Vehicle and Plate Detection for Intelligent Transport Systems: Performance Evaluation of Models YOLOv5 and YOLOv8

Intelligent transport systems aim to enhance efficiency and safety in urban mobility, employing technologies like computer vision to detect vehicles and license plates in images and footage. Regression-based algorithms such as you only look once (YOLO) can be applied in this context. Hence, this work assesses the performance of the YOLOv5 and YOLOv8 models in automatically detecting vehicle and license plates. The training and validation processes involved a curated dataset obtained through transfer learning techniques to enhance the quality and quantity of images, encompassing various locations and lighting conditions to ensure data diversity and representativeness. Confusion matrix analysis revealed that the YOLOv8 model slightly outperformed YOLOv5, with an accuracy of around 97.98% and a precision rating of 97.19%. In addition, the training time for YOLOv8 was lower than YOLOv5 based on the context.