基于数据融合的无人机自主择址技术

doi:10.12382/bgxb.2024.0751

摘要/Abstract

摘要：

无人机作为一种新型飞行器,正在逐步融入现代武器装备体系,成为军事领域中不可或缺的重要组成部分。为了使无人机具备安全的着陆决策系统,能够在没有地面标识的情况下自主地执行降落任务,提出一种基于多传感器数据融合从粗到精的分阶段自主择址技术。基于图像信息进行语义分割、实现粗糙落点搜索,在引导无人机降低飞行高度之后,基于点云信息的高程值计算地形参数、构建地形成本图,并考虑地形的类别融合图像语义信息,完成精细落点搜索。试验结果表明:该技术能够很好地划分出安全区域和危险区域,能够使无人机自主获取安全的着陆位置;在精细落点搜索阶段中通过与拟合点云平面实现决策的方式进行对比分析,验证了该技术能够较大程度地节省决策时间,提高择址效率。

关键词: 无人机, 自主择址, 数据融合, 地形成本图, 语义分割

Abstract:

As a new type of aircraft,the unmanned aerial vehicle (UAV) is gradually integrating into the modern weaponry system and becoming an indispensable and important part of the military field.In order to equip UAV with a safe landing decision-making system that can autonomously perform landing tasks without ground marking,this paper proposes a phased autonomous location selection technique based on multi-sensor data fusion from coarse to fine.The rough landing point search is realized based on the semantic segmentation of the image information.After guiding the UAV to reduce the flight altitude,the terrain parameters are calculated from the elevation value of point cloud information to construct a terrain cost map,and the semantic information of the image is fused by considering the category of a terrain to complete the fine landing point search.The experimental results show that the proposed location selection technique can well delineate the safe and dangerous areas,and enables UAVs to autonomously arrive at a safe landing position.Meanwhile,the comparative analysis of the decision-making in the fine landing point search stage with the fitted point cloud plane verifies that the technique can save decision-making time to a greater extent and improve the efficiency of location selection.

Key words: unmanned aerial vehicle, autonomous landing location selection, data fusion, terrain cost map, semantic segmentation

中图分类号:

TP391

周乐, 尹乔之, 钟沛霖, 魏小辉, 聂宏. 基于数据融合的无人机自主择址技术[J]. 兵工学报, 2025, 46(8): 240751-.

ZHOU Le, YIN Qiaozhi, ZHONG Peilin, WEI Xiaohui, NIE Hong. Autonomous UAV Location Selection Technique Based on Data Fusion[J]. Acta Armamentarii, 2025, 46(8): 240751-.

图/表 33

图1 UAV分阶段择址框架

Fig.1 Phased UAV landing location selection framework

图2 模型可视化结果(左为测试图,右为预测图)

Fig.2 Visualization results of the model(left :test images,right:predicted results)

图3 基于图像的落点搜索流程图

Fig.3 Flowchart of image-based landing point search

图4 地形成本图的构建

Fig.4 Construction of terrain cost map

图5 室外仿真环境

Fig.5 Outdoor simulation environment

图6 加装传感器的UAV模型

Fig.6 UAV model with sensors

表1 仿真中视觉传感器的性能参数

Table 1 Performance parameters of vision sensor in simulation

参数	数值
焦距/pixel	639.997649
视场角/(°)	69×42
分辨率	640×480
畸变系数	[0,0,0,0,0]
帧率/(帧·s^-1)	30

表2 仿真中激光雷达的性能参数

Table 2 Performance parameters of LiDAR in simulation

参数	取值
扫描模式	非重复式扫描
测量范围/m	190
视场角/(°)	70.4×77.2
输出/(点·次^-1)	24000
更新频率/Hz	10

图7 高空图像

Fig.7 Upper-air image

图8 高空语义图像

Fig.8 Upper-air semantic image

图9 二值图像

Fig.9 Binary image

图10 初步择址结果

Fig.10 Preliminary result of location selection

图11 低空下传感器采集的信息

Fig.11 The information collected at low altitude

图12 低空下的语义信息

Fig.12 The semantic information at low altitude

图13 基于平面拟合筛选最佳着陆点

Fig.13 The best landing point screened based on plane fitting

图14 点云高程值与地形参数网格示意图

Fig.14 The schematic diagrams of the elevation values of point cloud and the terrain parameter grid

图15 点云语义网格图

Fig.15 Point cloud semantic grid diagram

图16 地形成本图

Fig.16 Terrain cost map

图17 安全-危险网格图

Fig.17 Safe-hazard grid map

图18 安全距离图

Fig.18 Safety distance map

表3 精细落点搜索阶段的仿真结果对比

Table 3 Comparison of simulated results in the fine landing point search stage

参数	平面拟合	地形成本图
点云数量/个	225881	225881
最佳点坐标/m	(5.23,-1.07,-18.24)	(6.75,-1.35,-18.24)
运行时间/s	1.1787	0.1437

图19 具备择址系统的真实UAV

Fig.19 A real UVA with a location selection system

表4 部分传感器的详细信息

Table 4 Detailed information of some sensors

设备	类型	参数
视觉传感器	Realsense D455	RGB摄像头分辨率:640×480 视场角:90° × 65° 帧率:30帧/s
激光雷达	Livox avia	量程:190m @10%反射率扫描模式:非重复扫描视场角:70.4°×77.2° 点云输出:240000点/s

图20 场景1粗糙落点搜索

Fig.20 Scenario 1 Rough landing point search

图21 场景2粗糙落点搜索

Fig.21 Scenario 2 Rough landing point search

图22 场景3粗糙落点搜索

Fig.22 Scenario 3 Rough landing point search

图23 场景4粗糙落点搜索

Fig.23 Scenario 4 Rough landing point search

表5 粗糙落点搜索阶段的真实试验结果

Table 5 Real test results in the rough landing point search phase

场景	图像尺寸/pixel	像素坐标/pixel	时间/s
场景1	640×480	(284,344)	0.077
场景2	640×480	(224,344)	0.071
场景3	640×480	(124,264)	0.073
场景4	640×480	(504,124)	0.074

图24 场景1精细落点搜索

Fig.24 Scenario 1 Fine landing point search

图25 场景2精细落点搜索

Fig.25 Scenario 2 Fine landing point search

图26 场景3精细落点搜索

Fig.26 Scenario 3 Fine landing point search

图27 场景4精细落点搜索

Fig.27 Scenario 4 Fine landing point search

表6 精细落点搜索阶段的真实试验结果对比

Table 6 Comparison of real test results in the fine landing point search phase

场景	方式	点云坐标/m	时间/ s	时间对比/%
场景1 (105 394个点)	平面拟合	(4.36,-6.15, -22.26)	0.491	72.30
场景1 (105 394个点)	地形成本图	(4.25,-6.55, -22.26)	0.136	72.30
场景2 (84 640个点)	平面拟合	(1.26,-3.74, -19.19)	0.465	72.36
场景2 (84 640个点)	地形成本图	(1.15,-3.85, -19.21)	0.129	72.36
场景3 (89798个点)	平面拟合	(-6.43,-4.94, -20.52)	0.447	73.60
场景3 (89798个点)	地形成本图	(-6.45,-5.65, -20.54)	0.118	73.60
场景4 (84416个点)	平面拟合	(4.48,1.96, -19.51)	0.419	75.42
场景4 (84416个点)	地形成本图	(4.25,2.75, -19.52)	0.103	75.42

参考文献 30

[1]	刘闯, 鱼小军, 张婷, 等. 无人集群装备仿真试验关键技术现状及趋势[J]. 航空学报, 2022, 43(增刊1):726919.
	LIU C, YU X J, ZHANG T, et al. Research status and trend of key technologies for simulation test of unmanned swarm equipment[J]. Acta Aeronautica et Astronautica Sinica, 2022, 43(S1):726919. (in Chinese) doi: 10.7527/S1000-6893.2022.26919
[2]	NI X L, YIN Q Z, WEI X H, et al. Research on landing stability of four-legged adaptive landing gear for multirotor UAVs[J]. Aerospace, 2022, 9(12):776.
[3]	孙浩, 尹乔之, 魏小辉, 等. 新型自适应起落架的单支腿落震性能研究[J]. 北京航空航天大学学报, 2023, 49(4):990-998.
	SUN H, YIN Q Z, WEI X H, et al. Research of single leg drop performance of new adaptive landing gear[J]. Journal of Beijing University of Aeronautics and Astronautics, 2023, 49(4):990-998. (in Chinese)
[4]	王盼, 吴昊, 梁宇, 等. 轻型无人机起落架设计与强度分析[J]. 兵工学报, 2022, 43(增刊1):140-145.
	WANG P, WU H, LIANG Y, et al. Design and strength snalysis of light UAV landing gear[J]. Acta Armamentarii, 2022, 43(S1):140-145. (in Chinese)
[5]	LEE M, SHIN S G, JANG S, et al. Visual-based landing guidance system of UAV with deep learning technique for environments of visual-detection impairment[J]. International Journal of Control,Automation and Systems, 2022, 20(5):1735-1744.
[6]	马宁, 曹云峰. 面向无人机自主着陆的视觉感知与位姿估计方法综述[J]. 自动化学报, 2024, 50(7):1284-1304.
	MA N, CAO Y F. A survey on vision-based sensing and pose estimation methods for UAV autonomous landing[J]. Acta Automatica Sinica, 2024, 50(7):1284-1304. (in Chinese)
[7]	LEE M F R, NUGROHO A, LE T T, et al. Landing area recognition using deep learning for unammaned aerial vehicles[C]// Proceedings of 2020 International Conference on Advanced Robotics and Intelligent Systems. Taipei,Taiwan,China: IEEE, 2020:1-6.
[8]	DEHSHIBI M M, FAHIMI M S, MASHHADI M. Vision-based site selection for emergency landing of UAVs[C]// Recent Advances in Information and Communication Technology 2015:Proceedings of the 11st International Conference on Computing and Information Technology. Berlin,Germany: Springer, 2015:133-142.
[9]	袁俊. 基于无人机平台的目标跟踪和着陆位姿估计[D]. 哈尔滨: 哈尔滨工业大学, 2020.
	YUAN J. Target tracking and landing pose estimation based on UAV platform[D]. Harbin: Harbin Institute of Technology, 2020. (in Chinese)
[10]	黄建宇, 屈玉福, 姜吉祥. 无人飞行器自主降落区识别方法研究[J]. 液晶与显示, 2018, 33(3):238-244.
	HUANG J Y, QU Y F, JIANG J X. Method for identifying the landing area of unmanned aerial vehicle[J]. Chinese Journal of Liquid Crystals & Display, 2018, 33(3):238-244. (in Chinese)
[11]	ALAM M S, OLUOCH J. A survey of safe landing zone detection techniques for autonomous unmanned aerial vehicles (UAVs)[J]. Expert Systems with Applications, 2021, 179:115091.
[12]	LOUREIRO G, DIAS A, MARTINS A, et al. Emergency landing spot detection algorithm for unmanned aerial vehicles[J]. Remote Sensing, 2021, 13(10):1930.
[13]	SINGH D P, YADAV M. Suitable landing site selection for unmanned aerial vehicles using airborne laser scanning point cloud[J]. The International Archives of the Photogrammetry,Remote Sensing and Spatial Information Sciences, 2024, 48:343-348.
[14]	YANG L J, WANG L P. An optimization-based selection approach of landing sites for swarm unmanned aerial vehicles in unknown environments[J]. Expert Systems with Applications, 2022, 204:117582.
[15]	CHEN L J, YUAN X M, XIAO Y, et al. Robust autonomous landing of UAV in non-cooperative environments based on dynamic time camera-LiDAR fusion:arXiv:2011.13761[R]. Ithaca,NY,US: Cornell University, 2020:2011.13761.
[16]	SERRANO N. A bayesian framework for landing site selection during autonomous spacecraft descent[C]// Proceedings of the 2006 IEEE/RSJ International Conference on Intelligent Robots and Systems. Beijing,China: IEEE, 2006.
[17]	刘飞, 单佳瑶, 熊彬宇, 等. 基于多传感器融合的无人机可降落区域识别方法研究[J]. 航空科学技术, 2022, 33(4):19-27.
	LIU F, SHAN J Y, XIONG B Y, et al. Research on the identification method of UAV landing area based on multi-sensor fusion[J]. Aeronautical Science and Technology, 2022, 33(4):19-27. (in Chinese)
[18]	张炯, 牛欢, 陈雪. 基于多传感器融合的无人机应急着陆功能研究[J]. 航空科学技术, 2022, 33(11):11-20.
	ZHANG J, NIU H, CHEN X. Research on UAV emergency landing function based on multi-sensor fusion[J]. Aeronautical Science and Technology, 2022, 33(11):11-20. (in Chinese)
[19]	FANKHAUSER P, HUTTER M. A universal grid map library:Implementation and use case for rough terrain navigation[J]. Studies in Computational Intelligence, 2016, 625(5):99-120.
[20]	CHEN L C, PAPANDREOU G, SCHROFF F, et al. Rethinking atrous convolution for semantic image segmentation:arXiv:1706.05587[R]. Ithaca,NY,US: Cornell University, 2017:1706.05587.
[21]	CHEN L C, ZHU Y, PAPANDREOU G, et al. Encoder-decoder with atrous separable convolution for semantic image segmentation[C]// Proceedings of the European Conference on Computer Vision. Munich,Germany: Springer, 2018:801-818.
[22]	HINZMANN T, STASTNY T, CADENA C, et al. Free LSD:Prior-free visual landing site detection for autonomous planes[J]. IEEE Robotics and Automation Letters, 2018, 3(3):2545-2552.
[23]	张泽旭, 王卫东, 崔平远, 等. 一种行星软着陆地形风险评估方法[J]. 哈尔滨工业大学学报, 2011, 43(5):25-29.
	ZHANG Z X, WANG W D, CUI P Y, et al. An algorithm of terrain hazard assessment for planetary soft landing[J]. Journal of Harbin Institute of Technology, 2011, 43(5):25-29. (in Chinese)
[24]	WARREN M, MEJIAS L, YANG X, et al. Enabling aircraft emergency landings using active visual site detection[C]// Proceedings of the 9th International Conference on Field and Service Robotics. Berlin,Germany: Springer, 2015:167-181.
[25]	YANG L J, WANG C L, WANG L P. Autonomous UAVs landing site selection from point cloud in unknown environments[J]. ISA Transactions, 2022, 130:610-628. doi: 10.1016/j.isatra.2022.04.005 pmid: 35697539
[26]	CUI P Y, GE D T, GAO A. Optimal landing site selection based on safety index during planetary descent[J]. Acta Astronautica, 2017, 132:326-336.
[27]	FELZENSZWALB P F, HUTTENLOCHER D P. Distance transforms of sampled functions[J]. Theory of Computing, 2012, 8(1):415-428.
[28]	闫璐. 基于机载激光雷达的旋翼无人机自主择址方法研究[D]. 天津: 天津大学, 2020.
	YAN L. Research on autonomous landing site selection method of rotor UAV based on LiDAR[D]. Tianjin: Tianjin University, 2020. (in Chinese)
[29]	LIU F, SHAN J Y, XIONG B Y, et al. A real-time and multi-sensor-based landing area recognition system for uavs[J]. Drones, 2022, 6(5):118.
[30]	FISCHLER M A, BOLLES R C. Random sample consensus:a paradigm for model fitting with applications to image analysis and automated cartography[J]. Communications of the ACM, 1981, 24(6):381-395.