Skip to content

Traffic monitoring system using CNN

Related topic: Machine learning, Automation

May 11, 2022

Abstract

Kathmandu has seen an exponential increase in the number of publicly and privately owned vehicles over the past decade but, a lack in the advancement of technology to monitor the traffic has seen a number of rules being broken easily and flee without being fined or caught. As per current conditions, all the rules are being checked manually by the traffic personnel which is tedious with sub-par results. Our system “Traffic Monitoring System” aims to automate the process of detecting traffic rule violations by vehicles, classifying the vehicles, and identifying/capturing license plates of vehicles that violate the traffic rules.

Problem Statement

The existing traffic control and management system have developed a lot in the past decade but still lacks in a lot of aspects, especially considering the use of modern technology. The current system being used by the Nepal Traffic Police to monitor the roads follows a manual approach where a lot of human resources are being used. The results have also not been up to the mark as per the resources used as the process of tracking the vehicles has proved to be a very tedious task. As per our study, the Nepal Traffic Office has installed about 700 CCTVs inside Kathmandu Valley along with a few high-resolution cameras in the traffic-heavy places as well in order to monitor the day-to-day traffic flow inside the city. Through this, they have implemented a system to control illegal lane crossing in some places but alerting and notifying who/what broke the rule hasn’t been implemented yet.

Possible Outcome

Develop a system that will provide assistance in notifying whether the traffic rule has been broken or not. For example, accidents, lane violations, turning without indicators, etc. Our system will not only identify the vehicle but the number plate too if but is intended to track those vehicles as well once the number plate is recognized, with the help of the various CCTVs placed within the valley. Our system will not only lessen the burden currently being faced by the police but will help society and new riders making it a safe place to drive. Another advantage of the system is it’ll alert the system and save the clip and information of the vehicle when the rule has been breached, this will save time and energy of the traffic. The system will also aid in the proper utilization of traffic resources and is intended to bring a good balance between the technical and non-technical aspects of proper vehicle tracking.