SIA OpenIR  > 光电信息技术研究室
Alternative TitleResearch on Key Visual Measurement Technology for Aerial Autonomous Refueling of UAV
Thesis Advisor郝颖明
Keyword空中加油 位姿解算 测量系统标定 视觉测量
Degree Discipline模式识别与智能系统
Degree Name硕士
Degree Grantor中国科学院沈阳自动化研究所
Place of Conferral沈阳
Abstract随着航空技术的发展,无人机已经在各个领域发挥了重大作用。为了扩大无人机的飞行半径,提升其战斗力,无人机空中自主加油的需求越来越迫切。为实现空中自主加油,必须能够时刻获得无人机与加油锥套间的相对位置和姿态。本文以无人机空中自主加油为背景,开展基于视觉的加油锥套相对位姿测量方法研究。分别选择PSD相机和可见光面阵相机作为视觉传感器,采用基于点特征的位姿求解方法,研究其具体实现技术。首先,分别针对两种视觉传感器的特点,以提高测量精度为目标,研究视觉系统的标定技术。根据PSD相机的畸变特点,提出了基于多项式畸变模型的PSD相机标定方法。该方法首先获取理想模型参数初值,然后基于最小二乘获得多项式畸变参数的初值,以重投影误差最小对标定参数进行非线性优化求精。面向可见光面阵相机的标定,探究标定摆放位置对标定精度的影响。 其次,研究特征提取和相对位姿求解技术。针对PSD相机对入射光点敏感的特点,选择频率调制的红外LED作为合作目标,PSD相机能够直接获得红外LED在PSD相机上的成像进而获得特征点坐标。而对于可见光相机来说,基于标志器上标志点的几何分布特征与约束,利用轮廓查找对目标进行识别,并通过区域生长获得特征点圆心处的成像坐标。基于获得的特征点对,对EPNP与基于LM优化求解两种位姿解算方案进行仿真实验,选择解算精度更高、输出频率满足测量需求的LM优化求解相对位姿。 基于以上的测量技术,建立基于PSD传感器与普通可见光相机的无人机空中加油视觉测量系统,开发了测量系统的标定与测量软件,在多种场景下对测量系统进行实验验证与精度测试,并对基于两种传感器的测量系统测量结果进行对比。实验结果证明,在满足空中加油视觉测量精度要求的条件下,基于PSD相机的视觉测量方案对环境适应性更好、可靠性更高更能满足无人机空中加油的视觉测量需求。
Other AbstractWith the development of aviation technology, UAV have played a major role in various fields. In order to expand the flying radius of the drone and improve its combat effectiveness, the demand for drone airborne refueling is becoming more and more urgent. In order to achieve self-fueling in the air, it is necessary to be able to obtain the relative position and posture between the drone and the fuel cone. In this paper, based on the aerial self-fueling of UAV, the research on the relative pose measurement method of the vision-based fueling cone is carried out. PSD camera and visible light area camera are selected as visual sensors respectively, and the specific method of solving the pose is studied by using the point feature based pose method. Firstly, aiming at the characteristics of two kinds of vision sensors, aiming at improving measurement accuracy, the calibration technology of vision system is studied. According to the distortion characteristics of PSD camera, a calibration method of PSD camera based on polynomial distortion model is proposed. The method first obtains the initial value of the ideal model parameters, then obtains the initial value of the polynomial distortion parameter based on the least squares, and performs nonlinear optimization and refinement on the calibration parameters with the minimum reprojection error. For the calibration of visible light area camera, the number of characteristic points of the calibration plate and the influence of the calibration position on the calibration accuracy are explored. Secondly, the feature extraction and relative pose solving techniques are studied. In view of the sensitivity of the PSD camera to the incident light spot, the frequency-modulated infrared LED is selected as the cooperation target, and the PSD camera can directly obtain the imaging of the infrared LED on the PSD camera to obtain the feature point coordinates. For the visible light camera, based on the geometric distribution features and constraints of the markers on the marker, the contour is used to identify the target, and the imaging coordinates at the center of the feature point are obtained by region growth. Based on the obtained feature point pairs, the EPNP and LM-based optimization solution are used to simulate the two pose calculation schemes. The LM optimization with higher resolution and output frequency meets the measurement requirements is used to solve the relative pose. Based on the above measurement technology, the UAV aerial refueling visual measurement system based on PSD sensor and common visible light camera is established. The calibration and measurement software of the measurement system is developed. The test system is verified and tested in various scenarios. Compare the measurement results of the measurement system based on the two sensors. The experimental results show that under the condition of meeting the accuracy requirements of aerial refueling visual measurement, the visual measurement scheme based on PSD camera has better environmental adaptability and higher reliability, and can meet the visual measurement requirements of drone aerial refueling.
Contribution Rank1
Document Type学位论文
Recommended Citation
GB/T 7714
滕玉坤. 面向无人机空中加油的视觉测量关键技术研究[D]. 沈阳. 中国科学院沈阳自动化研究所,2019.
Files in This Item:
File Name/Size DocType Version Access License
面向无人机空中加油的视觉测量关键技术研究(3640KB)学位论文 开放获取CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[滕玉坤]'s Articles
Baidu academic
Similar articles in Baidu academic
[滕玉坤]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[滕玉坤]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.