Vehicle detection is one of the most important tasks of highway UAV traffic monitoring system, but the existing vehicle detection algorithms have low accuracy for vehicle detection in small target form under UAV aerial surveillance. This paper proposes a vehicle detection algorithm for highway UAV monitoring scenarios. The YOLOv8n algorithm has been enhanced by incorporating the BiFormer module, which improves the model's capability to capture detailed information of small targets. Additionally, the FasterNet fast network has been integrated to reduce the model's parameters and compensate for any speed loss caused by the algorithm. And at the same time, The algorithm's convergence speed is accelerated by adopting ECIoU instead of the original CIoU function. This enhancement allows for faster convergence of the algorithm. The experimental results show that the improved YOLOv8n algorithm has a detection accuracy of 87.6%, which effectively improves the vehicle detection accuracy in the highway surveillance scenario.
Improved Highway Vehicle Detection Algorithm for YOLOv8n
17.11.2023
1590403 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
YOLOv7-Tiny and YOLOv8n Evaluation for Face Detection
Springer Verlag | 2024
|Lightweight and Efficient Tiny-Object Detection Based on Improved YOLOv8n for UAV Aerial Images
DOAJ | 2024
|DOAJ | 2024
|