Vehicle Component Detection for Congestion Control using Image Processing and Machine Learning

Abstract

Traffic control systems should cope with the ever increasing demand by determining the situation on the road network and by controlling traffic flows. Here, the system contains CCTV surveillance footage which captures video photography of road and transmits the footage to the server. These are mounted from the sides of roads. The system gets counted whenever any vehicle passes/come forward to the coverage area of camera on. In Adaptive Traffic Control System which receives information from vehicle such as type, components, position and speed and then it utilize to optimize the traffic signal. Microcontroller controls Copyright © 2019 Daffodil International University page | 2 the Vehicle Component Detection [1] system and counts number of vehicles passing on road. Microcontroller also store vehicles count in its memory was processed by yolov3 object classifier [2]. Based on different vehicles count, the microcontroller takes decision and updates the traffic light delays as a result. The traffic light is situated at a certain distance. Thus based on vehicle count, microcontroller defines different ranges for traffic light delays and updates those accordingly.