High-precision Industrial Robot Teaching Method Based on 6D Light Pens

doi:10.3969/j.issn.1004-132X.2025.11.029

Abstract

Abstract:

Aiming at the problems of poor precision， sensitivity to light and restricted teaching range in current industrial robot vision teaching methods， a high-precision industrial robot teaching method was proposed based on a 6D light pen. The light pen was equipped with three infrared-emitting marker balls， and infrared binocular camera mounted on the robot's end-effector captured the marker balls to obtain their 3D position coordinates and 3D pose information. Two measures were implemented to enhance the vision teaching precision of the 6D light pens. Firstly， a hand-eye calibration method was proposed based on infrared light-emitting marker balls， which significantly improved the transformation precision from the camera coordinate system to the robot tool coordinate system. Secondly， an adaptive camera tracking method was proposed， allowing the robot to automatically track the position of the light pen through the camera mounted on the end-effector， ensuring that the marker balls remain centered in the camera's field of view， thus effectively improving the teaching precision. Finally， the captured teaching trajectory was transformed into the robot base coordinate system to achieve trajectory programming and tracking. Experimental results show that the maximum single-point error of the light pens is as 0.63 mm， the maximum pose error is as 2.1432°， and the maximum trajectory error is as 0.73 mm. The proposed teaching method may achieve high-precision and high-efficiency teaching for robots.

Key words: robot teaching, 6D light pen, infrared binocular camera, hand-eye calibration

摘要：

针对目前工业机器人视觉示教方法存在精度较低、对光照敏感和示教范围易受限等问题，提出一种基于6D光笔的工业机器人高精度示教方法。光笔设有三个红外发光标记球，通过固定于机器人末端的红外双目相机捕捉光笔标记球即可获得其3D位置坐标和3D姿态信息。采取了两种措施提高6D光笔的视觉示教精度，一是提出一种基于红外发光标记球的手眼标定方法，有效提高了相机坐标系到机器人工具坐标系的转换精度；二是提出了一种相机视野自适应跟踪方法，使机器人能够通过固定在末端的相机自动跟踪光笔的位置，从而保证光笔标记球始终处于相机视野中心区域，有效提高了示教精度。最后将捕捉到的示教轨迹转换到机器人基坐标系下实现轨迹的编程跟踪。实验结果表明，利用该光笔得到的单点最大误差为0.63 mm，姿态最大误差为2.1432°，示教轨迹最大误差为0.73 mm，所提出的示教方法能够实现机器人高精度、高效率示教。

关键词: 机器人示教, 6D光笔, 红外双目相机, 手眼标定

CLC Number:

TP242

Jihuang LIANG, Weifeng WANG, Haibin WU. High-precision Industrial Robot Teaching Method Based on 6D Light Pens[J]. China Mechanical Engineering, 2025, 36(11): 2710-2719.

梁继煌, 汪炜锋, 吴海彬. 基于6D光笔的工业机器人高精度示教方法[J]. 中国机械工程, 2025, 36(11): 2710-2719.

Figures/Tables 21

Fig.1 Industrial robot teaching system based on 6D light pen

Fig.2 Flowchart of vision-based teaching method

Fig.3 Subpixel circle center extraction algorithm

Fig.4 Binocular stereo vision measurement principle

Fig.5 System coordinate system transformation relationship

Fig.6 Light pen probe calibration

Fig.7 Adaptive tracking effect of the view

Fig.8 Teaching experiment platform

Tab.1 Vision system parameters

相机型号	MV-CE060-10UM
基线/mm	130
相机焦距/mm	12
分辨率	3072×2048
工作距离/mm	600
双目视野区域/（mm×mm）	299×211

Fig.9 Hand-eye calibration comparison experiment

Tab.2 Results of hand-eye calibration experiment

标定方法	最大误差/mm	平均误差/mm
本文方法	0.632	0.508
棋盘格标定法	1.068	0.960

Fig.10 Tracking teaching comparison experiment

Tab.3 Results of tracking teaching experiment

示教方法	最大误差/mm	平均误差/mm
跟踪示教	0.601	0.495
固定示教	0.896	0.685

Fig.11 Attitude measurement experiment

Tab.4 Results of attitude measurement

组数	$Δ E x$ /（°）	$Δ E y$ /（°）	$Δ E z$ /（°）
1	1.0545	2.1319	1.9704
2	1.9295	2.0772	1.6830
3	1.5376	1.2669	2.0890
4	2.0401	2.0154	1.9028
5	2.0536	1.9057	1.2350
6	1.7436	2.1306	1.7901
7	2.1239	1.7656	2.0860
8	1.7591	2.1432	1.8051

Tab.4 Results of attitude measurement

组数	$Δ E x$ /（°）	$Δ E y$ /（°）	$Δ E z$ /（°）
1	1.0545	2.1319	1.9704
2	1.9295	2.0772	1.6830
3	1.5376	1.2669	2.0890
4	2.0401	2.0154	1.9028
5	2.0536	1.9057	1.2350
6	1.7436	2.1306	1.7901
7	2.1239	1.7656	2.0860
8	1.7591	2.1432	1.8051

Fig.12 Dynamic teaching of saddle-shaped trajectory

Fig.13 3D trajectory comparison

Fig.14 Robot trajectory reproduction

Fig.15 Comparison of experiments under different light environments

Tab.5 Comparison of errors under different lighting environments

光照强度	最大误差/mm	平均误差/mm
正常光照（67 lx）	0.734	0.618
弱（153 lx）	0.715	0.634
中（376 lx）	0.748	0.613
强（742 lx）	0.704	0.638

Tab.6 Precision comparison of this paper's method with existing methods

示教方法	示教精度/mm	是否受可见光影响	允许的示教姿态范围
本文方法	0.618	否	较大
棋盘格^［7］	1.18	是	较小
多面体^［9］	0.8	是	大
发光点^［10］	1.3	是	中
球杆型^［11］	2.94	是	大

References 17

[1]	吴超群，赵松，雷艇. 曲线焊缝的机器人焊接轨迹规划与高频控制［J］. 中国机械工程， 2023， 34（14）： 1723-1728.
	WU Chaoqun， ZHAO Song， LEI Ting. Robot Welding Trajectory Planning and High Frequency Control for Curved Seams ［J］. China Mechanical Engineering， 2023， 34（14）： 1723-1728.
[2]	葛吉民，邓朝晖，王水仙，等. 基于点云的机器人焊缝自动化磨削系统与方法［J］. 中国机械工程， 2024， 35（7）： 1253-1262.
	GE Jimin， ZHENG Chaohui， WANG Shuixian， et al. Automated Grinding System and Method for Robotic Weld Seams Based on Point Cloud［J］. China Mechanical Engineering， 2024， 35（7）：1253-1262.
[3]	曾超，杨辰光，李强，等. 人-机器人技能传递研究进展［J］. 自动化学报， 2019， 45（10）： 1813-1828.
	ZENG Chao， YANG Chenguang， LI Qiang， et al. Research Progress on Human-Robot Skill Transfer ［J］. Acta Automatica Sinica， 2019， 45（10）： 1813-1828.
[4]	CHANG H T， CHANG J Y. Sensor Glove Based on Novel Inertial Sensor Fusion Control Algorithm for 3-D Real-time Hand Gestures Measurements ［J］. IEEE Transactions on Industrial Electronics， 2019， 67（1）： 658-666.
[5]	TOGIA T， GKOURNELOS C， ANGELAKIS P， et al. Virtual Reality Environment for Industrial Robot Control and Path Design ［J］. Procedia CIRP， 2021， 100： 133-138.
[6]	PAN Y， CHEN C， LI D， et al. Augmented Reality-based Robot Teleoperation System Using RGB-D Imaging and Attitude Teaching Device［J］. Robotics and Computer-Integrated Manufacturing， 2021， 71： 102167.
[7]	倪自强，王田苗，刘达．基于视觉引导的工业机器人示教编程系统［J］. 北京航空航天大学学报， 2016， 42（3）： 562-568.
	NI Ziqiang， WANG Tianmiao， LIU Da. Vision Guide Based Teaching Programming for Industrial Robot［J］． Journal of Beijing University of Aeronautics and Astronautics， 2016， 42（3）： 562-568.
[8]	应再恩，平雪良，李正洋，等.基于视觉跟踪的机器人复杂轨迹模拟再现［J］.机械设计与研究， 2014， 30（01）： 39-41.
	YING Zaien， PING Xuelian， LI Zhengyang， et al. The Simulation and Reconstruction of the Complex Robot Trajectories Based on Visual Tracking［J］. Mechanical Design and Research， 2014， 30（1）： 39-41.
[9]	LE T S， TRAN Q V， NGUYEN X L， et al. Solpen： an Accurate 6-DOF Positioning Tool for Vision-guided Robotics［J］. Electronics， 2022， 11（4）： 618.
[10]	LIN H I， LIN Y H. A Novel Teaching System for Industrial Robots ［J］. Sensors， 2014， 14（4）： 6012-6031.
[11]	ZHANG H D， LIU S B， LEI Q J， et al. Robot Programming by Demonstration： a Novel System for Robot Trajectory Programming Based on Robot Operating System［J］. Advances in Manufacturing， 2020， 8： 216-229.
[12]	LIU L， ZHU P， GUAN J， et al. A Binocular Reconstruction Method Fused with Laplacian Image Information for Pavement Texture Evaluation［J］. Measurement， 2021， 185： 110039.
[13]	毛成林，于瑞强，宋爱国. 一种结合TCP标定的深度相机手眼标定方法［J］. 仪器仪表学报， 2023， 44（3）： 280-286.
	MAO Chenglin， YU Ruiqiang， SONG Aiguo. A Hand-eye Calibration Method for Depth Camera Combined with TCP Calibration ［J］. Chinese Journal of Scientific Instrument， 2023， 44（3）： 280-286.
[14]	XU J， LI Q， WHITE BAI. A Novel Hand-eye Calibration Method for Industrial Robot and Line Laser Vision Sensor ［J］. Sensor Review， 2023， 43（4）： 259-265.
[15]	WU L， ZANG X， BAI M， et al. Simple and High‐precision Hand-eye Calibration for 3D Robot Measurement Systems ［J］. Advanced Intelligent Systems， 2023， 5（12）： 2300307.
[16]	PAN B， YU L， WU D， et al. Systematic Errors in Two-dimensional Digital Image Correlation due to Lens Distortion ［J］. Optics and Lasers in Engineering， 2013， 51（2）： 140-147.
[17]	TSAI R Y， LENZ R K. A New Technique for Fully Autonomous and Efficient 3D Robotics Hand-eye Calibration［J］. IEEE Transactions on Robotics and Automation， 1989， 5（3）： 345-358.