Helmet Detection and Monitoring For Two Wheeler Rider Using Advanced Machine Learning Based Automated Model

Abstract

Most of the two wheeler riders forget to wear helmet at the time of driving. As a result when an accident occurs, there are high chances of  head injuries to the riders which sometimes results in death. Conventional methods of detecting and monitoring the rider’s helmet during driving are limited to certain conditions like orientation of rider and the two wheeler. To overcome the limitations of conventional methods of detecting two wheeler helmet, a neural network based automated model is proposed in the present study. The proposed model can detect the helmet under varied conditions due to its properties like learning and generalization. The superiority of the proposed model over similar type of models is tested on the real time images which are acquired through automated model. The performance of the proposed model is evaluated using various parameters like user’s accuracy (UA), producer’s accuracy (PA), Kappa coefficient (KC), Overall accuracy (OA), average computational time (S).