作者：张林龙¹, 张伟², 胡昌华¹, 周志杰¹

作者单位：1. 火箭军工程大学302室, 陕西 西安 710025;
2. 火箭军工程大学403室, 陕西 西安 710025

关键词：单目测量;极线约束;光束法平差;形变测量

摘要：

针对大型结构形变的摄像测量中，双目系统结构复杂、成本高、特征点匹配度低，而传统单目测量系统不能有效获取目标深度信息等问题，借助平面反射镜提出一种新的单目像机三维形变测量方法。首先，基于平面反射镜的光学反射原理构建单目三维测量模型；其次，利用像机与平面镜的内在参数推导出单目像机视图下的极线约束关系；然后，在测量系统完成参数标定的情况下，将极线约束关系作为约束条件引入利用基于LM算法的改进光束法平差进行结果优化，完成三维目标点的位置精确测量，进而实现大型结构的形变测量。仿真实验与实际测量实验结果表明单目像机测量方法的有效性与优越性，具有实际工程应用价值。

The 3D structure deformation measurement method based on monocular camera

ZHANG Linlong¹, ZHANG Wei², HU Changhua¹, ZHOU Zhijie¹

1. 302 Unit, Rocket Force University of Engineering, Xi'an 710025, China;
2. 403 Unit, Rocket Force University of Engineering, Xi'an 710025, China

Abstract: Binocular measurement system has the problem of complexity,high price,low matching precision in the deformation measurement of large-scale structure.Meanwhile,traditional monocular system cannot obtain depth information of 3D(three dimensional) object.A new method for measuring the 3D deformation that uses a single video camera utilizing a planar reflector was proposed in this paper.Firstly a monocular measuring model of 3D object was founded utilizing the reflective principle of planar reflector.Then an epipolar geometry inhibition equation in the same video camera view was settled by inherent parameters of camera and reflector.The measured point was located accurately by the improved bundle adjustment based on LM algorithm after the intrinsic and extrinsic camera parameters of the measurement system were calibrated.So the deformation measurement of large-scale structure was achieved.Both simulation and indoor experiments results validate that the proposed measuring technique is effective and has the realism value in engineering applications.

Keywords: monocular;epipolar geometry inhibition;bundle adjustment;deformation measurement

2016, 42(11): 17-22  收稿日期: 2016-3-2;收到修改稿日期: 2016-3-2

基金项目: 国家杰出青年基金（61025014）；国家自然科学基金（61174030，61104223，61374120）

作者简介: 张林龙(1991-),男,河南驻马店市人,硕士研究生,专业方向为诊断技术及自动化装置。

参考文献

[1] 周翔,胡建,孙林涛,等.大型结构变形实时摄像测量方法[J].实验力学,2012,27(4):503-510.
[2] 于起峰,尚洋,伏思华,等.大型结构变形及形貌摄像测量技术研究进展[J].实验力学,2011,26(5):479-490.
[3] 黄小云,高峰,徐国艳,等.基于单幅立式标靶图像的单目深度信息提取[J].北京航空航天大学学报,2015,41(4):649-655.
[4] 晁志超,伏思华,姜广文,等.单目摄像机-激光测距传感器位姿测量系统[J].光学学报,2011,31(3):85-91.
[5] OYA Y,KAWASUE K.Three-dimensional quantitative visualization from a single image[J].Sensors & Transducers,2009(7):142-154.
[6] KAWASUE K,NAGATOMO S,OYA Y.Three dimensional measurement of aquatic organisms using a single video camera[J].Sensors &Transducers,2010(6):118-126.
[7] FENG X,PAN D.Research on the application of single camera stereo vision sensor in three-dimensional point measurement[J].Journal of Modern Optics,2015(62):1-7.
[8] 祝强,李少康,徐臻.LM算法求解大残差非线性最小二乘问题研究[J].中国测试,2016,42(3):13-15.
[9] 李德仁,袁修孝.误差处理与可靠性分析[M].武汉:武汉大学出版社,2002:240-255.
[10] 马友青,贾永红,刘少创,等.基于LM法的光束法平差巡视器导航定位[J].东北大学学报(自然科学版),2014,35(4):489-493.
[11] 朱焱丹,廖俊必,何子牛.基于LabVIEW的机器视觉系统的标定[J].中国测试,2009,35(6):54-56.