Deep Learning-Based Detection for Traffic Control
Yiou Yang · 2021
Often urban intersections have a problem with long queues, but the traffic flow during the green light is relatively small. In this paper, we propose an integrated, machine-vision control mode process for controlling traffic lights. The method proposed in this paper is to obtain real-time video images of the road through the monitoring camera, detect the number of vehicles in the different areas, and dynamically control the traffic signal. Our real-time system fundamentally guarantees the smooth traffic flow through the intersection to avoid frequent traffic-congestion.