[paper Id: 28037] a Comprehensive Survey on Helmet Detection Techniques
Automatic helmet detection systems are crucial because fatal head injuries from motorcycle and workplace accidents are often caused by riders and workers not wearing helmets. Traditional monitoring methods fail when faced with real-world challenges like occlusions, poor lighting, and varied helmet types. Deep learning models, particularly the YOLO family, CNNs, and Transformer-based architectures, have revolutionized the field, creating highaccuracy, real-time systems essential for safety. This survey systematically reviews these advancements, covering detection models like YOLOv3 through YOLOv11, classification techniques (including hybrid models), and generative models (like GANs for data augmentation). The paper also benchmarks state-of-the-art models and discusses practical applications in smart cities and industrial safety. The focus is identifying open research problems and proposing solutions for next-generation helmet detection systems.