International conference on construction engineering and project management (국제학술발표논문집)

  • 2024.07a
  • /
  • Pages.367-374
  • /
  • 2024
  • /
  • 2508-9048(eISSN)
Korea Institute of Construction Engineering and Management (한국건설관리학회)
권호 표지
DOI QR코드

DOI QR Code

DTR: A Unified Detection-Tracking-Re-identification Framework for Dynamic Worker Monitoring in Construction Sites

  • Nasrullah Khan (Department of Architectural Engineering, Chung-Ang University) ;
  • Syed Farhan Alam Zaidi (Department of Computer Science, Chung-Ang University) ;
  • Aqsa Sabir (Department of Computer Science, Chung-Ang University) ;
  • Muhammad Sibtain Abbas (Department of Architectural Engineering, Chung-Ang University) ;
  • Rahat Hussain (Department of Architectural Engineering, Chung-Ang University) ;
  • Chansik Park (Department of Architecture and Building Science, Chung-Ang University) ;
  • Dongmin Lee (Department of Architecture and Building Science, Chung-Ang University)
  • Published : 2024.07.29
Download PDF
⟨ Previous Next ⟩

Abstract

The detection and tracking of construction workers in building sites generate valuable data on unsafe behavior, work productivity, and construction progress. Many computer vision-based tracking approaches have been investigated and their capabilities for tracking construction workers have been tested. However, the dynamic nature of real-world construction environments, where workers wear similar outfits and move around in often cluttered and occluded regions, has severely limited the accuracy of these methods. Herein, to enhance the performance of vision-based tracking, a new framework is proposed which seamlessly integrates three computer vision components: detection, tracking, and re-identification (DTR). In DTR, a tracking algorithm continuously tracks identified workers using a detector and tracker in combination. Then, a re-identification model extracts visual features and utilizes them as appearance descriptors in subsequent frames during tracking. Empirical results demonstrate that the proposed method has excellent multi-object-tracking accuracy with better accuracy than an existing approach. The DTR framework can efficiently and accurately monitor workers, ensuring safer and more productive dynamic work environments.

Keywords

Acknowledgement

This work was supported by a National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT) (No. RS-2023-00217322).

References

  1. Teizer, J., T. Cheng, and Y. Fang, Location tracking and data visualization technology to advance construction ironworkers' education and training in safety and productivity. Automation in construction, 2013. 35: p. 53-68.
  2. Zhang, M., R. Shi, and Z. Yang, A critical review of vision-based occupational health and safety monitoring of construction site workers. Safety science, 2020. 126: p. 104658.
  3. Angah, O. and A.Y. Chen, Tracking multiple construction workers through deep learning and the gradient based method with re-matching based on multi-object tracking accuracy. Automation in Construction, 2020. 119: p. 103308.
  4. Teizer, J. and P.A. Vela, Personnel tracking on construction sites using video cameras. Advanced Engineering Informatics, 2009. 23(4): p. 452-462.
  5. Xu, S., et al., Computer vision techniques in construction: a critical review. Archives of Computational Methods in Engineering, 2021. 28: p. 3383-3397.
  6. Xu, Z., J. Huang, and K. Huang, A novel computer vision-based approach for monitoring safety harness use in construction. IET Image Processing, 2023. 17(4): p. 1071-1085.
  7. Zhong, B., et al., Mapping computer vision research in construction: Developments, knowledge gaps and implications for research. Automation in Construction, 2019. 107: p. 102919.
  8. Son, H. and C. Kim, Integrated worker detection and tracking for the safe operation of construction machinery. Automation in Construction, 2021. 126: p. 103670.
  9. Park, M.-W. and I. Brilakis, Continuous localization of construction workers via integration of detection and tracking. Automation in Construction, 2016. 72: p. 129-142.
  10. Brown, L., et al., Performance evaluation of surveillance systems under varying conditions. 2023.
  11. Dai, Y., et al., A survey of detection-based video multi-object tracking. Displays, 2022. 75: p. 102317.
  12. Amin, S.U., et al., Deep learning based active learning technique for data annotation and improve the overall performance of classification models. Expert Systems with Applications, 2023. 228: p. 120391.
  13. Amin, S.U., et al., An automated chest X-ray analysis for COVID-19, tuberculosis, and pneumonia employing ensemble learning approach. Biomedical Signal Processing and Control, 2024. 87: p. 105408.
  14. Neuhausen, M., J. Teizer, and M. Konig. Construction worker detection and tracking in bird's-eye view camera images. in ISARC. Proceedings of the International Symposium on Automation and Robotics in Construction. 2018. IAARC Publications.
  15. Yang, J., et al., Tracking multiple workers on construction sites using video cameras. Advanced Engineering Informatics, 2010. 24(4): p. 428-434.
  16. Konstantinou, E., J. Lasenby, and I. Brilakis, Adaptive computer vision-based 2D tracking of workers in complex environments. Automation in Construction, 2019. 103: p. 168-184.
  17. Milan, A., et al., MOT16: A benchmark for multi-object tracking. arXiv preprint arXiv:1603.00831, 2016.
  18. Ge, Z., et al., Yolox: Exceeding yolo series in 2021. arXiv preprint arXiv:2107.08430, 2021.
  19. Zhang, Y., et al. Bytetrack: Multi-object tracking by associating every detection box. in European Conference on Computer Vision. 2022. Springer.
  20. Zhou, K., et al. Omni-scale feature learning for person re-identification. in Proceedings of the IEEE/CVF international conference on computer vision. 2019.
  21. Hard-Hat. PPE-RD-v1 Dataset. Roboflow Universe [Open Source Dataset] 2023 Available from: https://universe.roboflow.com/hard-hat-ihqrk/ppe-rd-v1.
  22. developers, O.R. ONNX Runtime. 2021; Available from: https://onnxruntime.ai/.