Target Detection and Tracking Techniques for Vehicles in Real-Time Scenarios Based on YOLOv8

Vehicle detection and tracking techniques in real-time scenarios based on YOLOv8 are thoroughly investigated in this study. The relatively quick “One-stage” model is used for the vehicle target detection problem., the recently released YOLOv8 algorithm by Ultralytics is adopted., and the dataset UA-DETRAC is used as the input for training and testing on the MistGPU platform. The final model exhibits greater than 90% accuracy and recall after 300 training rounds and a batch size of 64. The total number of training rounds is 300., and the batch-size is 64. The final model exceeds 90% in terms of precision., recall., etc. ., and shows strong stability. For the vehicle target tracking problem., multi-target tracking techniques and real-time tracking are generally used in real-time scenarios.