Real-Time Traffic Light Management System with Manual Control

Congestion at road intersections is a widespread problem. Congestion is a result of improper management of traffic. To make this traffic management process less hectic and automated, Traffic Lights are used at intersections which are effective too. Various researches focus on making traffic lights more reliable and effective which include a) Survey System b) Induction Loops c) Proximity Sensors d) Image Classification. Every proposed technique computes the vehicle count using any technique and accordingly timing of lights is set. In this paper, a system has been proposed which primarily utilizes image classification and has three main parts: Vehicle Count using Image Classification, Decision making Algorithm and Manual Control. Real-time traffic is analysed using image processing and computed vehicle count is given as input to decision-making algorithm, in return algorithm sets the timing of green signal to a selected lane. The timing of signals can also be updated by utilising manual control unit of the system. The proposed system provides manual control to lights along with real-time analysis of traffic and performs 75% times better in comparison with the survey system.