Bike Authentication by Helmet ‘Using Faster RCNN Machine Learning

Abstract

Two-wheeled bikes are the most common way to get around in developing countries like India. Even though riding is easy, many accidents happen because riders don't follow the law that says they have to wear helmets. In 2017, at least 98 riders who weren't wearing helmets died every day, according to the Road Accident Report. In 2016, 10,135 people died because they didn't wear a helmet. In 2017, that number rose to 36,000. WHO has said that one reason for the rise in road accidents is that riders don't use safety equipment as they should. The government has many ways to catch people who break the law. But these methods need help from a person, which slows down the system and makes it less reliable. In this paper, the Faster RCNN machine learning algorithm is used to make a system that automatically finds helmets. The camera in the system takes a picture of the rider and sends it to the raspberry pi so it can be processed. To find the helmet, a faster RCNN algorithm is used. If the helmet is not found, the vehicle can only go 40kmph, but it can go as fast as it wants if the helmet is found.