siemens S7-1200PLC control core with deep learning object detection algorithm to construct a machine vision based automatic sorting system for industrial parts.Traditional industrial sorting systems require step-by-step classification and location information determination.In this paper,an end-to-end object detection algorithm is used to obtain the object’s position coordinates while classifying.The visual module transmits data position and category information to the robot control system for preparation for subsequent robotic arm grasping.At the same time,the visual module determines whether there are parts on the conveyor belt,inputs instructions to the PLC controller,and controls the working status of the conveyor belt and non-standard equipment.The experimental results show that the final sorting accuracy using the object detection algorithm is 91.1%,and the grasping speed can reach 1.3 s/piece.
作者
丁江涛
王帅
王强
DING Jiang-tao;WANG Shuai;WANG Qiang(Research Center for Integrated Circuits and Embedded Applications,Chizhou College,Chizhou,247000,Anhui;School of Artificial Intelligence Innovation,Ma’anshan College,Ma’anshan,243000,Anhui)
出处
《蚌埠学院学报》
2023年第5期82-87,共6页
Journal of Bengbu University
基金
安徽省高校自然科学研究重点项目(KJ2019ZD63)。
