Authors:
O. Jeba Singh, S. Rubin Bose, J. Angelin Jeba, B. Judy Flavia, M. Rehena Sulthana
Addresses:
Centre for Academic Research, Alliance University, Bengaluru, Karnataka, India. Department of Electronics and Communication Engineering, SRM Institute of Science and Technology, Ramapuram, Chennai, Tamil Nadu, India. Department of Electronics and Communication Engineering, S.A. Engineering College, Chennai, Tamil Nadu, India. Department of Computer Science and Engineering, SRM Institute of Science and Technology, Ramapuram, Chennai, Tamil Nadu, India. School of Information Technology and Engineering, Melbourne Institute of Technology, Melbourne, Victoria, Australia.
Abstract:
In the majority of developing countries, the traffic management system has emerged as a significant problem in the present day. Daily, the number of vehicles continues to rise, and with it, the number of infractions of traffic rules. It is necessary to have an automated surveillance system to efficiently monitor infractions of traffic rules. The purpose of this research is to present a method for the automatic detection of visual traffic offences by utilising a Deep Learning algorithm. In this study, moving cars that violate traffic rules are considered. Some examples of these violations include driving without a helmet and crossing a signal when the lights are red. In light of this, the vehicles that violate traffic rules are identified by the use of deep learning algorithms in three stages: (1) the detection of vehicle objects, (2) the identification of traffic violations, and (3) the recognition of the license plate of the vehicle that has violated the traffic rules. An average accuracy of 95.8% was found in the detection of vehicle objects, according to the results of the tests conducted using the public data set for Indian traffic signals. Concurrently, the method was able to obtain a precision of 98.5% in the detection of violations and 93.67% in the recognition of identification plates.
Keywords: Traffic Violation; Deep Learning; Traffic Management; Automated Detection, Region Proposal Network; Fully Convolutional Neural Network, Automatic License Plate Recognition; Radio frequency Identification.
Received: 29/09/2024, Revised: 20/12/2024, Accepted: 04/02/2025, Published: 07/06/2025
DOI: 10.64091/ATICS.2025.000134
AVE Trends in Intelligent Computing Systems, 2025 Vol. 2 No. 2 , Pages: 99-110