A deep learning framework based on improved self-supervised learning for ground-penetrating radar tunnel lining inspection

Huang，Jian1; Yang，Xi1; Zhou，Feng1,2; Li，Xiaofeng1; Zhou，Bin3; Lu，Song3; Ivashov，Sergey4; Giannakis，Iraklis5; Kong，Fannian6; Slob，Evert7

2023

10.1111/mice.13042

Computer-Aided Civil and Infrastructure Engineering   影响因子和分区

1093-9687

1467-8667

It is not practical to obtain a large number of labeled data to train a supervised learning network in tunnel lining nondestructive testing with ground-penetrating radar (GPR). To decrease the dependence of supervised learning on the number of labeled data, an improved self-supervised learning algorithm—self-attention dense contrastive learning (SA-DenseCL)—is proposed and incorporated with a mask region-convolution neural network (Mask R-CNN), which is trained by unlabeled and labeled GPR data. The proposed SA-DenseCL adds a self-attention-based relevant projection head to the DenseCL architecture of self-supervised learning, capturing the spatially continuing information between adjacent GPR traces. In the workflow, some unlabeled GPR images are used to pre-train the SA-DenseCL network for feature extraction and obtaining the backbone weights, which is superior to the conventional pre-training methods of supervised learning pre-trained by ImageNet images. The weights of the pre-trained backbone are then used to initialize the Mask R-CNN through transfer learning. Subsequently, a limited number of labeled GPR images are used to fine-tune the Mask R-CNN for automatically identifying the locations of the reinforcement bars and voids and estimating the secondary lining thickness. The experimental results show that the average precision reaches 96.70%, 81.04%, and 94.67% in identifying reinforcement bar locations, detecting void defects, and estimating secondary lining thickness, respectively, which outperform the conventional methods that use ImageNet-based supervised learning or GPR image-based DenseCL for initializing the Mask R-CNN backbone weights. It is observed that the improved self-supervised learning-based framework can improve the detection and estimation accuracy in GPR tunnel lining inspection.

SCI ; EI

英语

通讯

WOS:000991452400001

20232114126322

Deep learning ; Geological surveys ; Geophysical prospecting ; Learning algorithms ; Learning systems ; Nondestructive examination ; Radar imaging ; Reinforcement ; Supervised learning ; Tunnel linings

Tunnels and Tunneling:401.2 ; Ergonomics and Human Factors Engineering:461.4 ; Geology:481.1 ; Geophysical Prospecting:481.4 ; Radar Systems and Equipment:716.2 ; Machine Learning:723.4.2 ; Materials Science:951

ENGINEERING

2-s2.0-85159725752

Scopus

1.School of Mechanical Engineering and Electronic Information,China University of Geosciences (Wuhan),Wuhan,China
2.Guangdong Provincial Key Laboratory of Geophysical High-resolution Imaging Technology,Southern University of Science and Technology,Shenzhen,China
3.China Railway Southwest Research Institute Co. LTD,Chengdu,China
4.Remote Sensing Laboratory,Bauman Moscow State Technical University,Moscow,Russian Federation
5.School of Geosciences,University of Aberdeen,Aberdeen,United Kingdom
6.Norwegian Geotechnical Institute,Oslo,Norway
7.Department of Geoscience and Engineering,Delft University of Technology,Delft,Netherlands

Huang，Jian,Yang，Xi,Zhou，Feng,et al. A deep learning framework based on improved self-supervised learning for ground-penetrating radar tunnel lining inspection[J]. Computer-Aided Civil and Infrastructure Engineering,2023,39(6):814-833.

Huang，Jian.,Yang，Xi.,Zhou，Feng.,Li，Xiaofeng.,Zhou，Bin.,...&Slob，Evert.(2023).A deep learning framework based on improved self-supervised learning for ground-penetrating radar tunnel lining inspection.Computer-Aided Civil and Infrastructure Engineering,39(6),814-833.

Huang，Jian,et al."A deep learning framework based on improved self-supervised learning for ground-penetrating radar tunnel lining inspection".Computer-Aided Civil and Infrastructure Engineering 39.6(2023):814-833.