Quality Assurance of Tunnel Lining Construction Using Ground-Penetrating Radar and Convolution Neural Networks

Donghao Zhang,Hui Qin,Zhengzheng Wang,Junwei Huang

DOI: https://doi.org/10.1190/gpr2020-058.1

2020-01-01

Abstract:PreviousNext No Access18th International Conference on Ground Penetrating Radar, Golden, Colorado, 14–19 June 2020Quality assurance of tunnel lining construction using ground-penetrating radar and convolution neural networksAuthors: Donghao ZhangHui QinZhengzheng WangJunwei HuangDonghao ZhangSchool of Civil Engineering, Dalian University of Technology, ChinaSearch for more papers by this author, Hui QinSchool of Civil Engineering, Dalian University of Technology, ChinaSearch for more papers by this author, Zhengzheng WangSchool of Civil Engineering, Dalian University of Technology, ChinaSearch for more papers by this author, and Junwei HuangSchool of Civil Engineering, Dalian University of Technology, ChinaSearch for more papers by this authorhttps://doi.org/10.1190/gpr2020-058.1 SectionsAboutPDF/ePub ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InRedditEmail Abstract Ground penetrating radar (GPR) is one of the most recommended tool for routine inspection of tunnel linings, yet most of the data interpretation work is accomplished by manpower. In order to rapidly evaluate the arrangement of the steel ribs as well as the presence of voids in the initial lining structure of tunnels, we developed an automatic recognition algorithm based on the convolution neural network (CNN) that was designed to search for hyperbolic features from GPR B-scan images. In this work, we used the Faster-RCNN framework together with the VGG16 model for feature extraction, object classification, and position regression. The designed network was first pre-trained on ImageNet for a set of model parameters, then on 198 processed GPR images for fine tuning. Finally, the designed network was tested with 70 GPR images and the recognition accuracies reached 99% for steel ribs and 94% for voids, indicating that the designed method is capable of recognizing targets automatically in GPR images. Keywords: GPR, neural network, processing, algorithm, filteringPermalink: https://doi.org/10.1190/gpr2020-058.1FiguresReferencesRelatedDetails 18th International Conference on Ground Penetrating Radar, Golden, Colorado, 14–19 June 2020ISSN (online):2159-6832Copyright: 2020 Pages: 455 publication data© 2020 Published in electronic format with permission by the Society of Exploration GeophysicistsPublisher:Society of Exploration Geophysicists HistoryPublished Online: 11 Nov 2020 CITATION INFORMATION Donghao Zhang, Hui Qin, Zhengzheng Wang, and Junwei Huang, (2020), "Quality assurance of tunnel lining construction using ground-penetrating radar and convolution neural networks," SEG Global Meeting Abstracts : 216-219. https://doi.org/10.1190/gpr2020-058.1 Plain-Language Summary KeywordsGPRneural networkprocessingalgorithmfilteringPDF DownloadLoading ...