Smart Traffic Controller Using Machine Learning

Abstract

Traffic signals play a vital role in managing vehicular traffic, which can lead to severe congestions if mismanaged. It is designed to improve the efficiency and the effectiveness of the pre-existing system which is managed by traffic wardens. The system will be able to manage traffic independently, ensuring that there is a free flow of traffic at all times of the day. It employs two basic components, namely OpenCV which is a computer vision library which primarily deals with the video-source and preprocessing, followed by which, the feed is processed by TensorFlow Object Detection API, which detects the number of cars in a given frame of video. The system then determines the density of traffic at a given intersection and grants the signal which can clear maximum traffic for a given point in time. Implementation of such a system not only reduces the traffic by a large extent but also reduces the manpower required to manage the intersections. Thus, the implementation of an automated system is the need of the hour to ensure that the rising traffic density is effectively managed.

Keywords: Computer Vision ,Object Detection, Smart Traffic Management System, Tensorflow