无线电工程2025，Vol.55Issue(2)：230-245,16.DOI:10.3969/j.issn.1003-3106.2025.02.002
基于深度学习的车辆检测方法研究进展
Research Progress of Vehicle Detection Methods Based on Deep Learning
摘要
Abstract
Detection methods based on deep learning have made significant research progress in recent years.The introduction of deep learning models has made vehicle detection greatly improved in terms of accuracy and efficiency.This study reviews the research progress of vehicle detection methods based on deep learning.Firstly,the difficulties of vehicle detection are introduced and summarized.Then,it focuses on the current deep learning-based vehicle detection methods applied in different scenarios as well as those under the influence of other factors,including region-based methods,single-phase-based methods,and attention mechanism-based methods,and briefly introduces each method and analyzes the problems they can solve;it also presents the mainstream vehicle open source datasets and evaluation metrics for vehicle detection;Finally,the solved problems and difficulties to be improved of vehicle detection algorithm are summarized respectively,and future research directions are prospected,including the introduction of multimodal information,cross-temporal and spatial vehicle detection and other aspects of the research.These studies will further promote the development of deep learning-based vehicle detection technologies,making it play a greater role in practical applications.关键词
深度学习/车辆检测/计算机视觉/无锚点检测器/遥感图像/注意力机制Key words
deep learning/vehicle detection/computer vision/anchorless detectors/remote sensing images/attention mechanisms分类
信息技术与安全科学引用本文复制引用
游昊,吕文涛,叶丹,邓志江..基于深度学习的车辆检测方法研究进展[J].无线电工程,2025,55(2):230-245,16.基金项目
国家自然科学基金(U1709219,61601410) （U1709219,61601410）
浙江省科技厅重点研发计划项目(2022C01079,2024C01060)National Natural Science Foundation of China(U1709219,61601410) （2022C01079,2024C01060）
Key R&D Project of Zhejiang Provincial Department of Science and Technology(2022C01079,2024C01060) （2022C01079,2024C01060）
