The accuracy and robustness of vehicle object detection are critical in the development of intelligent transportation systems and autonomous driving technologies. This study proposes improvements to the YOLOv5 algorithm to enhance its vehicle detection performance in complex urban road environments. By introducing the CLOU _ Loss function to optimize bounding box loss calculation, employing Mosaic data augmentation techniques-specifically the Mosaic-9 strategy-and implementing adaptive anchor box calculation, the study significantly improves the model's ability to detect small objects and enhances its adaptability. Experimental results demonstrate that the improved YOLOv5 model achieves an accuracy of 0.96 for the Van category and 0.83 for the Car category, with notable improvements in the Bus category as well. Through evaluations based on epoch visualization plots, PR curves, F1-score curves, and confidence score curves, this paper not only enhances the performance of the YOLOv5 algorithm in vehicle detection tasks but also provides robust technical support for the practical applications of intelligent transportation systems and autonomous driving technologies.
Adaptive Anchor Box Calculation and Data Augmentation Improved YOLOv5 Vehicle Object Detection Algorithm
2025-05-16
6442434 byte
Conference paper
Electronic Resource
English
Research on Improved YOLOv5 Vehicle Target Detection Algorithm in Aerial Images
DOAJ | 2024
|Mixed Traffic Flow Detection Algorithm Based on Improved YOLOv5
Springer Verlag | 2025
|