Detection of TwoWheeler Riders Without Helmets Using YOLOv3 and Image Classifier

  • Dr. J.D. Dorathi Jayaseeli,Jayesh Joel, Shubham Sinha, Dr. D. Malathi

Abstract

In the current scenario of the traffic regulations and other measures for the safety on roads and work places, the use of safety equipment are essential parts for precautionary measures from the safety point of view. In the present paper, a technique has been developed for the safety measures to maintain the utmost precaution on road traffic or any other places where the helmets are mandatory to use. In this current paper, the algorithms You Only Look Once version 3 (YOLOv3) and LeNet are used for development of a system that can be used for Helmet detection and recognition of two-wheeler riders. The developed technique is able to detect any type of Helmet and it is tested on several cases. In-case, the rider is found not wearing a helmet, the instance is noted and the results are prepared on the basis of data obtained. The model shows an accuracy of 74% and reaches a speed of 1 FPS using CPU.

Published
2020-05-12
How to Cite
Dr. J.D. Dorathi Jayaseeli,Jayesh Joel, Shubham Sinha, Dr. D. Malathi. (2020). Detection of TwoWheeler Riders Without Helmets Using YOLOv3 and Image Classifier. International Journal of Advanced Science and Technology, 29(7), 899 - 907. Retrieved from http://sersc.org/journals/index.php/IJAST/article/view/14941
Section
Articles