计算机工程与应用 ›› 2020, Vol. 56 ›› Issue (17): 210-216.DOI: 10.3778/j.issn.1002-8331.1909-0274

• 图形图像处理 • 上一篇    下一篇

无人机单目视觉AOA三维定位方法

熊航,杨任农,梁晓龙,侯岳奇,王维佳   

  1. 空军工程大学 空管领航学院,西安 710051
  • 出版日期:2020-09-01 发布日期:2020-08-31

3D AOA Localization Using UAVs Monocular Vision

XIONG Hang, YANG Rennong, LIANG Xiaolong, HOU Yueqi, WANG Weijia   

  1. College of Air Traffic Control and Navigation, Air Force Engineering University, Xi’an 710051, China
  • Online:2020-09-01 Published:2020-08-31

摘要:

针对目前无人机单目视觉定位研究中对先验背景信息依赖性强、环境变化导致精度波动等问题,提出基于到达角(AOA)的无人机分时异地单目视觉三维定位方法,无人机搭载单目CCD相机依次飞行至目
标周围多个位置进行目标成像和数据采集,并基于AOA原理完成对目标三维坐标的估计。建立单目视觉和AOA三维定位量测模型,依次采用二维伪线性估计和三维加权最小二乘算法与非等角分布的量测点进行AOA
三维定位并设计任务流程,通过算法仿真与实验,在无环境与先验信息的条件下实现对目标的三维定位,验证方法有效性。

关键词: 无人机, 单目视觉, 到达角(AOA), 三维定位, 伪线性估计, 加权最小二乘

Abstract:

A method of UAV monocular vision localization at different points of time and places is presented based on AOA, aiming at the problems caused by high dependence on prior
background information and fluctuation in accuracy due to varied environment in recent studies on UAV monocular vision localization. Equipped with a monocular CCD camera, the UAV flies
to multiple positions around the target successively for target imaging and data collection. Then the 3D coordinate estimation of the target is completed based on the principle of AOA.
Measurement model is established for monocular vision and AOA 3D localization. Pseudolinear Estimator(PLE), Weighted Least-Squares(WLS) and Non-Uniform Angular Arrays(Non-UAA) of
measuring points are presented for AOA 3D localization, under which the task process is designed. The proposed method has been proved to be valid according to successful 3D positioning
of targets achieved by algorithm simulation and experimental verification without environmental or prior information.

Key words: Unmanned Aerial Vehicle(UAV), monocular vision, Angle of Arrival(AOA), 3D localization, Pseudolinear Estimator(PLE), Weighted Least-Squares(WLS)