Deep learning-based GPR clutter suppression for detecting backfill grout defects behind segmental linings

Younseo Kim; Chaemin Hwang; Hyoungju Kim; Hangseok Choi

doi:10.9711/KTAJ.2025.27.5.431

All Issue

2025 Vol.27, Issue 5 Preview Page Next Page

Research Paper

Deep learning-based GPR clutter suppression for detecting backfill grout defects behind segmental linings 터널 세그먼트 라이닝의 뒤채움 그라우트 결함 탐지를 위한 딥러닝 기반 GPR 철근 클러터 제거 모델

30 September 2025. pp. 431-441

PDF XML

Abstract

Ground-penetrating radar (GPR) is a non-destructive testing technique that employs high-frequency electromagnetic waves to detect subsurface defects. With its broad survey coverage and rapid data acquisition, GPR is well-suited for assessing the condition of backfill grout behind shield TBM segmental linings. However, reinforcing bars embedded in the lining generate strong signal interference that introduces clutter into GPR images, obscuring defect signatures in the grout and hindering reliable evaluation. To solve this problem, a deep learning-based image segmentation model was developed for clutter suppression, thereby enhancing the visibility of defect signals. A training dataset was created using numerical GPR simulations based on the finite-difference time-domain (FDTD) method, and two image segmentation architectures, a fully convolutional network (FCN) and Deeplab V3+, were trained. The model’s performance was evaluated using three image similarity metrics: PSNR, SSIM, and MS-SSIM. The results demonstrate that the Deeplab V3+ outperforms the FCN, achieving superior clutter elimination while preserving defect visibility. The proposed model is expected to improve defect detection in backfill grout and enhance the efficiency of segmental lining maintenance, thereby contributing to the durability and extended service life of TBM tunnels.

Keywords

TBM tunnel

GPR inspection

Defect detection

Machine learning

Image segmentation

지표투과 레이더(ground penetrating radar, GPR)는 고주파 전자기파를 이용하여 지중의 결함을 탐지하는 비파괴 탐사 기법이다. GPR은 넓은 탐지 범위와 짧은 데이터 획득 시간을 제공하므로, 쉴드 TBM 세그먼트 라이닝 배면의 뒤채움 그라우트에 대한 상태 평가에 적합하다. 그러나 철근이 포함된 세그먼트 라이닝에서는 철근의 차폐 효과로 인해 GPR 이미지에 클러터(clutter)가 발생하며, 이로 인해 그라우트 내 결함 신호가 가려져 신뢰성 있는 평가를 어렵게 한다. 본 연구에서는 이러한 한계를 극복하기 위해 딥러닝 기반 이미지 분할 기법을 적용한 철근 클러터 제거 모델을 개발하여 결함 신호의 가시성을 개선하였다. 이를 위해 유한차분 시간영역(finite-difference time-domain, FDTD) 기반 GPR 수치해석 모델을 통해 데이터베이스를 구축하고, FCN (fully convolutional networks)과 Deeplab V3+ 등의 이미지 분할 기법을 적용하여 모델을 학습하였다. 각 기법의 성능은 PSNR, SSIM, MS-SSIM의 세 가지 이미지 유사도 지표를 활용해 평가하였다. 그 결과, Deeplab V3+가 FCN보다 우수한 클러터 제거 성능을 보였으며 결함 신호의 가시성을 효과적으로 향상시키는 것으로 확인되었다. 제안된 모델은 세그먼트 라이닝 뒤채움 그라우트의 결함 식별률을 개선하고 세그먼트 유지관리의 효율성을 높여, TBM 터널의 내구성 확보와 사용 수명 연장 등에 기여할 수 있을 것으로 기대된다.

키워드

TBM 터널

GPR 탐사

결함 탐지

머신 러닝

이미지 분할

References

Chen, L.C., Zhu, Y., Papandreou, G., Schroff, F., Adam, H. (2018), “Encoder-decoder with atrous separable convolution for semantic image segmentation”, Proceedings of the European Conference on Computer Vision (ECCV), Munich, Germany, pp. 801-818.

10.1007/978-3-030-01234-2_49

Go, G.H., Lee, S.J. (2021), “A study on numerical analysis for GPR signal characterization of tunnel lining cavities”, Journal of the Korean Geotechnical Society, Vol. 37, No. 10, pp. 65-76.

10.7843/KGS.2021.37.10.65

Kingma, D.P., Ba, J. (2014), “Adam: A method for stochastic optimization”, arXiv preprint arXiv:1412.6980.

Li, J., Liu, C., Zeng, Z., Chen, L. (2015), “GPR signal denoising and target extraction with the CEEMD method”, IEEE Geoscience and Remote Sensing Letters, Vol. 12, No. 8, pp. 1615-1619.

10.1109/LGRS.2015.2415736

Long, J., Shelhamer, E., Darrell, T. (2015), “Fully convolutional networks for semantic segmentation”, Proceedings of the IEEE conference on computer vision and pattern recognition, Boston, USA, pp. 3431-3440.

10.1109/CVPR.2015.7298965

Meguid, M.A., Dang, H.K. (2009), “The effect of erosion voids on existing tunnel linings”, Tunnelling and Underground Space Technology, Vol. 24, No. 3, pp. 278-286.

10.1016/j.tust.2008.09.002

Ruder, S. (2016), “An overview of gradient descent optimization algorithms”, arXiv preprint arXiv:1609.04747.

Terrasse, G., Nicolas, J.M., Trouvé, E., Drouet, E. (2017), “Application of the curvelet transform for clutter and noise removal in GPR data”, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, Vol. 10, No. 10, pp. 4280-4294.

10.1109/JSTARS.2017.2717960

Wang, J., Huang, H., Xie, X., Bobet, A. (2014), “Void-induced liner deformation and stress redistribution”, Tunnelling and Underground Space Technology, Vol. 40, pp. 263-276.

10.1016/j.tust.2013.10.008

Wang, Z., Simoncelli, E.P., Bovik, A.C. (2003), “Multiscale structural similarity for image quality assessment”, Proceedings of the Thirty-Seventh Asilomar Conference on Signals, Systems & Computers 2003, Pacific Grove, USA, Vol. 2, pp. 1398-1402.

10.1109/ACSSC.2003.1292216

Warren, C., Giannopoulos, A., Giannakis, I. (2016), “gprMax: Open source software to simulate electromagnetic wave propagation for Ground Penetrating Radar”, Computer Physics Communications, Vol. 209, pp. 163-170.

10.1016/j.cpc.2016.08.020

Zeng, L., Zhang, X., Xie, X., Zhou, B., Xu, C., Lambot, S. (2023), “Measuring annular thickness of backfill grouting behind shield tunnel lining based on GPR monitoring and data mining”, Automation in Construction, Vol. 150, 104811.

10.1016/j.autcon.2023.104811

Zhou, Y., Chen, W. (2019), “MCA-based clutter reduction from migrated GPR data of shallowly buried point target”, IEEE Transactions on Geoscience and Remote Sensing, Vol. 57, No. 1, pp. 432-448.

10.1109/TGRS.2018.2855728

Information

Publisher :Korean Tunneling and Underground Space Association
Publisher(Ko) :한국터널지하공간학회
Journal Title :Journal of Korean Tunnelling and Underground Space Association
Journal Title(Ko) :한국터널지하공간학회 논문집
Volume : 27
No :5
Pages :431-441
Received Date : 2025-08-22
Revised Date : 2025-09-16
Accepted Date : 2025-09-19
DOI :https://doi.org/10.9711/KTAJ.2025.27.5.431

Journal of Korean Tunnelling and Underground Space Association ISSN:2233-8292(Print) 2287-4747(Online) 한국터널지하공간학회 논문집

All Issue