面向无人机自主着陆的视觉感知与位姿估计方法综述  

A Survey on Vision-based Sensing and Pose Estimation Methods for UAV Autonomous Landing

作　　者：马宁 曹云峰 MA Ning;CAO Yun-Feng(College of Astronautics,Nanjing University of Aeronautics and Astronautics,Nanjing 211106)

机构地区：[1]南京航空航天大学航天学院,南京211106

出　　处：《自动化学报》2024年第7期1284-1304,共21页Acta Automatica Sinica

基　　金：国家自然科学基金(U2033201)资助。

摘　　要：自主着陆技术是制约无人机(Unmanned aerial vehicle,UAV)自主性等级提升中极具挑战性的一项技术.立足于未来基于视觉的无人机自主着陆技术的发展需求,围绕其中的核心问题——着陆场检测与位姿估计,对近十年来国内外无人机自主着陆领域中基于视觉的着陆场检测与位姿估计方法研究进展进行总结.首先,在分析无人机自主着陆应用需求的基础上,指出机器视觉在无人机自主着陆领域的应用优势,并凝练出存在的科学问题;其次,按不同应用场景划分对着陆场检测算法进行梳理;然后,分别对纯视觉、多源信息融合的位姿估计技术研究成果进行归纳;最后,总结该领域有待进一步解决的难点,并对未来的技术发展趋势进行展望.Autonomous landing technology is the most challenging technology to restrict the improvement of unmanned aerial vehicle(UAV)autonomy level.Focusing on the core problem of autonomous landing technology of UAV——Landing site detection and pose estimation,the research progress in this field in the past decade is summarized in this paper,which is on the basis of the development demand of vision-based autonomous landing technology in the future.First of all,on the basis of analyzing the application demand for UAV autonomous landing,the advantages of the application of machine vision in the field of UAV autonomous landing are pointed out,and the existing scientific problems are summarized.Secondly,the landing site detection algorithms are sorted according to different application scenarios.Then,the research progress of pose estimation technology based on vision and multiinformation fusion are summarized,respectively.Finally,the challenges to be solved in this field are summarized,and future technical development trends are introduced.

关 键 词：自主着陆 无人机 机器视觉 着陆场检测 位姿估计 信息融合 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] V279[自动化与计算机技术—计算机科学与技术]