Automatic Number Plate and Helmet Detection Using Yolo

Abstract

The use of a helmet in the operation of a motorcycle is a necessary safety measure that significantly
reduces the likelihood of fatal cranial injuries in the event of accidents. The automation of the helmet
identification has become a critical and comprehensive effort in intelligent supervision systems of transport to
comply with the road safety laws. This research is a strong deep architecture of learning to identify a helmet in
real time and an extraction of license plates. The technology combines descriptors of elements with a deep neural
network classifier (DNN) to distinguish between cyclists on each other and those who do not. The methodology
uses a systematic workflow that includes pre -processing images, extraction of elements using convolution
operations and classification through a trained deep neural network model. The system is trained using a marked
data set, including a diverse photo of a motorcycle rider in a real world taken in various circumstances of lighting
and surrounding circumstances to guarantee generality. After identifying offenses, the system uses algorithms to
optical character recognition (OCR) to load the vehicle license plate for possible intervention of coercive organs.
Experimental findings indicate the accuracy, accuracy and recall of the model, which is to verify the effectiveness
of the proposed strategy in various circumstances. This platform provides a scalable and efficient method for
expanding urban traffic rights and supporting road safety through automated observance monitoring.