首页|基于YOLOv5s算法的制动器防错装检测方法

基于YOLOv5s算法的制动器防错装检测方法

扫码查看
针对制动器零件装配过程中零件易混淆、有错装漏装风险等问题,设计了一种基于YOLOv5s算法的制动器防错装检测方法.在制动器零件装配过程中,通过摄像头监测装配画面,提取视频帧图片输送给检测系统,运用YOLOv5s算法识别图片中的零件,实时检测所装配零件类别,将检测信息储存并判断装配流程是否正确.根据生产线实验结果可知,该检测系统可以有效检测出零件装配顺序,检测错装漏装问题,起到防错装效果,降低装配失误率.
Detection Method of Brakes Error-proof Assembly based on YOLOv5s Algorithm
Aimed at the problems such as the confusion and the risk of mis-assembly of brake parts assembly,it was de-signed that a brake error-proof mis-assembly detection method based on YOLOv5 algorithm.In the assembly process of brake parts,the assembly picture was monitored by the camera,the video frame picture was extracted and was sent to the detection system,and the parts in the picture were identified by the YOLOv5s algorithm,so the categories of assembled parts were detected in real time.The detection information was stored and whether the assembly process was right or not was determined.According to the experimental results of the production line,the detection system could effectively detect the assembly sequence of parts,detect the problem of wrong assembly and missing assembly,which played a role of preven-ting wrong assembly,and reduced the assembly error rate.

target detectionYOLOv5sparts assemblyerror-proof detectioncomputer vision

张欣宇、张楠、王希彦、吴玉、刘鹏、谢斌、张文硕、王延忠

展开 >

北京航空航天大学机械工程与 自动化学院,北京 100191

中国兵器工业新技术推广研究所,北京 100089

重庆铁马变速箱有限公司,重庆 400050

目标检测 YOLOv5s 零件装配 防错检测 计算机视觉

国防基础科研项目

JCKY2021601B006

2024

新技术新工艺
中国兵器工业新技术推广研究所

新技术新工艺

影响因子:0.294
ISSN:1003-5311
年,卷(期):2024.435(3)
  • 13