Dynamic Traffic Signal using Image Processing

Abstract

In cities, traffic on road is increasing rapidly resulting in traffic jams and reduced traffic flow. Traffic lights play a major role in regulation and management of traffic flow at junctions. These lights are controlled by traffic light controllers (TLCs) that are programmed to assign timely directions to drivers in Red, Yellow and Green signals. Current TLC(s) are based on programmable micro-controllers, which have limitations as their signal timings function according to a program that needs to be modified manually and does not have flexibility to accommodate the functionality to modify signal timings based on variable traffic. In this paper, inspired by advanced technologies, we have studied and tested a method to improve traffic flow at squares run by traffic lights in real-time. To make TLC(s) more efficient, we are proposing a technique which controls green, yellow, red signal switching based on the real-time calculation of traffic density through capturing and processing images of all incoming lanes at any junction. This paper also focuses on the algorithm for scheduling these lights according to vehicle density on the road, thereby aiming at reducing the traffic congestion on squares which will help people reach their destinations safe and in the least time possible.