Road Helmet Detection with YOLOv3: Ensuring Rider Safety

This work addresses the significant issue of safety helmet recognition in industrial settings using deep learning models, with a focus on YOLOv3 architecture. Since it takes a lot of resources, it is challenging to verify helmet compliance in real time, even with CCTV cameras. Several object identification models are assessed using numerous benchmark datasets, including SSD, RCNN, YOLOv3, YOLOv4, and YOLOv5. A novel method called FD-YOLOv3 M combines YOLOv3 with a fuzzy-based photo enhancement module. It expands on YOLOv3’s success. The performance of the model is enhanced by this fusion, particularly in identifying safety helmets in noisy CCTV footage and distinguishing them from other kinds of headgear. The results of the study show the benefit of the FD-YOLOv3 model above its prior versions and show that it may be applied in practical industrial settings. The study’s findings demonstrate how useful YOLOv3-based models are for automatic number plate and helmet recognition in industrial environments. By offering a stable and cost-effective solution, these models significantly contribute to guaranteeing worker safety through the early and accurate identification of safety equipment.