Vehicle-Integrated Camera Based Rider Helmet Detection & Warning System

Abstract

<div class="section abstract"><div class="htmlview paragraph">Non-usage of helmets does not cause accidents but is critical for averting fatal and grievous injuries in the event of road occurrence accidents. Currently, traffic police use the helmet detection solution on surveillance videos to identify the vehicle number plate of a person who is not wearing a helmet and issue challan. But on the vehicle side, it is not yet implemented. At present, vehicles are neither equipped to issue warnings nor there are any safety measures taken to minimize the risk when the rider is not wearing a helmet. This paper suggests a passive safety system for two-wheelers that uses an integrated camera to detect if the rider is wearing a helmet or not by utilizing image processing techniques. Based on the result, if a helmet is not detected, then the vehicle can send control frames to vehicle HMI for alerts. This paper suggests two approaches to implement the solution. One is Machine Learning Model deployment, and another is OpenCV-based helmet detection. Each approach comes with certain constraints, which are discussed in this paper.</div></div>