Helmet Detection using Machine Learning Approach

The safety of riders depends greatly on their wearing helmets. Any intelligent traffic system must automatically detect those who are breaking the rules of the roadway. A popular method of transportation in a nation like India, where there is a significant population density in all major cities, is the motorcycle. It has been noted that the majority of motorcycle riders choose not to wear helmets on city streets or even on highways. In most motorcycle accident situations, wearing a helmet can lower the risk of head and serious brain injuries for the riders. In this paper, a framework for helmet detection while riding is proposed. To identify riders who are not wearing helmets, a cascade classifier based on machine learning and HAAR characteristics are used. If the result is negative, the rider is informed right away so they can use helmets and ride safely. After a few additional warnings, a relay switch connected to the Raspberry Pi and the DC motor stops the two-wheeler correctly if this is ignored.The experimental results show that the evaluation results of this method are highly consistent with helmet detection and we got the experimental result of accuracy 97.6%.