面向智能制造的高精度三维视觉测量系统研究

随着我国制造业走向智能化和精密化，对生产线零件在线精密三维测量的需求也日益增长。传统的接触式三维测量技术，如三坐标测量仪，需要人工辅助进行测量，较为繁琐，难以满足零件在线全检的需求；而以 3D 视觉为代表的非接触式测量技术有着精度高、速度快、不损伤工件等优点，在工业三维测量中逐步得到运用。本文面向复杂零件的高精度测量需求，设计了一套基于工业机器人和面阵结构光相机的三维视觉测量系统方案。

本文首先研究了测量系统的系统架构、系统通信方式、软件架构等技术细节。研究基于模型的视点规划方法，研究系统约束条件并构建传感器模型，生成并筛选视点，对局部特征进行视点规划。研究了基于手眼系统和标记点的拼接方案，并提出了一种基于标定块的多视点拼接方法，实现多视点点云拼接，并提出了一种基于标定块的多视点拼接方法。研究了三维测量领域常用的滤波、分割、拟合和配准等点云处理算法，并通过理论结合实验分析算法场景适用性。以汽车零部件为测量对象进行了测量实验。搭建测量系统实验平台，进行了视点规划、多视点点云拼接、特征点云提取和拟合等实验，实现了管口三维位置度的测量，并基于 QT开发测量软件实现结果显示。经过以三坐标测量仪测量值为真值的重复测量实验，测量系统对最大尺寸为245 mm 的汽车零配件进行位置度测量的重复精度为 0.06 mm，比例误差为0.025%；绝对精度为 0.2 mm，比例误差为 0.082 %。结果表明，本课题基于工业机器人设计的三维视觉测量系统方案对智能制造在线精密检测具有技术可行性及实用价值。

[1] DENG Z P, LI S G, HUANG X. A flexible and cost¬-effective compensation method for leveling using large¬scale coordinate measuring machines and its application in aircraft digital assembly[J]. Measurement Science & Technology, 2018.
[2] ZHU S, GAO Y. Noncontact 3­D coordinate measurement of cross­cutting feature points on the surface of a large­scale workpiece based on the machine vision method[J]. IEEE Transactions on Instrumentation & Measurement, 2010, 59(7): 1874­1887.
[3] FRANCESCHINI F, MAISANO D, MASTROGIACOMO L. Mobile spatial coordinate measuring system (MScMS) and CMMs: a structured comparison[J]. International Journal of Advanced Manufacturing Technology, 2009, 42(11­12): 1089­1102.
[4] 余苏, 运金芬, 张进, 等. 激光跟踪仪在重型装备在线测量中的应用[J]. 中国重型装备,2020(1): 3.
[5] KAMALI K, JOUBAIR A, BONEV I A, et al. Elasto­geometrical calibration of an industrial robot under multidirectional external loads using a laser tracker[C]//2016 IEEE International Conference on Robotics and Automation (ICRA). 2016.
[6] XU X, BUREICK J, HAO Y, et al. TLS­based composite structure deformation analysis validated with laser tracker[J]. Composite Structures, 2017, 202(OCT.): 60­65.
[7] 衣本强, 孙慧兰, 于涛. 论 3D 视觉在汽车行业的应用和发展趋势[Z]. 2021.
[8] ZHANG S. High­speed 3D shape measurement with structured light methods: A review[J/OL]. Optics and Lasers in Engineering, 2018, 106: 119­131. https://www.sciencedirect.com/science/article/pii/S0143816617313246. DOI: https://doi.org/10.1016/j.optlaseng.2018.02.017.
[9] EL­HAKIM S F, BERALDIN J A, BLAIS F. Comparative evaluation of the performance of passive and active 3D vision systems[C]//Digital Photogrammetry and Remote Sensing’95: Vol. 2646. SPIE, 1995: 14­25.
[10] 卢荣胜, 史艳琼, 胡海兵. 机器人视觉三维成像技术综述[J]. 激光与光电子学进展, 2020, 57(9­27).
[11] FANG C, ZOU G M, HOU Y, et al. Measurement method of glass gob volume based on binocular vision[J]. Jisuanji Xitong Yingyong = Computer Systems and Applications, 2022(8): 402.
[12] MEHRANIAN A, WOLLENWEBER S D, WALKER M D, et al. Deep learning–based timeof­flight (ToF) image enhancement of non­ToF PET scans[J]. European journal of nuclearmedicine and molecular imaging, 2022, 49(11): 3740­3749.
[13] 黄远建, 李晓银, 叶文怡, 等. 基于共聚焦亚像素扫描的高分辨三维成像[J]. 光学学报, 2023, 43(8): 0822014.
[14] ZHANG R, LIU W, LU Y, et al. Local high precision 3D measurement based on line lasermeasuring instrument[C]//Vol. 10710. SPIE, 2018: 107101L­107101L­6.
[10] 卢荣胜, 史艳琼, 胡海兵. 机器人视觉三维成像技术综述[J]. 激光与光电子学进展, 2020,57(9¬27).
[11] FANG C, ZOU G M, HOU Y, et al. Measurement method of glass gob volume based on binocular vision[J]. Jisuanji Xitong Yingyong = Computer Systems and Applications, 2022(8): 402.
[12] MEHRANIAN A, WOLLENWEBER S D, WALKER M D, et al. Deep learning–based timeof¬flight (ToF) image enhancement of non¬ToF PET scans[J]. European journal of nuclearmedicine and molecular imaging, 2022, 49(11): 3740¬3749.
[13] 黄远建, 李晓银, 叶文怡, 等. 基于共聚焦亚像素扫描的高分辨三维成像[J]. 光学学报, 2023, 43(8): 0822014.
[14] ZHANG R, LIU W, LU Y, et al. Local high precision 3D measurement based on line lasermeasuring instrument[C]//Vol. 10710. SPIE, 2018: 107101L¬107101L¬6. 57(9¬27).
[11] FANG C, ZOU G M, HOU Y, et al. Measurement method of glass gob volume based on binocular vision[J]. Jisuanji Xitong Yingyong = Computer Systems and Applications, 2022(8): 402.
[12] MEHRANIAN A, WOLLENWEBER S D, WALKER M D, et al. Deep learning–based timeof¬flight (ToF) image enhancement of non¬ToF PET scans[J]. European journal of nuclearmedicine and molecular imaging, 2022, 49(11): 3740¬3749.
[13] 黄远建, 李晓银, 叶文怡, 等. 基于共聚焦亚像素扫描的高分辨三维成像[J]. 光学学报 2023, 43(8): 0822014.
[14] ZHANG R, LIU W, LU Y, et al. Local high precision 3D measurement based on line lasermeasuring instrument[C]//Vol. 10710. SPIE, 2018: 107101L¬107101L¬6.
[15] WU H, CAO Y, AN H, et al. High¬precision 3D shape measurement of rigid moving objectsbased on the Hilbert transform[J]. Applied optics (2004), 2021, 60(27): 8390¬8399.
[16] VAN DER JEUGHT S, DIRCKX J J. Real¬time structured light profilometry: a review[J].Optics and lasers in engineering, 2016, 87: 18¬31.
[17] LI J, LIU G, LIU Y. A dynamic volume measurement system with structured light vision[C] 2016 31st Youth Academic Annual Conference of Chinese Association of Automation (YAC). IEEE, 2016: 251¬255.
[18] LIN H, ZHANG H, LI Y, et al. 3D point cloud capture method for underwater structures in turbid environment[J]. Measurement Science and Technology, 2020, 32(2): 025106.
[19] YU H, HUANG Y, ZHENG D, et al. Three¬dimensional shape measurement technique for largescale objects based on line structured light combined with industrial robot[J]. Optik, 2020, 202: 163656.
[20] PENG W, WANG Y, MIAO Z, et al. Viewpoints Planning for Active 3¬D Reconstruction of Profiled Blades Using Estimated Occupancy Probabilities (EOP)[J/OL]. IEEE Transactions on Industrial Electronics, 2021, 68(5): 4109¬4119. DOI: 10.1109/TIE.2020.2987286.
[21] SHI Q, XI N, SPAGNULUO C. A feedback design to a CAD¬guided area sensor planning system for automated 3D shape inspection[J/OL]. Computer¬aided design and applications, 2007, 4(1¬4): 209¬218. https://doi.org/10.1080/16864360.2007.10738541.
[22] CHEN S, LI Y. Automatic sensor placement for model¬based robot vision[J/OL]. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 2004, 34(1): 393¬408. DOI: 10.1109/TSMCB.2003.817031.
[23] TARBOX G H, GOTTSCHLICH S N. Planning for Complete Sensor Coverage in Inspection[J]. Computer Vision & Image Understanding, 1995, 61(1): 84¬111.
[24] SCOTT W R. Model¬based view planning[J]. Machine Vision & Applications, 2009, 20(1): 47¬69.
[25] MOROOKA K, ZHA H, HASEGAWA T. Next best viewpoint (NBV) planning for active object modeling based on a learning¬by¬showing approach[C/OL]//Proceedings. Fourteenth International Conference on Pattern Recognition: Vol. 1. 1998: 677¬681 vol.1. DOI: 10.1109/ICPR.1998.711234.
[26] SAKANE S, SATO T, KAKIKURA M. Model¬based planning of visual sensors using a handeye action simulator system[C]//International Conference on Advanced Robotics. 1987.
[27] SAKANE S, ISH M, KAKIKURA M. Occlusion avoidance of visual sensors based on a handeye action simulator system: HEAVEN[J]. Advanced Robotics, 1986, 2(2): 149¬165.
[28] SAKANE S, SATO T, KAKIKURA M, et al. Automatic planning of light source placement for an active photometric stereo system[C]//IEEE International Conference on Robotics and Automation. 2009.
[29] SAKANE S, SATO T, KAKIKURA M. Planning focus of attentions for visual feedback control [J]. Transactions of the Society of Instrument and Control Engineers, 2009, 24(6): 608-615.
[30] COWAN C K. Model¬based synthesis of sensor location[C]//IEEE International Conference on Robotics & Automation. 1988.
[31] XU H, YU L, HOU J, et al. Automatic reconstruction method for large scene based on multi¬site point cloud stitching[J]. Measurement, 2019, 131: 590¬596.
[32] FU J, DING Y, HUANG T, et al. Hand– eye calibration method based on three-dimensional visual measurement in robotic high¬precision machining[J/OL]. International Journal of Advanced Manufacturing Technology, 2022, 119(5¬6): 3845 – 3856. https://www.scopus.com/inward/record.uri?eid=2¬s2.0-85122668609anddoi=10.1007%2fs00170¬021¬08591¬1andpartn erID=40andmd5=5debb925d23628726b48dc3ff3b9fc46. DOI: 10.1007/s00170¬021¬08591¬1.
[33] YIN S, REN Y, GUO Y, et al. Development and calibration of an integrated 3D scanning system for high¬accuracy large¬scale metrology[J/OL]. Measurement, 2014, 54: 65¬76. https://www.sciencedirect.com/science/article/pii/S0263224114001675. DOI: https://doi.org/10.1016/j.measurement.2014.04.009.
[34] WANG J, TAO B, GONG Z, et al. A mobile robotic measurement system for large¬scale complex components based on optical scanning and visual tracking[J/OL]. Robotics and ComputerIntegrated Manufacturing, 2021, 67: 102010. https://www.sciencedirect.com/science/article/pii/S0736584520302210. DOI: https://doi.org/10.1016/j.rcim.2020.102010.
[35] HAN L Y, CHENG X, LI Z W, et al. A robot¬driven 3D shape measurement system for automaticquality inspection of thermal objects on a forging production line[J/OL]. Sensors, 2018, 18(12):4368. DOI: 10.3390/s18124368.
[36] 孟祥丽, 刘瑶, 吴涛. 一种圆形编码标志点的设计及解码算法研究[J/OL]. 传感器与微系统, 2018, 37(42¬44). DOI: 10.13873/j.1000¬9787(2018)07¬0042¬03.
[37] 伏燕军, 翁凌霄, 胡茗. 基于单目结构光的大物体三维测量关键方法的研究[J]. 应用光学,2018, 39(212¬219).
[38] 梁云波, 邓文怡, 娄小平, 等. 基于标志点的多视三维数据自动拼接方法[J/OL]. 北京信息科技大学学报 (自然科学版), 2010, 25(30¬33). DOI: 10.16508/j.cnki.11-5866/n.2010.01.020.
[39] GENG J. Structured¬light 3D surface imaging: a tutorial[J/OL]. Adv. Opt. Photon., 2011, 3(2):128¬160. https://opg.optica.org/aop/abstract.cfm?URI=aop¬3¬2¬128. DOI: 10.1364/AOP.3.000128.
[40] YUAN M, MUHAMMAD A, RUKSHAN H, et al. A collaborative robotic approach for inspection and anomaly detection in industrial applications[M]//Lecture Notes in Computer Science:Social Robotics. Cham: Springer International Publishing, 2021: 757¬762.
[41] DEINZER F, DENZLER J, NIEMANN H. Viewpoint selection–planning optimal sequences of views for object recognition[C]//Lecture Notes in Computer Science: Computer Analysis of Images and Patterns. Berlin, Heidelberg: Springer Berlin Heidelberg, 2003: 65¬73.
[42] 王彦科. 三角网格模型顶点法矢研究[J]. 科技风, 2020(20): 183.
[43] 马元魁, 白晓亮. 三角网格模型体素特征分割[J]. Ji suan ji ke xue, 2015, 42(10): 13¬15.
[44] XIE Z X X Q. Computer¬Aided Inspection Planning—The state of the art[J]. Computers inIndustry, 2009.
[45] KAMRANI A, ABOUEL NASR E, AL¬AHMARI A, et al. Feature¬based design approach for integrated CAD and computer¬aided inspection planning[J]. International Journal of Advanced Manufacturing Technology, 2015, 76(9¬12): 2159¬2183.
[46] LEE R, AHN Y J. Limit curve of H¬Bézier curves and rational Bézier curves in standard form with the same weight[J]. Journal of computational and applied mathematics, 2015, 281: 1¬9.
[47] 杨轶焬, 刘建群, 高伟强. 基于非均匀 B 样条曲线的机器人轨迹过渡算法[J]. Zu he ji chuang yu zi dong hua jia gong ji shu, 2022(2): 20¬28.
[48] 王龙权, 陈小雕, 陈立庚. 点到 NURBS 曲线最近距离的快速计算方法[J]. Ji suan ji fu zhu she ji yu tu xing xue xue bao, 2019, 31(1): 26¬30.
[49] 张胜文, 尤炎炎, 李滨城, 等. 数字化设计与制造三维 MBD 模型尺寸完备性检查及分析技术研究[J]. 图学学报, 2020, 41(4): 7.
[50] 轩左晨, 李智, 周丹晨. 交互式 MBD 模型工艺审查方法研究[J]. 现代制造工程, 2021(008): 000.
[51] WANG, XIAOMIN, XIE, et al. Research on a handheld 3D laser scanning system for measuring large¬sized objects[J]. Sensors, 2018.
[52] HOU X, SONG Y, HAO Q, et al. Measurement of parallelism and perpendicularity of largesize workpieces by laser alignment method[C]//Optical Test & Measurement Technology & Equipment. 2005.
[53] LIU Z, LIU X, DUAN G, et al. Precise hand–eye calibration method based on spatial distance and epipolar constraints[J]. Robotics and Autonomous Systems, 2021, 145: 103868.
[54] ZHANG Z, ZHANG L, YANG G Z. A computationally efficient method for hand– eye calibration[J]. International Journal for Computer Assisted Radiology and Surgery, 2017, 12(10): 1775¬1787.
[55] 李祥云, 向民志, 范百兴, 等. 工业机器人运动学参数标定精度分析与改进[J]. 测绘科学技术学报, 2018, 35(3): 5.
[56] 张敏, 隋永新, 杨怀江. 提取标记点中心在子孔径拼接检测中的应用[J]. 中国光学, 2014, 7(5): 830¬836.
[57] YANG H, CHEN J. Point cloud data enhancement based on homomorphic filter[C]//2014 IEEE 5th International Conference on Software Engineering and Service Science. 2014: 509-512.
[58] XIE Y, CHEN K, LI W, et al. An improved adaptive threshold RANSAC method for medium tillage crop rows detection[C]//2021 6th International Conference on Intelligent Computing and Signal Processing (ICSP). 2021: 1282¬1286.
[59] PANG S, KENT D, CAI X, et al. 3D scan registration based localization for autonomous vehicles ¬ a comparison of NDT and ICP under realistic conditions[C]//2018 IEEE 88th Vehicular Technology Conference (VTC¬Fall). IEEE, 2018: 1¬5.
[60] YANG J, LI H, CAMPBELL D, et al. Go¬ICP: A globally optimal solution to 3D ICP point¬set registration[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2016, 38(11): 2241¬2254.
[61] 安维胜, 余让明, 伍玉铃. 基于 FAST 和 SURF 的图像配准算法[J]. 计算机工程, 2015, v.41;No.456(10): 238¬241+245.
[62] 秦红星, 刘镇涛, 谭博元. 深度学习刚性点云配准前沿进展[J]. 中国图象图形学报, 2022, 27(329¬348).
[63] LU J, LIU W, DONG D L, et al. Point cloud registration algorithm based on NDT with variable size voxel[C]//2015 34th Chinese Control Conference (CCC). Technical Committee on Control Theory, Chinese Association of Automation, 2015: 3707¬3712.
[64] WANG J, LU Y, ZHANG J, et al. Optimization and performance verification of high efficiency ICP registration for laser point clouds[J]. Hongwai Yu Jiguang Gongcheng = Infrared and Laser Engineering, 2021(10): 301.