Vehicle Classification Using YOLOv8: A Deep Learning Framework for Real-Time Traffic Monitoring

Accurate and real-time vehicle classification is important for traffic control, smart transportation, and autonomous surveillance. The current research presents a new deep learning architecture based on YOLOv8, a cutting-edge object detection model, for effective vehicle detection and multiclass classification. Compared with its predecessors, such as YOLOv5 and YOLOv7, YOLOv8 brings an anchor-free decoupled head and enhanced feature extraction, a feature that makes it superior in fine-grained vehicle classification. We have trained the model on an augmented and varied sample of Roboflow's Vehicles-2 dataset, consisting of cars, trucks, buses, and motorcycles. Empirical comparisons show mAP @ 0.5 of 0.89, with F1 scores of more than 0.95 for significant classes. Our Web deployment in Flask facilitates real-time video stream detection at inference speeds of under 0.25 seconds per frame. The system is lightweight, scalable, and deployable in smart cities for traffic monitoring and law enforcement assistance.