摘要
由于航天仪器外型结构不规则、难以准确定位,其装配过程主要依赖人工进行定位引导,自动化程度低。因此,提出一种基于机器视觉辅助定位的自动装配引导系统,解决非规则表面装配定位难题。首先,针对不规则装配表面特征,提出基于视觉在线扫描的孔特征提取以及中心点位识别方法,实现非确定性定位下的装配孔位识别以及圆心辨识;然后,搭建了原型样机并开展了模拟卫星舱板零件安装的验证实验。实验结果表明:本文提出的基于机器视觉辅助定位的自动装配引导系统在非确定环境下能保持高精度的定位检测效果,定位精度在X方向最大定位误差为0.242 mm,在Y方向的最大定位误差为0.232 mm,满足航天仪器的自动化装配精度需求。
Due to the irregular structure of aerospace instruments and the difficulty of accurate positioning,the assembly process mainly relies on manual positioning guidance,with relatively low level of automation.Therefore,this paper proposes an automatic assembly guidance system based on machine vision-assisted positioning to solve the positioning problem of assembling irregular surfaces.Firstly,a method of hole feature extraction and center point identification based on vision online scanning for irregular assembly surface features is proposed to realize the identification of assembly holes and circle center recognition under non-deterministic positioning.Then,a prototype is built and a validation experiment is carried out to simulate the installation of satellite hatch parts.The experimental results show that the proposed automatic assembly guidance system based on machine vision-assisted positioning can maintain high accuracy in the non-deterministic environment,and the maximum positioning error is 0.242 mm in the Xdirection and 0.232 mm in the Y-direction,meeting the automated assembly accuracy requirements of aerospace instruments.
作者
田森
莫威
刘顺
张继昌
时轮
金永乔
TIAN Sen;MO Wei;LIU Shun;ZHANG Jichang;SHI Lun;JIN Yongqiao(School of Mechanical Engineering,Shanghai Jiao Tong University,Shanghai 200240,China;Shanghai Smartstate Technology Co.,Ltd.,Shanghai 201306,China;Shanghai Hanghe Intelligent Technology Co.,Ltd.,Shanghai 200240,China)
出处
《机械设计与研究》
CSCD
北大核心
2022年第3期122-126,共5页
Machine Design And Research
基金
上海市人工智能创新发展专项(2019-RGZN-01026)
上海市产业协同创新项目(JMRHCY-2020-037)。
关键词
航天仪器
不规则表面结构
非确定性定位
装配孔位识别
自动化装配
aerospace instruments
irregular surface structure
non-deterministic positioning
assembly hole recognition
automated assembly
