Detection of structural defect and deformation based on multi-view geometric three-dimensional reconstruction method

被引：0
作者：
Liu Y.-F. [1 ]
Fan J.-S. [1 ]
Kong S.-Y. [1 ]
Wei X.-C. [1 ]
机构：
[1] Department of Civil Engineering, Tsinghua University, Key Laboratory of Civil Engineering Safety and Durability of China Ministry of Education, Tsinghua University, Beijing
来源：
Gongcheng Lixue/Engineering Mechanics | 2020年 / 37卷 / 09期
关键词：
Bending deformation; Damage detection; Damage localization; Digital image processing; Multi-view geometric three-dimensional reconstruction; Surface crack;
D O I：
10.6052/j.issn.1000-4750.2019.10.0598
中图分类号：
学科分类号：
摘要：
Surface damage and deformation are important indicators for testing, appraisal and monitoring of structural safety. The adoption of digital image method can effectively overcome the shortcomings of current manual inspection methods. However in engineering practice, challenges are encountered such as the difficulty in correcting the geometric deformation of the image during quantitative detection, the difficulty of damage localization of images in the overall structure, and the inability to measure the spatial deformation of the steel structure. By combining the research and application of multi-view geometric 3D reconstruction method, the above problems can be effectively solved. This paper discusses the principle and algorithm for the implementation of multi-view geometric 3D reconstruction, and introduces the classic and efficient algorithms for practice. Aiming at the problem of the digital image detection method in engineering practice, a surface projection method for the correction of imaging geometric deformation and damage localization, and a reverse engineering modeling and feature extraction method for the geometric deformation damage detection are proposed. Through three application studies including structural surface crack identification, large structure surface damage localization and steel structure component deformation identification, the specific operation methods and comparative advantages of multi-view geometric 3D reconstruction method for structural surface damage and deformation identification are discussed. The multi-view geometry three-dimensional reconstruction method has the benefits of low equipment requirements, convenient operation, rich model color and relatively high point cloud precision, After combined with the digital image method, it shows great research and application potential in the field of engineering structure inspection and monitoring. Copyright ©2020 Engineering Mechanics. All rights reserved.
引用
收藏
页码：103 / 111
页数：8
相关论文
共 22 条
[1]  
Zisserman A., Multiple view geometry in computer vision, (2003)
[2]  
Huang T S, Netravali A N., Motion and structure from feature correspondences: A review [J], Proceedings of the IEEE, 82, 2, pp. 252-268, (1994)
[3]  
Lowe D G., Distinctive image features from scale-invariant keypoints [J], International Journal of Computer Vision, 60, 2, pp. 91-110, (2004)
[4]  
Alcantarilla P F, Bartoli A, Davison A J., KAZE features [C], European Conference on Computer Vision, pp. 214-227, (2012)
[5]  
Esteban I, Dijk J, Groen F., FIT3D toolbox: multiple view geometry and 3d reconstruction for matlab, Electro-Optical Remote Sensing, Photonic Technologies, and Applications IV, 7835, (2010)
[6]  
Wu C, Agarwal S, Curless B, Et al., Multicore bundle adjustment [C], CVPR 2011, pp. 3057-3064, (2011)
[7]  
Wu C., Visualsfm: A visual structure from motion system, (2011)
[8]  
Cheng J, Leng C, Wu J, Et al., Fast and accurate image matching with cascade hashing for 3D reconstruction [C], Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, (2014)
[9]  
Triggs B, McLauchlan P F, Hartley R I, Et al., Bundle adjustment — a modern synthesis [C], International Workshop on Vision Algorithms, pp. 298-372, (1999)
[10]  
Moulon P, Monasse P, Marlet R., Adaptive structure from motion with a contrario model estimation [C], Asian Conference on Computer Vision, pp. 257-270, (2012)
123