摘要
为了降低交通隐患,提高交通系统的可靠性与安全性,保证车辆检测的实时性和准确性。采用了深度学习领域的目标检测YOLOv8算法,该算法是基于YOLOv5的进一步提升与改进,在Backbone部分使用了梯度流更加丰富的C2f结构,Head部分换成了目前主流的解耦头结构,损失函数使用了Task-Aligned Assigner正负样本匹配方式以及其他一些改变。将车辆划分成7种类型的数据集为例进行车辆检测,首先将数据集进行拆分和归一化,然后调整相关的参数,最后训练模型。实验结果表明,YOLOv8算法在该实验数据集的车辆类型识别上mAP达到了95%,该方法在车辆类型的检测中结果较好,能够应用在实际的交通系统中。
In order to reduce traffic hazards,improve the reliability and safety of the transportation system,and ensure the real-time and accuracy of vehicle detection.This article adopts the YOLOv8 algorithm for object detection in the field of deep learning,which is further enhanced and improved based on YOLOv5.In the Backbone section,a C2f structure with richer gradient flow is used,the Head section is replaced with the current mainstream decoupling head structure,and the loss function uses Task Aligned Assignor positive and negative sample matching method,among other changes.Taking the dataset of dividing vehicles into 7 types as an example for vehicle detection,the dataset is first split and normalized,relevant parameters are adjusted,and finally the model is trained.The experimental results show that the YOLOv8 algorithm achieves 95%mAP in vehicle type recognition on the experimental dataset.This method has good results in vehicle type detection and can be applied in practical transportation systems.
作者
马明扬
宋淑彩
赵一航
张博
李鸿鑫
MA Mingyang;SONG Shucai;ZHAO Yihang;ZHANG Bo;LI Hongxin(Hebei University of Architecture,Zhangjiakou,Hebei 075000)
出处
《河北建筑工程学院学报》
CAS
2024年第1期247-252,共6页
Journal of Hebei Institute of Architecture and Civil Engineering
