• Skip to primary navigation
  • Skip to main content
Logo

Work Zone Safety Information Clearinghouse

Library of Resources to Improve Roadway Work Zone Safety for All Roadway Users

  • About
  • Join Listserv
  • Contact
  • Twitter
  • Facebook
  • LinkedIn
  • Work Zone Data
    • At a Glance
    • National & State Traffic Data
    • Work Zone Traffic Crash Trends and Statistics
    • Worker Fatalities and Injuries at Road Construction Sites
  • Topics of Interest
    • Commercial Motor Vehicle Safety
    • Smart Work Zones
    • Transportation Management Plans
    • Accommodating Pedestrians
    • Worker Safety and Welfare
    • Project Coordination in Work Zones
  • Training
    • Flagger
    • Online Courses
    • Toolboxes
    • FHWA Safety Grant Products
    • Certification and
      Accreditation
  • Work Zone Devices
  • Laws, Standards & Policies
    • COVID-19 Guidance
  • Public Awareness
  • Events
  • About
  • Listserv
  • Contact
  • Search
Publication

Deep Learning-Based Workers Safety Helmet Wearing Detection on Construction Sites Using Multi-Scale Features

Author/Presenter: Han, Kun; Zeng, Xiangdong
Abstract:

Wearing safety helmets can effectively protect the safety of workers on construction sites. However, workers often take off the helmets because of weak security-conscious and discomfort, then hidden dangers will be brought by this behaviour. Workers without safety helmets will suffer more injuries in accidents such as falling human body and vertical falling matter. Hence, detecting safety helmet wearing is a vital step of construction sites safety management and a safety helmet detector with high speed and accuracy is urgently needed. However, traditional manual monitor is labour intensive and methods of installing sensors on safety helmet are difficult to popularize. Therefore, this paper proposes a deep learning-based method to detect safety helmet wearing at a satisfactory accuracy with high detection speed. This method chooses YOLO v5 as the baseline, then the fourth detection scale is added to predict more bounding boxes for small objects and the attention mechanism is adopted in the backbone of the network to construct more informative features for following concatenation operations. In order to overcome the defects caused by insufficient data, targeted data augmentation and transfer learning are used. Improvements caused by every modification are discussed in this paper. Finally, the model achieves 92.2% mean average precision, up 6.3% compared to the original algorithm, and it only takes 3.0 ms to detect an image at 640×640 . These results demonstrate the robustness and feasibility of the model. Meanwhile, the size of the trained model is only 16.3 m, which means the model is easy to be deployed. At last, after obtaining a satisfactory model, a graphical user interface (GUI) is designed to make the chosen algorithm more user-friendly.

Source: IEEE Access
Volume: 10
Publication Date: 2022
Full Text URL: Link to URL
Publication Types: Books, Reports, Papers, and Research Articles
Topics: Detection and Identification; Hard Hats; Machine Learning; Worker Safety

Copyright © 2023 American Road & Transportation Builders Association (ARTBA). The National Work Zone Safety Information Clearinghouse is a project of the ARTBA Transportation Development Foundation. It is operated in cooperation with the U.S. Federal Highway Administration and Texas A&M Transportation Institute. | Copyright Statement · Privacy Policy · Disclaimer
American Road and Transportation Builders Association Transportation Development Foundation, American Road and Transportation Builders Association U.S. Department of Transportation Federal Highway Administration Texas A&M Transportation Institute