DOI QR코드

DOI QR Code

기계학습 기반 회절파 분리 적용을 통한 GPR 탐사 자료의 도로 하부 공동 및 구조물 탐지 성능 향상

Improvement of Underground Cavity and Structure Detection Performance Through Machine Learning-based Diffraction Separation of GPR Data

  • 김수윤 (한양대학교 자원환경공학과) ;
  • 변중무 (한양대학교 자원환경공학과)
  • Sooyoon Kim (Department of Earth Resources and Environmental Engineering, Hanyang University) ;
  • Joongmoo Byun (Department of Earth Resources and Environmental Engineering, Hanyang University)
  • 투고 : 2023.08.16
  • 심사 : 2023.09.27
  • 발행 : 2023.11.30

초록

최근 도심지 도로에서 빈번하게 발생하는 도로 파임의 주원인인 지하 공동의 발생을 파악하기 위해, 차량 부착형 지표투과레이더(GPR)를 통해 얻은 대량의 취득 자료를 효율적으로 처리하기 위한 기계학습 기반 공동 탐지 기술이 활발하게 연구되고 있다. 그러나 기계학습 자료 생성 시 단순한 영상 처리 기법들만 활용되고 있고, 탄성파 탐사나 GPR 자료 처리에 시도되었던 여러 기법들은 충분히 활용되지 못하고 있다. 이 연구에서는 지하 공동의 탐지가 대부분 회절파의 탐지에 의해 이루어진다는 점에 착안하여 GPR 자료로부터 회절파를 분리하여 YOLO v5 모델을 이용한 도로 하부 공동 탐지 모델의 성능을 향상시켰다. 탄성파에서 개발된 기계학습 기반 회절파 분리 기법을 GPR 자료에 맞게 변형한 후, GPR 현장 자료에서 회절파를 분리하여 공동 탐지 모델의 입력으로 사용하였다. 서울시 공공 개방 GPR 자료를 이용하여 제안된 방법의 성능을 검증한 결과, 회절파 분리를 이용했을 때 더 정확하게 공동 및 지하 구조물을 탐지하는 것을 확인하였다. 또한 제안된 회절파 분리 기법은 향후 GPR 탐사가 이용되는 다양한 분야에서 활용될 수 있을 것으로 기대된다.

Machine learning (ML)-based cavity detection using a large amount of survey data obtained from vehicle-mounted ground penetrating radar (GPR) has been actively studied to identify underground cavities. However, only simple image processing techniques have been used for preprocessing the ML input, and many conventional seismic and GPR data processing techniques, which have been used for decades, have not been fully exploited. In this study, based on the idea that a cavity can be identified using diffraction, we applied ML-based diffraction separation to GPR data to increase the accuracy of cavity detection using the YOLO v5 model. The original ML-based seismic diffraction separation technique was modified, and the separated diffraction image was used as the input to train the cavity detection model. The performance of the proposed method was verified using public GPR data released by the Seoul Metropolitan Government. Underground cavities and objects were more accurately detected using separated diffraction images. In the future, the proposed method can be useful in various fields in which GPR surveys are used.

키워드

과제정보

이 논문은 2022년도 정부(교육부, 산업통상자원부)의 재원으로 K-CCUS 추진단의 지원을 받아 수행된 연구입니다(KCCUS20220001, 온실가스 감축 혁신인재양성사업). 또한, 이 연구를 위해 귀중한 현장 자료를 제공해주신 한국지능정보사회진흥원의 AI 허브와 서울시 도로관리과에 감사를 드립니다.

참고문헌

  1. AI Hub 3D GPR data, 2023, https://aihub.or.kr/ (June 20, 2023 Accessed)
  2. Arosio, D., 2016, Rock fracture characterization with GPR by means of deterministic deconvolution, J. Appl. Geophy., 126, 27-34, https://doi.org/10.1016/j.jappgeo.2016.01.006
  3. Baker, G. S., Steeples, D. W., Schmeissner, C., Pavlovic, M., and Plumb, R., 2001, Near-surface imaging using coincident seismic and GPR data, Geophys. Res. Lett., 28(4), 627-630, https://doi.org/10.1029/2000GL008538
  4. Benedetto, A., Tosti, F., Ciampoli, L. B., and D'amico, F., 2017, An overview of ground-penetrating radar signal processing techniques for road inspections, Signal processing, 132, 201-209, https://doi.org/10.1016/j.sigpro.2016.05.016
  5. Chae, J., Ko, H.-y., Lee, B.-g., and Kim, N., 2019, A study on the pipe position estimation in GPR images using deep learning based convolutional neural network, Journal of Internet Computing and Services, 20(4), 39-46 (in Korean with English abstract), https://doi.org/10.7472/jksii.2019.20.4.39
  6. Chlaib, H. K., Mahdi, H., Al-Shukri, H., Su, M. M., Catakli, A., and Abd, N., 2014, Using ground penetrating radar in levee assessment to detect small scale animal burrows, J. Appl. Geophy., 103, 121-131, https://doi.org/10.1016/j.jappgeo.2014.01.011
  7. Choi, B., Pyun, S., Choi, W., Jo, C.-h., and Yoon, J., 2022, Deep-learning-based GPR Data Interpretation Technique for Detecting Cavities in Urban Roads, Geophys. and Geophys. Explor., 25(4), 189-200 (in Korean with English abstract), https://doi.org/10.7582/GGE.2022.25.4.189
  8. Choo, J.-H., Yoo, C.-K., Oh, Y.-C., and Lee, I.-M., 2019, Assessment of NATM tunnel lining thickness and its behind state utilizing GPR survey, Journal of Korean Tunnelling and Underground Space Association, 21(5), 717-733 (in Korean with English abstract), https://doi.org/10.9711/KTAJ.2019.21.5.717
  9. De Figueiredo, J., Oliveira, F., Esmi, E., Freitas, L., Schleicher, J., Novais, A., Sussner, P., and Green, S., 2013, Automatic detection and imaging of diffraction points using pattern recognition, Geophys. Prospect., 61, 368-379, https://doi.org/10.1111/j.1365-2478.2012.01123.x
  10. Gao, Y., Pei, L., Wang, S., and Li, W., 2021, Intelligent detection of urban road underground targets by using ground penetrating radar based on deep learning, Journal of Physics: Conference Series, expanded abstracts, 1757(1), 012081, https://doi.org/10.1088/1742-6596/1757/1/012081
  11. Girshick, R., Donahue, J., Darrell, T., and Malik, J., 2014, Rich feature hierarchies for accurate object detection and semantic segmentation, IEEE conference on computer vision and pattern recognition, expanded abstracts, 580-587, https://doi.org/10.48550/arXiv.1311.2524
  12. Hong, G., Kim, S. M., and Park, J. J., 2022, A Study on the Calculation of Cavity Filling Amount Using Ground Penetrating Radar and Cavity Shaping Equipment, Journal of the Society of Disaster Information, 18(2), 261-268 (in Korean with English abstract), https://doi.org/10.15683/kosdi.2022.6.30.261
  13. Huang, H., Lin, L., Tong, R., Hu, H., Zhang, Q., Iwamoto, Y., Han, X., Chen, Y.-W., and Wu, J., 2020, Unet 3+: A full-scale connected unet for medical image segmentation, ICASSP 2020-2020 IEEE international conference on acoustics, speech and signal processing (ICASSP), expanded abstracts, 1055-1059, https://doi.org/10.1109/ICASSP40776.2020.9053405
  14. Jeong, D., Yu, H., Cha, J., Seoyun, K., Ahn, H., Chi, Y., Lee, G., Noh, M., and Park, E., 2022, A Study on Building Ground Penetrating Radar Data and Artificial Intelligence-based Cavity Detection, Proceedings of the Korea Software Congress 2022, expanded abstracts, 300-302, https://www.dbpia.co.kr/journal/articleDetail?nodeId=NODE11224084
  15. Kang, M.-S., Kim, N., Lee, J. J., and An, Y.-K., 2020, Deep learning-based automated underground cavity detection using three-dimensional ground penetrating radar, Structural Health Monitoring, 19(1), 173-185, https://doi.org/10.1177/1475921719838081
  16. Kim, B., and Seol, S. J., 2017, Application of image processing techniques to GPR data for the reliability improvement in subsurface void analysis, Geophys. and Geophys. Explor., 20(2), 61-71 (in Korean with English abstract), http://doi.org/10.7582/GGE.2017.20.2.061
  17. Kim, N., Kim, S., An, Y.-K., and Lee, J.-J., 2021, A novel 3D GPR image arrangement for deep learning-based underground object classification, International Journal of Pavement Engineering, 22(6), 740-751, https://doi.org/10.1080/10298436.2019.1645846
  18. Kim, S., Seol, S. J., Byun, J., and Oh, S., 2022, Extraction of diffractions from seismic data using convolutional U-net and transfer learning, Geophysics, 87(2), V117-V129, https://doi.org/10.1190/geo2020-0847.1
  19. Lee, S. Y., Song, K.-I., Kang, K. N., and Ryu, H. H., 2022, Comparison of performance of automatic detection model of GPR signal considering the heterogeneous ground, Journal of Korean Tunnelling and Underground Space Association, 24(4), 341-353 (in Korean with English abstract), https://doi.org/10.9711/KTAJ.2022.24.4.341
  20. Li, F., Yang, F., Qiao, X., Xing, W., Zhou, C., and Xing, H., 2023, 3D ground penetrating radar cavity identification algorithm for urban roads using transfer learning, Measurement Science and Technology, 34(5), 055106, https://doi.org/10.1088/1361-6501/acb6e3
  21. Lin, T.-Y., Dollar, P., Girshick, R., He, K., Hariharan, B., and Belongie, S., 2017, Feature pyramid networks for object detection, Proceedings of the IEEE conference on computer vision and pattern recognition, expanded abstracts, 2117-2125, https://doi.org/10.48550/arXiv.1612.03144
  22. Liu, S., Qi, L., Qin, H., Shi, J., and Jia, J., 2018, Path aggregation network for instance segmentation, IEEE conference on computer vision and pattern recognition, expanded abstracts, 8759-8768, https://openaccess.thecvf.com/content_cvpr_2018/papers/Liu_Path_Aggregation_Network_CVPR_2018_paper.pdf
  23. Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C.-Y., and Berg, A. C., 2016, Ssd: Single shot multibox detector, Computer Vision-ECCV 2016: 14th European Conference, expanded abstracts, 21-37, https://doi.org/10.1007/978-3-319-46448-0_2
  24. Ma, G. H., Lee, J. W., Oh, H. J., Lee, S., and Kim, K., 2022, Evaluation of Concrete Bridge Deck Condition Using 3D GPR System, International Journal of Highway Engineering, 24(1), 9-15 (in Korean with English abstract), https://doi.org/10.7855/IJHE.2022.24.1.009
  25. Mehta, R., Fazeel, A., Rama, P., Danner, M., Bajcinca, N., Riedel, P.-B., and Schwabe, J., 2021, CNN-Based SubSurface Object Detection Using Ground Penetrating Radar, 11th International Workshop on Advanced Ground Penetrating Radar (IWAGPR), expanded abstracts, 1-5, https://doi.org/10.1109/IWAGPR50767.2021.9843163
  26. Oh, H., Kwon, M., and Jang, H., 2019, GPR Analysis on Underground Features and Foundation Structure of Cheomseongdae, Gyeongju, Korean Journal of Cultural Heritage Studies, 52(4), 264-271 (in Korean with English abstract), https://doi.org/10.22755/KJCHS.2019.52.4.264
  27. Ozkaya, U., Melgani, F., Bejiga, M. B., Seyfi, L., and Donelli, M., 2020, GPR B scan image analysis with deep learning methods, Measurement, 165, 107770, https://doi.org/10.1016/j.measurement.2020.107770
  28. Pan, S. J., and Yang, Q., 2009, A survey on transfer learning, IEEE Transactions on knowledge and data engineering, 22(10), 1345-1359, https://doi.org/10.1109/TKDE.2009.191
  29. Redmon, J., Divvala, S., Girshick, R., and Farhadi, A., 2016, You only look once: Unified, real-time object detection, Proceedings of the IEEE conference on computer vision and pattern recognition, expanded abstracts, 779-788, https://doi.org/10.48550/arXiv.1506.02640
  30. Schwarz, B., 2019, Coherent wavefield subtraction for diffraction separation, Geophysics, 84(3), V157-V168, https://doi.org/10.1190/geo2018-0368.1
  31. Son, J.-W., Moon, G.-S., and Kim, Y., 2021, Automatic Detection System of Underground Pipe Using 3D GPR Exploration Data and Deep Convolutional Neural Networks, Journal of the Korea Society of Computer and Information, 26(2), 27-37 (in Korean with English abstract), http://doi.org/10.9708/jksci.2021.26.02.027
  32. Ultralytics, https://github.com/ultralytics/yolov5/ (June 30, 2023 Accessed).
  33. Wang, C.-Y., Liao, H.-Y. M., Wu, Y.-H., Chen, P.-Y., Hsieh, J.-W., and Yeh, I.-H., 2020, CSPNet: A new backbone that can enhance learning capability of CNN, IEEE/CVF conference on computer vision and pattern recognition workshops, expanded abstracts, 390-391, https://doi.org/10.48550/arXiv.1911.11929
  34. Yoon, J.-S., Baek, J., Choi, Y. W., Choi, H., and Lee, C. M., 2016, Signal pattern analysis of ground penetrating radar for detecting road cavities, International Journal of Highway Engineering, 18(6), 61-67 (in Korean with English abstract), https://doi.org/10.7855/IJHE.2016.18.6.061
  35. Yue, G., Liu, C., Li, Y., Du, Y., and Guo, S., 2022, GPR Data Augmentation Methods by Incorporating Domain Knowledge, Applied Sciences, 12(21), 10896, https://doi.org/10.3390/app122110896
  36. Zhu, J., Zhao, D., and Luo, X., 2023, Evaluating the optimised YOLO-based defect detection method for subsurface diagnosis with ground penetrating radar, Road Materials and Pavement Design, 1-18, https://doi.org/10.1080/14680629.2023.2199880