现代制造工程 ›› 2026, Vol. 549 ›› Issue (6): 135-142.doi: 10.16731/j.cnki.1671-3133.2026.06.014

• 仪器仪表/检测/监控 • 上一篇    下一篇

基于3D机器人视觉的飞机零件制孔位姿修正*

陈家祥1, 闫星合2, 薛雷1, 贺元敏3, 殷传霞2, 张刚2,3   

  1. 1 上海飞机制造有限公司,上海 201324;
    2 无锡黎曼机器人科技有限公司,无锡 214174;
    3 华中科技大学无锡研究院,无锡 214174
  • 收稿日期:2025-06-21 出版日期:2026-06-18 发布日期:2026-07-02
  • 通讯作者: 张刚,博士,高级工程师,主要研究方向为机器人加工、移动机器人测量。E-mail:zgrobot@163.com
  • 作者简介:陈家祥,硕士,工程师,主要研究方向为机器人系统研发及应用。E-mail:chenjiaxiang@comac.cc
  • 基金资助:
    *国家商用飞机制造工程技术研究中心创新基金项目(COMAC-SFGS-2024-509);无锡市产业创新研究院先导技术预研项目(XD23018)

Drilling pose correction for aircraft components based on 3D robotic vision

CHEN Jiaxiang1, YAN Xinghe2, XUE Lei1, HE Yuanmin3, YIN Chuanxia2, ZHANG Gang2,3   

  1. 1 Shanghai Aircraft Manufacturing Company Ltd., Shanghai 201324, China;
    2 Wuxi Riemann Robot Technology Co., Ltd., Wuxi 214174,China;
    3 Wuxi Research Institute of Huazhong University of Science and Technology, Wuxi 214174, China
  • Received:2025-06-21 Online:2026-06-18 Published:2026-07-02

摘要: 随着航空航天、汽车制造等领域对高精度装配要求的不断提升,机器人自动化制孔技术以其高效性与灵活性成为关键工艺环节。而在复杂工况下,机器人制孔精度会被视觉定位偏差、视觉系统拍照局限性及机器人运动学误差等多源误差影响。现有方法大多依赖离线编程或刚性工装被动补偿,难以适应柔性生产场景下小批量、多品种的实时误差修正需求,因此,提出了一种融合在线检测与主动补偿的基于3D机器人视觉的飞机零件制孔位姿修正方法,通过3D机器人视觉反馈闭环控制,基于刚性变换补偿的视觉-工具坐标系统一修正框架,协同优化法向信息,并使用三阶段分层渐进点云配准策略,实现动态工况下的亚毫米级制孔精度保障,满足航空航天制造业、汽车制造业等严苛领域的制孔任务要求。

关键词: 机器人制孔, 刚性变换, 精度补偿, 点云配准

Abstract: With the increasing demand for high-precision assembly in aerospace,automotive manufacturing,and related fields,robotic automated drilling technology has emerged as a critical process due to its efficiency and flexibility. However,in complex working conditions,drilling accuracy is compromised by multi-source errors such as visual positioning deviations,imaging limitations of vision systems,and robotic kinematic errors. Existing methods predominantly rely on offline programming or passive compensation via rigid tooling,which struggle to adapt to real-time error correction requirements in flexible production scenarios characterized by small batches and diverse product types,so a 3D robot vision-based drilling pose correction method for aircraft components that integrates online detection and active compensation was proposed. By establishing a closed-loop control system with 3D robotic visual feedback and implementing a vision-tool coordinate unified correction framework based on rigid transformation compensation,the method cooperatively optimized normal vectors and employed a three-stage hierarchical progressive point cloud registration strategy.This approach ensures sub-millimeter drilling accuracy in dynamic working conditions,effectively addressing the stringent requirements of drilling tasks in aerospace and automotive manufacturing industries.

Key words: robotic drilling, rigid transformation, accuracy compensation, point cloud registration

中图分类号: 

版权所有 © 《现代制造工程》编辑部 
地址:北京市东城区东四块玉南街28号 邮编:100061 电话:010-67126028 电子信箱:2645173083@qq.com
本系统由北京玛格泰克科技发展有限公司设计开发 技术支持:support@magtech.com.cn