An Intelligent Realtime Traffic Control Based on Vehicle Density

Abstract

These days traffic congestion is becoming a serious problem. Roads and streets are getting over crowded, mainly in large cities leading to long vehicle queue, long hour of traffic on daily basis. This can evoke the day-to-day travellers to violate or break the traffic rules which can sometimes lead to accidents. Controlling and managing traffic system is the most demanding work in present days and the tradition traffic system cannot manage it efficiently. Nowadays worldwide in many big cities Intelligent system is been used for traffic surveillance and controlling. Therefore, we proposed a traffic control and management system which will detect the movement of vehicles, identify, track and count the numbers of vehicles in the lane by analysing a real-time live video feed form the camera with the help of computer vision and after detecting, classifying and counting the numbers of vehicle on a specific lane it will control traffic light according to the set threshold value, threshold value will be based on two criteria one is the density count of the lane and other is priority of that lane. This is done by using OpenCV and (YOLOV3) You Only Look Once real-time object detection algorithm based on CNN (Convolutional Neural Network).