Intelligent Traffic Monitoring System Using Deep Learning: Triple Riding, Automatic License Plate Recognition, and Helmet Detection

One of the main causes of death for people is traffic accidents. Motorcycle accidents are among the many kinds of traffic accidents that frequently result in serious casualties. The motorcyclist’s first line of defense is their helmet. The current state of affairs presents a number of challenges. In Indian traffic laws, which can be resolved in a variety of ways. Since it’s against the law to ride a motorcycle without a helmet and triple riding in India, there have been more crashes and fatalities. It is highly desirable for bike riders to utilize helmets and no triple ride are solutions. Seeing the value of helmets, governments have declared riding a bike without helmet is illegal and have implemented slow-moving manual enforcement methods to apprehend offenders. The suggested method uses street-level video monitoring to automatically determine if a bike rider is wearing a helmet and triple riding without assistance from a human. Four-level deep learning object detection is the core idea at play. A person, a motorcycle at level one using YOLOv5, a helmet at level two using YOLOv5, a triple riding at level three using YOLOv5, and a license plate at level four using Web API are the things observed. After that, web automation is used to obtain the license plate registration number. When a bike rider is seen without a helmet, the license plate of the bike is read and recorded. Details of the offender will also be generated in the e-challan. The police authorities will therefore have access to a database for analysis. Records will be created in a database to enable exact tracking of each offender.