Real-Time Helmet Compliance Detection for Motorcycle Riders Using YOLO
Deployment on Low-Cost Traffic Cameras
Keywords:
object detection, YOLO, traffic safety, helmet detection, deep learningAbstract
Helmet non-compliance remains a major factor in motorcycle fatalities, yet manual enforcement covers only a fraction of traffic. This paper presents a helmet compliance detector based on the YOLO family of single-stage object detectors, trained on 14,200 annotated frames from intersection cameras with varied lighting and weather. The best model reached a mean average precision of 0.89 at an IoU threshold of 0.5 and ran at 27 frames per second on an embedded GPU module. Most errors involved riders partly hidden by other vehicles or wearing caps that resemble helmets. A four-week pilot at two intersections produced aggregate compliance statistics that local authorities used to plan targeted safety campaigns. The system is designed to report anonymous counts rather than identify individual riders.
References
Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016). You only look once: Unified, real-time object detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 779–788.
He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 770–778.
LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436–444. https://doi.org/10.1038/nature14539
Published
Issue
Section
License
Copyright (c) 2026 The Authors
This work is licensed under a Creative Commons Attribution 4.0 International License.
