Enhancing Traffic Management Through Advanced Vehicle Detection for Congestion Prevention

Urban areas worldwide face severe traffic congestion and transportation inefficiencies due to pop-ulation growth and increased vehicle ownership. Traditional traffic management struggles with dynamic conditions, driving a shift towards advanced technologies like machine learning and computer vision for improved traffic control. This study explores YOLOv8s deep learning models for real-time traffic vehicle detection to mitigate urban congestion. Using a dataset of nearly 920 images across various vehicle classes (cars, two-wheelers, autos, buses, trucks), the YOLOv8s model (225 layers, 11.1 million parameters) achieved notable performance in identifying cars (>80% precision, [email protected] 85.8%) and two-wheelers (69% precision, [email protected]%). Challenges remain with autos, buses, and trucks, in-dicating room for improvement through techniques like data augmentation. Overall, the YOLOv8s model shows promise for enhancing urban traffic efficiency and safety through real-time vehicle detection and integration with adaptive traffic light control systems.