Original Article
Abstract
References
Information
Purpose: This paper proposes a method detecting the falling of a maintenance worker in the underground utility tunnel, by applying deep learning techniques using CCTV video, and evaluates the applicability of the proposed method to the worker monitoring of the utility tunnel. Method: Each rule was designed to detect the falling of a maintenance worker by using the inference results from pre-trained YOLOv5 and OpenPose models, respectively. The rules were then integrally applied to detect worker falls within the utility tunnel. Result: Although the worker presence and falling were detected by the proposed model, the inference results were dependent on both the distance between the worker and CCTV and the falling direction of the worker. Additionally, the falling detection system using YOLOv5 shows superior performance, due to its lower dependence on distance and fall direction, compared to the OpenPose-based. Consequently, results from the fall detection using the integrated dual deep learning model were dependent on the YOLOv5 detection performance. Conclusion: The proposed hybrid model shows detecting an abnormal worker in the utility tunnel but the improvement of the model was meaningless compared to the single model based YOLOv5 due to severe differences in detection performance between each deep learning model
연구목적: 본 논문은 CCTV 영상을 활용한 딥러닝 객체 인식 기술을 적용해 지하공동구 내 쓰러진 관리인력의 검출 방법을 제시하고, 제안 방법의 관리인력 모니터링 적용성을 평가한다. 연구방법: 사람 검출 목적으로 사전 훈련된 YOLOv5와 OpenPose 모델의 추론 결과로부터 쓰러짐을 판별할 수 있는 규칙을 제안하고, 각 모델의 결과를 통합해 지하공동구 내 작업자 쓰러짐 검출에 적용하였다. 연구결과: 제안된 모델로 작업인력의 감지 및 쓰러짐을 판단할 수 있었으나, CCTV와 작업자 간격 및 작업자가 쓰러진 방향에 의존해 검출성능이 영향을 받았다. 또한 지하공동구 작업자에 대해 YOLOv5 기반 쓰러짐 판별 규칙 적용 모델이 거리 및 쓰러짐 방향 의존성이 낮아 OpenPose 기반 모델에 비해 우수한 성능을 보였다. 그 결과 통합된 이중 딥러닝 모델의 쓰러짐 검출 결과는 YOLOv5 결과에 종속되었다. 결론: 제안모델을 통해 지하공동구 작업자의 이상상황 검출이 가능함을 보였으나, 개별 딥러닝 모델별 사람 감지 성능 차이로 인해 YOLOv5 기반 모델 대비 통합 모델의 쓰러짐 검출 성능 개선은 미미하였다.
- Cao, Z., Hidalgo, G., Simon, T., Wei, S.-E., Sheikh, Y. (2017). "OpenPose: Real time multi-person 2D pose estimation using part affinity fields." In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, US, arXiv:1812.08008, pp. 7291-7299. 10.1109/CVPR.2017.143 29416696 PMC5787440
- Choi, S., Youm, S. (2019). "A study on a fall detection monitoring system for falling elderly using open source hardware." Multimedia Tools and Applications, Vol. 78, pp. 28423-28434. 10.1007/s11042-017-5452-9
- Jocher, G. (2020). YOLOv5 by Ultralytics (Ver. 7.0) https://github.com/ultralytics/yolov5
- Kang, K.-W., Park, S.-Y. (2020) "The modified fall detection algorithm based on YOLO-KCF for elderly living alone care." Journal of the Korea Institute of Convergence Signal Processing, Vol. 21, No. 2, pp. 86-91. (In Korean)
- Kim, J.-H., Choi, J.-H., Park, Y.-H., Nasridinov, A. (2021). "Abnormal situation detection on surveillance video using object detection and action recognition." Journal of Korea Multimedia Society, Vol. 24, No. 2, pp. 186-198.
- Kim, K., Lee, S., Kim, S., Kim, J., Shin, D., Shin, D.-K. (2020). "Deep learning-based abnormal behavior detection system for dementia patients." Journal of Internet Computing and Services, Vol. 21, No. 3, pp. 133-144. (In Korean)
- Lee, J., Kang, T. (2019). "Fire fighters’s asphyxiation incidents during confined space rescue in Korea." Journal of Korean Society of Occupational and Environmental Hygiene, Vol. 29, No. 4, pp. 590-602. (In Korean)
- Li, Q., Stankovic, J., Hanson, A.M., Barth, T.A. (2009). "Accurate, fast fall detection using gyroscopes and accelerometer-derived posture information." In Proceeding of 16th International Workshop on Wearable and Implantable Body Sensor Networks, Berkeley, CA, US, pp. 138-143. 10.1109/BSN.2009.46
- Lim, K.M. (2021). Sejong fire station, Dramatic rescue of 2 workers suffering from gas poisoning in underground utility tunnel, Chungchung newspaper (In Korean) https://www.dailycc.net/news/articleView.html?idxno=657401
- Ministry of Employment and Labor (2023). Warning of suffocation risk in sealed space (In Korean) https://www.moel.go.kr/news/enews/report/enewsView.do?news_seq=15037
- Ministry of Land, Infrastructure and Transport (2019). Comprehensive Research for Promotion of Utility Tunnels. pp. 7-8. (In Korean)
- Park, J.-T., Han, K.-P., Park, Y.-W. (2021). "A dangerous situation recognition system using human behavior analysis." Journal of Korea Multimedia Society, Vol. 24, No. 3 pp. 345-354. (In Korean)
- Park, K.-M., Kim, S.-D., Bae, C.-O. (2020). "A study on the detection of fallen workers in shipyard using deep learning." Journal of the Korean Society , Vol. 26, No. 6, pp. 601-605. (In Korean) 10.7837/kosomes.2020.26.6.601
- Serkan, S., Park, J., Song, J., Choi, Y., Marshall, T. (2018). "Video based fall down detection with deep learning." In Proceeding of ICROS 2018 Conference, Buan, pp. 273-274. (In Korean)
- Youm, S., Kim, C., Choi, S., Kang, Y.-S. (2019). "Development of a methodology to predict and monitor emergency situations of the elderly based on obejct detection." Multimedia Tools and Applications, Vol. 78, pp. 5427-5444. 10.1007/s11042-018-6660-7
- Yun, K., Park, J., Cho, J. (2020). "Robust human pose estimation for rotation via self-supervised learning." In IEEE Access, Vol. 8, pp. 32502-32517. 10.1109/ACCESS.2020.2973390
- Publisher :The Korean Society of Disaster Information
- Publisher(Ko) :한국재난정보학회
- Journal Title :Journal of the Society of Disaster Information
- Journal Title(Ko) :한국재난정보학회논문집
- Volume : 19
- No :3
- Pages :498-509
- DOI :https://doi.org/10.15683/kosdi.2023.9.30.498
Journal of the Society of Disaster Information
