Geomechanics and Engineering

  • 제33권2호
  • /
  • Pages.203-209
  • /
  • 2023
  • /
  • 2005-307X(pISSN)
  • /
  • 2092-6219(eISSN)
테크노프레스 (Techno-Press)
권호 표지
DOI QR코드

DOI QR Code

Subsurface anomaly detection utilizing synthetic GPR images and deep learning model

  • Ahmad Abdelmawla (University of Georgia (College of Engineering, School of Environmental, Civil, Agricultural and Mechanical Engineering) ;
  • Shihan Ma (University of Georgia (College of Engineering, School of Environmental, Civil, Agricultural and Mechanical Engineering) ;
  • Jidong J. Yang (University of Georgia (College of Engineering, School of Environmental, Civil, Agricultural and Mechanical Engineering) ;
  • S. Sonny Kim (University of Georgia (College of Engineering, School of Environmental, Civil, Agricultural and Mechanical Engineering)
  • 투고 : 2022.11.28
  • 심사 : 2023.04.07
  • 발행 : 2023.04.25
⟨ 이전 논문 다음 논문 ⟩

초록

One major advantage of ground penetrating radar (GPR) over other field test methods is its ability to obtain subsurface images of roads in an efficient and non-intrusive manner. Not only can the strata of pavement structure be retrieved from the GPR scan images, but also various irregularities, such as cracks and internal cavities. This article introduces a deep learning-based approach, focusing on detecting subsurface cracks by recognizing their distinctive hyperbolic signatures in the GPR scan images. Given the limited road sections that contain target features, two data augmentation methods, i.e., feature insertion and generation, are implemented, resulting in 9,174 GPR scan images. One of the most popular real-time object detection models, You Only Learn One Representation (YOLOR), is trained for detecting the target features for two types of subsurface cracks: bottom cracks and full cracks from the GPR scan images. The former represents partial cracks initiated from the bottom of the asphalt layer or base layers, while the latter includes extended cracks that penetrate these layers. Our experiments show the test average precisions of 0.769, 0.803 and 0.735 for all cracks, bottom cracks, and full cracks, respectively. This demonstrates the practicality of deep learning-based methods in detecting subsurface cracks from GPR scan images.

키워드

참고문헌

  1. Abdelmawla A. and Kim, S.S. (2020), "Application of ground penetrating radar to estimate subgrade soil density", Infrastruct., 5(2), 12. https://doi.org/10.3390/infrastructures5020012.
  2. Al-Nuaimy W., Huang, Y., Nakhkash, M., Fang, M.T.C., Nguyen, V.T. and Eriksen, A. (2000), "Automatic detection of buried utilities and solid objects with GPR using neural networks and pattern recognition", J. Appl. Geophys., 43(2-4), 157-165. https://doi.org/10.1016/S0926-9851(99)00055-5.
  3. Amaghani, D., Mamou, A., Maraveas, C., Roussis, P., Siorikis, V., Skentou, A. and Asteris, P. (2021). "Predicting the unconfined compressive strength of granite using only two non-destructive test indexes", Geomech. Eng., 25(4), 317-330. https://doi.org/10.12989/gae.2021.25.4.317.
  4. Benedetto A. and Pensa, S. (2007), "Indirect diagnosis of pavement structural damages using surface GPR reflection techniques", J. Appl. Geophys., 62(2), 107-123. https://doi.org/10.1016/j.jappgeo.2006.09.001.
  5. Chen S., Wang, H., Xu, F. and Jin, Y.Q. (2016), "Target classification using the deep convolutional networks for SAR images", IEEE T. Geosci. Remote Sens., 54(8), 4806-4817. https://doi.org/10.1109/TGRS.2016.2551720.
  6. Diamanti N., Redman, D. and Giannopoulos, A. (2010), "A study of GPR vertical crack responses in pavement using field data and numerical modelling", Proceedings of the XIII International Conference on Ground Penetrating Radar, Castello Carlo V, Lecce, Italy, June.
  7. Ding J., Chen, B., Liu, H. and Huang, M. (2016), "Convolutional neural network with data augmentation for SAR target recognition", IEEE Geosci. Remote Sens. Lett., 13(3), 364-368. https://doi.org/10.1109/LGRS.2015.2513754.
  8. Gamba P. and Lossani, S. (2000), "Neural detection of pipe signatures in ground penetrating radar images", IEEE T. Geosci. Remote Sens., 38(2), 790-797. https://doi.org/10.1109/36.842008.
  9. Girshick R., Donahue, J., Darrell, T. and Malik, J. (2014), "Rich feature hierarchies for accurate object detection and semantic segmentation", Proceedings of the 2014 IEEE Conference on Computer Vision and Pattern Recognition, Columbus, OH, USA, June.
  10. Goodfellow I.J., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A. and Bengio, Y. (2014), "Generative adversarial networks", arXiv:1406.2661.
  11. Karras T., Aittala, M., Hellsten, J., Laine, S., Lehtinen, J. and Aila, T. (2020), "Training generative adversarial networks with limited data", arXiv:2006.06676.
  12. Karras T., Aittala, M., Laine, S., Harkonen, E., Hellsten, J., Lehtinen, J. and Aila, T. (2021), "Alias-free generative adversarial networks", arXiv:2106.12423
  13. Karras T., Laine, S. and Aila, T. (2018), "A style-based generator architecture for generative adversarial networks.", arXiv:1812.04948.
  14. Kingma D.P. and Welling, M. (2013), "Auto-encoding variational bayes", arXiv preprint arXiv:13126114.
  15. Kong, S.M., Kim, D.M., Lee, D.Y., Jung, H.S. and Lee, Y.J. (2018), "Field and laboratory assessment of ground subsidence", Geomech. Eng., 16(3), 285-293. https://doi.org/10.12989/gae.2018.16.3.285.
  16. Krizhevsky, A., Sutskever, I. and Hinton, G.E. (2012), "ImageNet classification with deep convolutional neural networks", Proc. Advances in Neural Information Processing Systems, 25, 1090-1098.
  17. Kwon, S.Y., Yoo, M. and Hong, S. (2020), "Earthquake risk assessment of underground railway station by fragility analysis based on numerical simulation", Geomech. Eng., 21(2), 143-152. https://doi.org/10.12989/gae.2020.21.2.143.
  18. Lee, K.H., Park, J.H., Park, J., Lee, I.M. and Lee, S.W. (2022), "Experimental verification for prediction method of anomaly ahead of tunnel face by using electrical resistivity tomography", Geomech. Eng., 20(6), 475-484. https://doi.org/10.12989/gae.2020.20.6.475.
  19. Li, J., Gu, J., Huang, Z. and Wen, J. (2019), "Application research of improved YOLO V3 algorithm in PCB electronic component detection.", Appl. Sci., 9, 3750.
  20. Mirza M. and Osindero, S. (2014), "Conditional generative adversarial nets", arXiv:1411.1784.
  21. Saarenketo T. and Scullion, T. (2000), "Road evaluation with ground penetrating radar", J. Appl. Geophys., 43(2-4), 119-138. https://doi.org/10.1016/S0926-9851(99)00052-X.
  22. Shaw, M.R., Millard, S.G., Molyneaux, T.C.K., Taylor, M.J. and Bungey, J.H. (2005), "Location of steel reinforcement in concrete using ground penetrating radar and neural networks", NDT & E Int., 38(3), 203-212. https://doi.org/10.1016/j.ndteint.2004.06.011.
  23. Sudyka J., Krysinski, L., Zofka, A., Mechowski, T. and Harasim, P. (2018), "Identification of deep-rooted transverse cracks using Ground Penetrating Radar", IOP Conf. Ser.: Mater. Sci. Eng., 356, 012022.
  24. Wang, C.Y., Yeh, I.H. and Liao, H.Y.M. (2021), "You only learn one representation: Unified network for multiple tasks", arXiv preprint arXiv:2105.04206.