Geophysics and Geophysical Exploration (지구물리와물리탐사)

Korean Society of Earth and Exploration Geophysicists (한국지구물리·물리탐사학회)
권호 표지
DOI QR코드

DOI QR Code

Deep-learning-based GPR Data Interpretation Technique for Detecting Cavities in Urban Roads

도심지 도로 지하공동 탐지를 위한 딥러닝 기반 GPR 자료 해석 기법

  • Byunghoon, Choi (Department of Energy Resources Engineering, Inha University) ;
  • Sukjoon, Pyun (Department of Energy Resources Engineering, Inha University) ;
  • Woochang, Choi (Department of Energy Resources Engineering, Inha University) ;
  • Churl-hyun, Jo (Subsurface Information Technologies, Inc.) ;
  • Jinsung, Yoon (Seoul Metropolitan Government)
  • 최병훈 (인하대학교 에너지자원공학과) ;
  • 편석준 (인하대학교 에너지자원공학과) ;
  • 최우창 (인하대학교 에너지자원공학과) ;
  • 조철현 (지하정보기술(주)) ;
  • 윤진성 (서울특별시청)
  • Received : 2022.08.03
  • Accepted : 2022.11.23
  • Published : 2022.11.30
Download PDF
⟨ Previous Next ⟩

Abstract

Ground subsidence on urban roads is a social issue that can lead to human and property damages. Therefore, it is crucial to detect underground cavities in advance and repair them. Underground cavity detection is mainly performed using ground penetrating radar (GPR) surveys. This process is time-consuming, as a massive amount of GPR data needs to be interpreted, and the results vary depending on the skills and subjectivity of experts. To address these problems, researchers have studied automation and quantification techniques for GPR data interpretation, and recent studies have focused on deep learning-based interpretation techniques. In this study, we described a hyperbolic event detection process based on deep learning for GPR data interpretation. To demonstrate this process, we implemented a series of algorithms introduced in the preexisting research step by step. First, a deep learning-based YOLOv3 object detection model was applied to automatically detect hyperbolic signals. Subsequently, only hyperbolic signals were extracted using the column-connection clustering (C3) algorithm. Finally, the horizontal locations of the underground cavities were determined using regression analysis. The hyperbolic event detection using the YOLOv3 object detection technique achieved 84% precision and a recall score of 92% based on AP50. The predicted horizontal locations of the four underground cavities were approximately 0.12 ~ 0.36 m away from their actual locations. Thus, we confirmed that the existing deep learning-based interpretation technique is reliable with regard to detecting the hyperbolic patterns indicating underground cavities.

도심지 도로에서의 지하공동 붕괴로 인한 지반침하 문제는 인명 및 재산 피해로 이어질 수 있기 때문에 이를 예방하기 위해서는 사전에 지하공동을 탐지하고 복구하는 과정이 필요하다. 지하공동 탐지는 주로 지표투과레이더(ground penetrating radar, GPR) 탐사를 통해 이루어지는데, 방대한 탐사 자료로 인해 해석에 많은 시간이 소모되고 전문가의 숙련도와 주관에 따라 해석 결과가 달라질 수 있다. 이러한 문제를 해결하기 위해 GPR 자료 해석 자동화 및 정량화 기법들이 연구되어 왔으며, 최근에는 딥러닝 기반의 해석 기법들이 많이 활용되고 있다. 이 연구에서는 딥러닝 기반의 GPR 자료해석 기법 중 쌍곡선(hyperbola) 신호를 탐지하는 과정에 대해 기존 연구에서 개발된 기법을 단계별로 실증 예제를 통해 설명하였다. 먼저, 쌍곡선 신호를 자동으로 탐지하기 위해서 딥러닝 기반 YOLOv3 객체탐지 기법을 적용했다. 다음으로는 column-connection clustering (C3) 알고리즘을 통해 쌍곡선 신호만을 추출하였고, 최종적으로 회귀분석을 통해 지하공동의 수평위치를 결정했다. YOLOv3 객체탐지 기법을 이용한 쌍곡선 신호 탐지 성능은 AP50 기준으로 정밀도 84%, 재현율 92%를 달성했다. 지하공동 수평위치 정확도는 4개 샘플에 대해 실제 위치와 약 0.12 ~ 0.36 m 정도의 차이를 보였다. 이를 통해 지하공동에 의해 나타나는 쌍곡선 신호에 대한 딥러닝 기반 탐지 기법의 적용성을 확인할 수 있었다.

Keywords

Acknowledgement

본 결과물은 환경부의 재원으로 환경산업기술원의 지중환경 오염위해관리기술개발사업의 지원을 받아 연구되었습니다 (과제번호: 2018002440005).

References

  1. Bae, Y.-S., Kim, K.-T., and Lee, S.-Y., 2017, The Road Subsidence Status and Safety Improvement Plans, J. Korea Acad.-Ind. cooperation Soc., 18(1), 545-552 (in Korean with English abstract), doi: 10.5762/kais.2017.18.1.545.
  2. Borgioli, G., Capineri, L., Falorni, P., Matucci, S., and Windsor, C. G., 2008, The detection of buried pipes from time-of-flight radar data, IEEE Trans. Geosci. Remote Sens., 46(8), 2254-2266, doi: 10.1109/TGRS.2008.917211.
  3. Capineri, L., Grande, P., and Temple, J., 1998, Advanced image-processing technique for real-time interpretation of ground-penetrating radar images, Int. J. Imaging Syst. Technol., 9(1), 51-59, doi: 10.1002/(SICI)1098-1098(1998)9:1<51::AIDIMA7>3.0.CO;2-Q
  4. Chen, H., and Cohn, A. G., 2010, Probabilistic robust hyperbola mixture model for interpreting ground penetrating radar data, 2010 Int. Joint Conf. Neural Networks (IJCNN), Expanded Abstracts, 1-8, doi: 10.1109/IJCNN.2010.5596298.
  5. Chomdee, P., Boonpoonga, A., and Prayote, A., 2014, Fast and efficient detection of buried object for GPR image, 20th AsiaPacific Conf. Commun. (APCC2014), Expanded Abstracts, 350-355, doi: 10.1109/APCC.2014.7092835.
  6. Dou, Q., Wei, L., Magee, D. R., and Cohn, A. G., 2016, Realtime hyperbola recognition and fitting in GPR data, IEEE Trans. Geosci. Remote Sens., 55(1), 51-62, doi: 10.1109/TGRS.2016.2592679.
  7. Girshick, R., 2015, Fast r-cnn. Proc. IEEE Int. Conf. Comput. Vis., Expanded Abstracts, 1440-1448, doi: 10.1109/ICCV.2015.169.
  8. He, K., Zhang, X., Ren, S., and Sun, J., 2016, Deep residual learning for image recognition, Proc. IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit., Expanded Abstracts, 770-778, doi: 10.1109/CVPR.2016.90.
  9. Hou, F., Lei, W., Li, S., and Xi, J., 2021, Deep learning-based subsurface target detection from GPR scans, IEEE Sens. J., 21(6), 8161-8171, doi: 10.1109/JSEN.2021.3050262.
  10. Kang, M. S., Kim, N., Im, S. B., Lee, J. J., and An, Y. K., 2019, 3D GPR image-based UcNet for enhancing underground cavity detectability, Remote Sens., 11(21), 2545, doi: 10.3390/rs11212545.
  11. Kim, B., Seol, S. J., and Byun, 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), doi: 10.7582/GGE.2017.20.2.061.
  12. Kim, Y. T., Kim, B., Kim, J. W., Park, H. M., and Yoon, J. S., 2016, Determining the Optimal Frequency of Ground Penetrating Radar for Detecting Voids in Pavements, Int. J. Highw. Eng., 18(2), 37-42 (in Korean with English abstract), doi: 10.7855/ijhe.2016.18.2.037.
  13. Lee, C. M., Yoon, J. S., Baek, J. E., and Kim, S. T., 2019, Evaluation of GPR application to survey utilities on sidewalks, Int. J. Highw. Eng., 21(1), 35-41 (in Korean with English abstract), doi: 10.7855/IJHE.2019.21.1.035.
  14. Lee, K. L., and Mokji, M. M., 2014, Automatic target detection in GPR images using Histogram of Oriented Gradients (HOG), 2014 2nd Int. Conf. Electronic Design (ICED), Expanded Abstracts, 181-186, doi: 10.1109/ICED.2014.7015795.
  15. Lei, W., Hou, F., Xi, J., Tan, Q., Xu, M., Jiang, X., Liu, G., and Gu, Q., 2019, Automatic hyperbola detection and fitting in GPR B-scan image, Autom. Constr., 106, 102839, doi: 10.1016/j.autcon.2019.102839.
  16. Lei, W., Luo, J., Hou, F., Xu, L., Wang, R., and Jiang, X., 2020, Underground cylindrical objects detection and diameter identification in GPR B-scans via the CNN-LSTM framework, Electronics, 9(11), 1804, doi: 10.3390/electronics9111804.
  17. Li, S., Gu, X., Xu, X., Xu, D., Zhang, T., Liu, Z., and Dong, Q., 2021, Detection of concealed cracks from ground penetrating radar images based on deep learning algorithm, Constr. Build Mater., 273, 121949, doi: 10.1016/j.conbuildmat.2020.121949.
  18. Li, Y., Zhao, Z., Luo, Y., and Qiu, Z., 2020, Real-time patternrecognition of GPR images with YOLO V3 implemented by Tensor Flow, Sensors, 20(22), 6476, doi: 10.3390/s20226476.
  19. Liu, H., Lin, C., Cui, J., Fan, L., Xie, X., and Spencer, B. F., 2020, Detection and localization of rebar in concrete by deep learning using ground penetrating radar, Autom. Constr., 118, 103279, doi: 10.1016/j.autcon.2020.103279.
  20. Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C. Y., and Berg, A. C., 2016, Ssd: Single shot multibox detector, Europ. Conf. Comput. Vis., Expanded Abstracts, 21-37, doi: 10.48550/arXiv.1512.02325.
  21. Liu, Z., Gu, X., Yang, H., Wang, L., Chen, Y., and Wang, D., 2022, Novel YOLOv3 Model With Structure and Hyperparameter Optimization for Detection of Pavement Concealed Cracks in GPR Images, IEEE Trans. Intell. Transp. Syst., 1-11, doi: 10.1109/TITS.2022.3174626.
  22. Luo, T. X., and Lai, W. W., 2020, GPR pattern recognition of shallow subsurface air voids, Tunn. Undergr. Space Technol., 99, 103355, doi: 10.1016/j.tust.2020.103355.
  23. Luo, T. X., Yuan, P., and Zhu, S., 2022, Automatic modelling of urban subsurface with ground-penetrating radar using multiagent classification method, Geo. Spat. Inf. Sci., 1-12, doi: 10.1080/10095020.2022.2040924.
  24. Merkle, D., Frey, C., and Reiterer, A., 2021, Fusion of ground penetrating radar and laser scanning for infrastructure mapping, J. Geod., 15(1), 31-45, doi: 10.1515/jag-2020-0004.
  25. Mertens, L., Persico, R., Matera, L., and Lambot, S., 2015, Automated detection of reflection hyperbolas in complex GPR images with no a priori knowledge on the medium, IEEE Trans. Geosci. Remote Sens., 54(1), 580-596, doi: 10.1109/TGRS.2015.2462727.
  26. Ni, Z. K., Zhao, D., Ye, S., and Fang, G., 2020, City road cavity detection using YOLOv3 for ground-penetrating radar,18th Int. Conf. Ground Penetrating Radar, Expanded Abstracts, 384-387, doi: 10.1190/gpr2020-100.1.
  27. Padilla, R., Netto, S. L., and Da Silva, E. A., 2020, A survey on performance metrics for object-detection algorithms, IEEE Int. Conf. on Syst., Signals and Image Process. (IWSSIP), Expanded Abstracts, 237-242, doi: 10.1109/IWSSIP48289.2020.9145130.
  28. Park, J. J., and Kim, I. D., 2020, Analysis of the under Pavement Cavity Growth Rate using Multi-Channel GPR Equipment. J. Soc. Disaster Inf., 16(1), 60-69 (in Korean with English abstract), doi: 10.15683/kosdi.2020.3.31.060.
  29. Park, S., Kim, J., Jeon, K., Kim, J., and Park, S., 2021, Improvement of gpr-based rebar diameter estimation using yolo-v3, Remote Sens., 13(10), 2011, doi: 10.3390/rs13102011.
  30. Pham, M. T., and Lefevre, S., 2018, Buried object detection from B-scan ground penetrating radar data using Faster-RCNN, IGARSS 2018-2018 IEEE Int. Geosci. and Remote Sens. Symposium, Expanded Abstracts, 6804-6807, doi: 10.1109/IGARSS.2018.8517683.
  31. Qiu, Z., Zhao, Z., Chen, S., Zeng, J., Huang, Y., and Xiang, B., 2022, Application of an Improved YOLOv5 Algorithm in Real-Time Detection of Foreign Objects by Ground Penetrating Radar, Remote Sens., 14(8), 1895, doi: 10.3390/rs14081895.
  32. Redmon, J., Divvala, S., Girshick, R., and Farhadi, A., 2016, You only look once: Unified, real-time object detection, Proc. IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit., Expanded Abstracts, 779-788, doi: 10.48550/arXiv.1506.02640.
  33. Redmon, J., and Farhadi, A., 2017, YOLO9000: better, faster, stronger, Proc. IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit., Expanded Abstracts, 7263-7271, doi: 10.48550/arXiv.1612.08242.
  34. Redmon, J., and Farhadi, A., 2018, Yolov3: An incremental improvement, arXiv preprint, arXiv:1804.02767, doi: 10.48550/arXiv.1804.02767.
  35. Ren, S., He, K., Girshick, R., and Sun, J., 2015, Faster r-cnn: Towards real-time object detection with region proposal networks, Adv. Neural Inf. Process. Syst., 28, doi: 10.48550/arXiv.1506.01497.
  36. Rezatofighi, H., Tsoi, N., Gwak, J., Sadeghian, A., Reid, I., and Savarese, S., 2019, Generalized intersection over union: A metric and a loss for bounding box regression, Proc. IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit., Expanded Abstracts, 658-666, doi: 10.1109/CVPR.2019.00075.
  37. Thitimakorn, T., Kampananon, N., Jongjaiwanichkit, N., and Kupongsak, S., 2016, Subsurface void detection under the road surface using ground penetrating radar (GPR), a case study in the Bangkok metropolitan area, Thailand, Int. J. Geotech. Eng., 7(1), 1-9, doi: 10.1186/s40703-016-0017-8.
  38. Tomita, H., Tada, H., Nanbu, T., Chou, K., Nakamura, T., and McGregor, T., 1995, Nature and detection of void-induced pavement failures, Transp. Res. Rec., 1505, 9-16, http://onlinepubs.trb.org/Onlinepubs/trr/1995/1505/1505-002.pdf
  39. Wahab, W. A., Jaafar, J., Yassin, I. M., and Ibrahim, M. R., 2013, Interpretation of Ground Penetrating Radar (GPR) image for detecting and estimating buried pipes and cables, 2013 IEEE Int. Conf. Control Syst., Computing and Engineering, Expanded Abstracts, 361-364, doi: 10.1109/ICCSCE.2013.6719990.
  40. 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. Int. J. Highw. Eng., 18(6), 61-67 (in Korean with English abstract), doi: 10.7855/IJHE.2016.18.6.061.
  41. Zhou, X., Chen, H., and Li, J., 2018, An automatic GPR B-scan image interpreting model. IEEE Trans. Geosci. Remote Sens., 56(6), 3398-3412, doi: 10.1109/TGRS.2018.2799586.
  42. Zou, Z., Shi, Z., Guo, Y., and Ye, J., 2019, Object detection in 20 years: A survey, arXiv preprint, arXiv:1905.05055, doi: 10.48550/arXiv.1905.05055.