摘要
三维数字图像相关方法具有全场、非接触和观测尺度易调节等优势,是目前科学研究和工程实践中应用广泛的变形场测量手段.在某些工业和国防领域,为了实现对大型复杂结构的全方位三维变形测量,常常需要使用多台摄像机从不同角度进行拍摄.由于现场条件限制,这些实验场景中经常难以精确地布置用于相机标定的特殊靶标或标记,这给相机成像参数标定、统一多视角下相机的坐标系、将各个视角下的测量数据有效拼接带来了重大挑战.本研究发展了一种自标定多视角三维数字图像相关测量方法,其利用待测结构自身纹理信息,以基于运动恢复结构算法的自标定方法,解算得到所有相机在世界坐标系中的内外部成像参数,然后结合多视角三角化方法获得世界坐标系中的三维变形场测量结果.本文介绍了自标定多视角三维数字图像相关方法的原理.通过高精度6自由度平台驱动的圆筒平移、旋转实验,以及气囊放气过程变形实验验证了方法的有效性.分析和实验结果表明自标定多视角三维数字图像相关方法在测量时无需进行繁琐的标定操作和坐标系统一操作,可显著简化测量流程,同时还可避免传统方法中多个双目系统拼接接缝和重叠问题,因此本文方法尤其适用于大型复杂工程结构的现场变形测量.
Three-dimensional(3D)digital image correlation(DIC)methods are extensively utilized for deformation field measurement in scientific research and engineering practices for their advantages such as full-field,noncontact,and adjustable observation scales.In scenarios involving large and complex structures in industrial and defense fields,the comprehensive acquisition of 3D deformation measurements often necessitates the deployment of multiple cameras to capture images from diverse angles.However,experimental limitations often render it difficult to accurately position specific targets or markers for camera calibration,presenting substantial challenges in calibrating camera imaging parameters,establishing a unified coordinate system across multiple views,and effectively integrating data from different perspectives.In this study,a self-calibration multi-view(SCM)3DDIC measurement method was developed.Firstly,utilizing the texture information of the structure itself and a self-calibration method with structure from motion,the intrinsic and extrinsic parameters of all cameras were ob-tained for the world coordinate system.Then,combined with multi-view triangulation method,the measurement results of the 3D deformation field were obtained for the world coordinate system.And also,the principle of the method was described and its effectiveness was verified through experiments involving high-precision six-degree-of-freedom platform-driven cylinder translation and rotation,as well as deformation experiments of airbag defla-tion.Analysis and experimental results show that the self-calibration multi-view 3D-DIC method can eliminate the need for tedious calibration operations and coordinate system unification steps during measurements.This method can greatly simplify the measurement process and avoid seam and overlap issues encountered in tradi-tional methods and multiple stereo systems.Therefore,the proposed method is particularly well-suited for the de-formation measurements of large and complex engineering structures.
作者
程向楠
马沁巍
马少鹏
CHENG Xiangnan;MA Qinwei;MA Shaopeng(School of Ocean and Civil Engineering,Shanghai Jiao Tong University,Shanghai 200240,China;School of Aerospace Engineering,Beijing Institute of Technology,Beijing 100081,China)
出处
《北京理工大学学报》
EI
CAS
CSCD
北大核心
2024年第12期1297-1309,共13页
Transactions of Beijing Institute of Technology
基金
国家自然科学基金资助项目(12132009,12272046)
国家重点研发计划资助项目(2023YFF0616800)
上海交通大学“深蓝计划”基金资助项目(SL2022ZD101)。
关键词
三维数字图像相关
相机自标定
运动恢复结构
多视图几何
3D digital image correlation
camera self-calibration
structure from motion
multi-view geometry