航空学报 > 2025, Vol. 46 Issue (10): 31251-031251   doi: 10.7527/S1000-6893.2024.31251

机场净空区飞鸟与无人机多源探测技术综述

陈唯实(), 牛红闯, 王鑫, 万健, 卢贤锋, 张洁, 王青斌   

  1. 中国民航科学技术研究院 机场研究所,北京 100028
  • 收稿日期:2024-09-23 修回日期:2024-10-14 接受日期:2024-11-22 出版日期:2024-12-24 发布日期:2024-11-29
  • 通讯作者: 陈唯实 E-mail:wishchen@buaa.edu.cn
  • 基金资助:
    国家重点研发计划(2023YFB2604103);国家自然科学基金(U2433211);中国民航科学技术研究院基本科研业务费项目(x242060302216)

Review on multi-source detection technologies for birds and drones in airport clearance area

Weishi CHEN(), Hongchuang NIU, Xin WANG, Jian WAN, Xianfeng LU, Jie ZHANG, Qingbin WANG   

  1. Airport Research Institute,China Academy of Civil Aviation Science and Technology,Beijing 100028,China
  • Received:2024-09-23 Revised:2024-10-14 Accepted:2024-11-22 Online:2024-12-24 Published:2024-11-29
  • Contact: Weishi CHEN E-mail:wishchen@buaa.edu.cn
  • Supported by:
    National Key Research and Development Program of China(2023YFB2604103);National Natural Science Foundation of China(U2433211);Basic Research Project of CAST(x242060302216)

摘要:

飞鸟和无人机是威胁机场净空区的两大危险源。通过梳理雷达、光电、无线电侦测、声学等主流的飞鸟与非合作无人机目标探测技术,总结国内外典型系统及目标识别分类算法的最新研究进展。在目标分类模型方面,深度学习网络已经脱颖而出,成为一个备受关注的研究方向。基于对各类传感器数据特点的分析,提出机场净空区飞鸟与无人机多源融合探测方案。该方案以雷达数据为主体、其他几类数据为辅,通过数据级、特征级和决策级的分级融合,实现飞鸟与无人机目标的多源融合探测识别。

关键词: 机场净空区, 飞鸟, 无人机, 多源融合探测, 深度学习

Abstract:

Flying birds and drones are the two major sources of hazards threatening the airport clearance area. The mainstream detection technologies for the targets of flying birds and non-cooperative Unmanned Aerial Vehicles (UAVs) are reviewed, such as detection via radar, optoelectronics, radio, and acoustics, and the latest research progress on the typical systems and target recognition and classification algorithms at home a nd abroad is also discussed. In terms of target classification models, deep learning networks have emerged as a highly regarded research direction. Based on the analysis of the characteristics of various sensor data, a multi-source fusion detection scheme for birds and drones in the airport clearance area is proposed, which is mainly based on radar data, and is supplemented by several other types of data. Multi-source fusion detection and recognition of flying birds and UAV targets is achieved through hierarchical fusion on data level, feature level, and decision level.

Key words: airport clearance area, flying bird, UAV, multi-source fusion detection, deep learning

中图分类号: