Article Details

Abstract

With the accelerated urbanization process and the vigorous development of the construction industry, the safety management problem at construction sites has become increasingly prominent. Traditional safety supervision methods are low in efficiency and difficult to adapt to the complex and ever-changing construction environments. To address this issue, this paper proposes a safety helmet wearing detection system for construction sites based on the YOLOv5 algorithm, which aims to automatically identify workers' safety helmet wearing status by virtue of Deep Learning technology and thus improve the intelligence level of safety management at construction sites. The system is developed with the Python language and integrates functional modules such as image detection, real-time video stream monitoring and historical video detection, with a user-friendly interface and a convenient operational experience. Through data annotation, model training and prediction, experimental results show that the system can accurately identify the safety helmet wearing status in complex construction environments and effectively improve the safety supervision level at construction sites.

Keywords

YOLOv5; safety helmet detection; data annotation; model training; model prediction

References

1. [1] Dai Yongshuang, Zhang Zhiwei A safety management system for subway construction personnel based on building information modeling technology and smart construction site concept [J]. Urban Rail Transit Research, 2024, 27 (1): 229-233
2. [2] Hao Yongqiang How to build a safe and standardized construction site in municipal engineering [J]. Yunnan Hydroelectric Power, 2024, 40 (1): 156-158
3. [3] MERCADO R J，GARCIA H L，ESPINO S C H，et al.Detection of Helmet Use in Motorcycle Drivers Using Convolutional Neural Network [J/OL].Applied Sciences，2023，13（10）：5882（2023-05-10）.https://doi.org/10.3390/app13105882.
4. [4] AHATSHAM H，DIAS M F.Deep Learning-Based Automatic Safety Helmet Detection System for Construction Safety [J/OL].Applied Sciences，2022，12（16）：8268（2022-08-16）.https://doi.org/10.3390/app12168268.
5. [5] Han Feiteng, Liu Yongqiang, Fang Yudong, etc A model for detecting the wearing status of safety helmets based on attention mechanism [J]. China Safety Production Science and Technology, 2024, 20 (8): 196-202
6. [6] Han Zhongli Detection of mining safety helmet wearing based on convolutional neural network [J]. Industrial and Mining Automation, 2024, 50 (S1): 82-87
7. [7] Ling Yingjie Research on Optimization of Video Surveillance Object Detection Algorithm Based on YOLO, SSD and Faster R-CNN [J]. Modern Computer, 2024, 30 (21): 49-53, 58
8. [8] Li Na, Jiang Jingwei Smart Training Room Detection System Based on YOLOv9+PyQt5 [J]. Journal of Luohe Vocational and Technical College, 2024, 23 (4): 25-30
9. [9] Zhang Yizhi Research on Optimization Algorithm for Safety Helmet Detection Based on YOLO [D]. Chengdu: University of Electronic Science and Technology of China, 2025
10. [10] Cheng Xiaoya, Wang Song, Dai Bin Analysis of semi-automatic labeling tools based on Labelimg [J]. Electronic Technology, 2025, 54 (8): 210-211

Funding

2025 Hubei Province College Students Innovation and Entrepreneurship Training Program Project (S202513239013)

About the authors

Yu Qin (1981-), female, Han ethnicity, from Xiantao, Hubei, associate professor, master's degree, research direction: computer application technology, machine learning; Tang Jun (2001-), male, Han ethnicity, from Quanzhou, Guangxi, bachelor's degree, research direction: Computer Science and Technology; Corresponding author: Ye Yang (1984-), female, Hui ethnic, from Wuhan, Hubei, associate professor, master's degree, research direction: data analysis.