基于数据增广的车辆目标检测系统优化研究  

Research on Optimization of Vehicle Target Detection System Based on Data Augmentation

作　　者：任培华[1] 孙晓东 徐希炜 王智明 REN Pei-hua;SUN Xiao-dong;XU Xi-wei;WANG Zhi-ming(Weifang Vocational College,Weifang,Shandong 262737)

机构地区：[1]潍坊职业学院,山东潍坊262737

出　　处：《潍坊工程职业学院学报》2025年第1期92-95,共4页Journal of Weifang Engineering Vocational College

基　　金：2024年中国高校产学研创新基金新一代信息技术创新项目“人工智能在无人驾驶领域的关键技术研究”(2022IT226)。

摘　　要：本文基于YOLO v8预训练模型,提出一种智能交通场景下车辆目标检测的系统构建与优化方法。在进行YOLO v8模型预训练前,加入数据增广模块,模拟更复杂的场景信息,旨在优化提高模型的鲁棒性。实验结果表明,融合数据增广策略的YOLO v8模型具有更佳的问题处理能力,泛化性能显著提高,为自动驾驶和智能交通管理提供了可靠的技术保障。Based on the YOLOv8 pre-training model,this article proposes a system construction and optimization method for vehicle target detection in intelligent transportation scenarios.Before pre-training the YOLOv8 model,a data augmentation module is added to simulate more complex scene information,aiming to optimize and improve the robustness of the model.Experimental results show that the YOLOv8 model integrated with the data augmentation strategy has better problem processing capabilities and significantly improved generalization performance,providing reliable technical support for autonomous driving and intelligent traffic management.

关 键 词：车辆检测 YOLO v8 数据增广 智能交通 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]