无人机视角下的车辆检测模型

随着经济发展和人们生活水平的不断提高，我国汽车数量不断增加，为了缓解日益严重的交通问题，发展智能交通系统、利用各种新技术进行交通信息采集已经成为交通领域的研究热点，其中，基于无人机的交通信息采集技术凭借着检测范围广、机动灵活、成本低的特点得到了广泛应用。无人机视角下的车辆检测算法是基于无人机的交通信息采集技术中的核心与基础，是进行交通参数提取和车辆行为分析等过程的前提。本团队提出了无人机视角下的车辆检测模型，可以在无人机视角下对生活中常见的车辆进行实时、精确的检测，并且在天气恶劣、画质模糊的情况下有很强的鲁棒性，进而为刑侦、灾难救助、环境监测等领域提供技术支持。

With the continuous growth of the economy and improvement of people's living standards, the number of automobiles in China has been steadily increasing. To address the increasingly severe traffic issues, developing intelligent transportation systems (ITS) and leveraging various new technologies for traffic information collection have emerged as a prominent research focus in the transportation domain. Among these technologies, Unmanned Aerial Vehicle (UAV)-based traffic information collection technology has been widely adopted, owing to its merits of wide detection range, high flexibility, and low cost. Vehicle detection algorithms from the UAV perspective serve as the core and foundation of UAV-based traffic information collection technology, and act as the prerequisite for subsequent processes including traffic parameter extraction and vehicle behavior analysis. Our team proposes a UAV perspective-based vehicle detection model, which can perform real-time and accurate detection of common daily vehicles from the UAV view, and demonstrates strong robustness under harsh weather conditions and blurry image scenarios, thereby providing technical support for fields such as criminal investigation, disaster rescue, and environmental monitoring.