Dynamic Traffic Flow Optimization with YOLO-Based Object Detection

There is a direct need for effective traffic management solutions owing to the growing urbanization and vehicular volume. Most conventional traffic control systems only use static timing to adjust the flow of vehicles, which can be manual and result in circuit delays. This paper proposed a smart traffic signal management system, where the input images are used to change the timings of signals with dynamically managed flow in real-time. This estimates the traffic density, vehicle count, lane occupancy and classifies them vehicles like cars/ buses/trucks/motorcycles with a machine learning algorithm using advanced image processing techniques. Some of these features include emergency vehicle preemption, and adaptive signal control for smooth traffic operations. It uses object detection models to interpret the state of traffic and adapt signal timings based on live conditions. These data inputs do away with fixed signal timing and sequence that make you wait longer, increasing congestion. Simulation results show that traffic throughput increased, delay decreased, and safety was improved with image-driven smart traffic management as a scalable solution for right now cities.