Journal of the Korea institute for structural maintenance and inspection (한국구조물진단유지관리공학회 논문집)

Korea Institute for Structural Maintenance Inspection (한국구조물진단유지관리공학회)
권호 표지
DOI QR코드

DOI QR Code

Estimation of Bridge Vehicle Loading using CCTV images and Deep Learning

CCTV 영상과 딥러닝을 이용한 교량통행 차량하중 추정

  • 배숙경 ((주)대우건설 싱가포르도시철도CR108현장) ;
  • 정우영 (서울시립대학교 토목공학과) ;
  • 최수현 (한국공항공사 토목조경부) ;
  • 김병현 (University of Illinois at Urbana-Champaign 토목환경공학과) ;
  • 조수진 (서울시립대학교 토목공학과/도시빅데이터융합학과)
  • Received : 2024.03.18
  • Accepted : 2024.04.23
  • Published : 2024.06.30
Download PDF
⟨ Previous Next ⟩

Abstract

Vehicle loading is one of the main causes of bridge deterioration. Although WiM (Weigh in Motion) can be used to measure vehicle loading on a bridge, it has disadvantage of high installation and maintenance cost due to its contactness. In this study, a non-contact method is proposed to estimate the vehicle loading history of bridges using deep learning and CCTV images. The proposed method recognizes the vehicle type using an object detection deep learning model and estimates the vehicle loading based on the load-based vehicle type classification table developed using the weights of empty vehicles of major domestic vehicle models. Faster R-CNN, an object detection deep learning model, was trained using vehicle images classified by the classification table. The performance of the model is verified using images of CCTVs on actual bridges. Finally, the vehicle loading history of an actual bridge was obtained for a specific time by continuously estimating the vehicle loadings on the bridge using the proposed method.

차량 하중은 교량의 열화를 일으키는 주된 원인 중 하나이다. 현재 WiM(Weigh-in-Motion)을 사용하여 통행 차량의 하중을 측정하고 있으나, WiM은 접촉식 센서로 설치 및 유지관리 비용이 큰 단점이 있다. 본 연구에서는 딥러닝과 CCTV 영상을 이용하여 비접촉식으로 교량 통행 차량 하중 이력을 추정하는 방법을 제안하였다. 제안된 방법은 물체 탐지 딥러닝 모델을 이용하여 통행 차종을 인식하고, 해당 차량의 하중을 국내 주요 차량 모델들의 공차중량에 근거하여 작성된 하중기반 7차종 분류표에 근거하여 추정한다. 물체 탐지 딥러닝 모델로는 Faster R-CNN 모델이 사용되었으며, Faster R-CNN 모델을 7차종 분류표에 따라 구축된 영상 학습데이터를 이용하여 학습시켰다. 학습된 딥러닝 모델의 성능은 교량 CCTV로 취득한 영상을 이용하여 검증하였다. 최종적으로 실제 교량 상부에 설치된 CCTV에서 취득한 영상을 이용하여 교량을 통행중인 차량 하중을 연속으로 추정함으로써 특정 시간동안 통행 차량의 하중 이력 그래프를 획득할 수 있음을 보였다.

Keywords

Acknowledgement

이 연구는 정부(과학기술정보통신부)의 재원으로 한국연구재단의 지원을 받아 수행된 연구입니다(No. 2021R1C1C1011119).

References

  1. Ministry of Land, Infrastructure and Transport. (2019), Road bridge and tunnel status, Report, Road Bridge and Tunnel Statistics Publication No.116050, Korea (in Korean).
  2. National Assembly Research Service. (2019), Road maintenance status and tasks - focusing on road asset management, Report 37, NARS Publication ISSN 2586-5668, Korea (in Korean).
  3. Ministry of Land, Infrastructure and Transport. (1998), Development of Bridge Live and Fatigue Load Model, Report, Korea (in Korean).
  4. Oh, J. S., Park, J. Y., Kim, J. K., and Park, S. H. (2015). "A study on Weight Estimation of Moving Vehicles using Bridge Weigh-in-Motion Technique." J. of Computational Structural Engineering Institute of Korea, 28(1), 29-37 (in Korean). https://doi.org/10.7734/COSEIK.2015.28.1.29
  5. Yoo, D. G., Kyung, K. S., Lee, S. J., Lee, H. H., and Jeon J. C. (2014). "A Study on Influencing Factors in BWIM System and Its Field Applicability." J. of Korean Society of Steel Construction, 26(4), 251-262 (in Korean). https://doi.org/10.7781/kjoss.2014.26.4.251
  6. Zhou, Y., Pei, Y., Li, Z., Fang, L., Zhao, Y., and Yi, W. (2020), Vehicle weight identification system for spatiotemporal load distribution on bridges based on non-contact machine vision technology and deep learning algorithms, Measurement, 159, 107801.
  7. Shaoqing, R.., Kaiming, H., Ross, G., and Jian, S. (2017). "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks." IEEE Transactions on Pattern Analysis and Machine Intelligence, 39(6), 1137-1149. https://doi.org/10.1109/TPAMI.2016.2577031
  8. Ministry of Land, Infrastructure and Transport. (2008), 12 Types of Traffic Volume Survey Vehicle Type Classification Guide, Available at: http://www.molit.go.kr/USR/policyTarget/dtl.jsp?idx=161 (Accessed: November 1, 2023) (in Korean).
  9. Korea Transport Institute. (2013), National traffic survey and DB construction project - Car Usage Survey, Report, Publication, Korea (in Korean).
  10. Ministry of Environment. (2021), Regulations on manufacturing vehicle certification and inspection methods and procedures, Report, Publication No. 2021-28, Korea (in Korean).
  11. MATLAB&Simulink. (2023), MathWorks, Available at: https://kr.mathworks.com/help/vision/ug/get-started-with-the-image-labeler.html (November 1, 2023).
  12. Korea Commercial Truck. (2023), Truck, Available at: https://kctruck.co.kr/%EC%86%8C%ED%98%95%ED%8A%B8%EB%9F%AD (November 1, 2023) (in Korean).
  13. Robbins, H., and Monro, S. (1951). "A stochastic Approximation Method." Ann. Math. Statist., 22(3), 400-407. https://doi.org/10.1214/aoms/1177729586
  14. Kaiming, H., Xiangyu, Z., Shaoqing, R., and Jian, S. (2015), Deep Residual Learning for Image Recognition, Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, IEEE, Piscataway, 770-778.
  15. Kim, G., Son, S., Kim, M., Jeon, J., Lee, I., Cha, J., and Choi, H. (2019), Object Tracking Method using Deep Learning and Kalman Filter, J. of Broadcast Engineering, 24(3), 495-505 (in Korean).