HELPR: Helmet Enforcement and License Plate Reconnaissance Using YOLOv8 and EasyOCR for Real-Time Traffic Surveillance

Traffic accidents that result in death of people are mainly caused by motorcycle accidents, especially in those areas where wearing of helmets is not a mandatory practice. To solve the problem of manual surveillance by the traffic authorities, this paper suggests an automated, real-time possession to identify the helmet violation and motorcycle number plates on the roadside surveillance streams. The framework incorporates the YOLOv8 object detector model to localize riders, helmets, and number plates, and uses the EasyOCR to do strong optical character recognition on plate regions extracted under a variety of lighting and traffic conditions. Experimental results on controlled datasets and on real-world video footage show that the system can be characterized by a high detection rate, achieving 93% accuracy in helmet detection, 96% accuracy in rider detection, and 88% accuracy in number plate detection, whereas EasyOCR gets a maximum of 96% accuracy in character-level OCR. a highquality text extraction rate, and a high generalization ability to a variety of complex traffic scenes, which means that the system has an opportunity to provide scalable, data-driven traffic enforcement within the smart city environment.