An Intelligent Traffic Monitoring System for Non-Helmet Wearing Motorcyclists Detection
Sumisha Samuel, Saanu Reghunadh, M K Ashwin, Sandeep Sabu, Sharath S Nair, Renju Rachel Varghese · 2020 International Conference on Data Analytics for Business and Industry: Way Towards a Sustainable Economy (ICDABI) · 2020
Riding without a helmet is the most common road offense in India. Road safety rules and regulations have to be more strictly followed and enforced. Currently, police officers identify offenders with human surveillance. This existing system is very inefficient and labor-intensive. At present, all major cities are already under a large video surveillance network to keep a watch on a wide variety of threats. However, these networks are not intelligent enough to do tasks such as non-helmet wearing motorcyclist detection. Hence, we propose an intelligent traffic monitoring system, which can detect non-helmet wearing motorcyclists from video streams using a convolutional object detector in real-time. The system will detect a violation, then attempt to extract the registration number of the vehicle and then persist the violation with evidence. An office of the law can review and take appropriate actions on the violations. This system is designed to integrate into existing video surveillance networks. The convolutional object detector was able to detect non-helmet wearing motorcyclists with an accuracy of 84% on the validation set.