Helmet Detection on Two Wheeler Riders using Machine Learning and Automatic Licence Plate Recognition for Identification

Abstract

To ensure safety measures on road, detection of traffic offenders is a highly desirable but a very challenging task due to various difficulties such as closure, lighting, low video surveillance, various weather conditions, etc. this violation is a challenge due to the population and the low level of access caused mainly by the lack of an automatic system for detecting violations and taking necessary action. The growing number of people and the increasing number of vehicles make it impossible for manual systems to prevent this problem. The latest developments in Deep Learning and Image Processing provide an opportunity to solve this problem. This manuscript introduces the implementation of the three-component system which is a car, the non-use of a helmet and the number of the vehicle being monitored using Tensorflow. In-depth learning using SSD MobileNet V2 is the main method used to use the system. In this paper, we present a framework for automatic detection of motorcycle riders who drive barefoot in surveillance videos.