Traffic Infraction and Alert System for Two-wheelers using Deep Learning and YOLO v3

In a country like India with a large population, traffic laws are frequently broken. These offences' accidents result in significant loss of life and property. Since bikes are used more frequently than other vehicles, there are more accidents involving bikes than with other types of vehicles. Motorcycle helmet use is one of the major contributors to this. So, using real-time surveillance cameras, we suggest a method dubbed "Traffic Infraction and Alert System for Two-wheelers using Deep Learning and YOLO v3" using deep learning which automatically delivers challans or SMSs to people when bicycle riders are identified as being triple riders and riding without a helmet. The suggested method uses object segmentation and backdrop subtraction to initially identify motorbike riders. At that point, we classify violations using an object classifier. Our main goal is to lessen the risk of injury in case of an accident because wearing a helmet while driving is essential. We can therefore improve the safety of motorcycle riders on the road by identifying those who are riding three at once or otherwise breaking the law. By automating, we can lessen the effort on the traffic control crew and effectively share the evidence with them so they can fine offenders.