宇航学报

宇航学报 ›› 2012, Vol. 33 ›› Issue (6): 728-735.doi: 10.3873/j.issn.1000-1328.2012.06.007

• 制导、导航与控制 • 上一篇    下一篇

非合作大目标位姿测量的线结构光视觉方法

高学海1, 梁斌1,2, 潘乐2, 徐文福3   

  1. 1. 哈尔滨工业大学空间智能系统研究所，哈尔滨 150001；2. 深圳航天东方红海特卫星有限公司，深圳 518057；3. 哈尔滨工业大学深圳研究生院，深圳 518055
  • 收稿日期:2011-09-26 修回日期:2011-11-23 出版日期:2012-06-15 发布日期:2012-06-20
  • 作者简介:1000-1328（2012）06-0728-08
  • 基金资助:
    国家自然科学基金（60805033）

Pose Measurement of Large Non-Cooperative Target using Line Structured Light Vision

GAO Xue-hai1, LIANG Bin1,2, PAN Le2, XU Wen-fu3   

  1. 1. Institute of Space Intelligent System, Harbin Institute of Technology, Harbin 150001, China;2. Aerospace Dongfanghong Development Ltd. Shenzhen, Shenzhen 518057, China;3. Shenzhen Graduate School, Harbin Institute of Technology, Shenzhen 518055, China
  • Received:2011-09-26 Revised:2011-11-23 Online:2012-06-15 Published:2012-06-20

PDF

793

摘要： 空间机器人与非合作大目标交会接近最终段，单目相机不能获取完整的特征图像而无法完成相对位姿测量。针对此问题，提出基于线结构光和单目视觉的相对位姿测量方法。以非合作大目标上的局部矩形特征为测量对象，首先，建立相对位姿测量模型并给出四个测量坐标系之间的关系；其次，通过相机对不完整矩形和线结构光的约束获得四个特征点在相机坐标系下的坐标；然后，利用四个特征点计算相机坐标系与目标坐标系之间的转移矩阵；最后，将转移矩阵分解得到矩形特征的相对位姿。通过改变影响测量精度的输入误差和标定误差等因素对该方法进行仿真验证，结果表明该测量方法是有效的。

关键词: 单目相机, 线结构光, 非合作大目标, 位姿测量, 空间机器人

Abstract: In final approach of rendezvous between a space robot and a large non-cooperative target, a monocular camera of the space robot can not observe a complete feature image of the target to determine relative position and attitude. To overcome this problem, a line structured light is introduced to aid measurement. A partial rectangular framework of the target is chosen as measurement object. Firstly, a measurement model is built and four measurement coordinate systems are presented. Secondly, according to camera projection constraints on the partial rectangular framework and the line structured light, four feature points are calculated in the camera coordinate system. Thirdly, using the four feature points, transform matrix between camera coordinate system and target coordinate system is computed. Lastly, the relative position and attitude of the partial rectangular framework is derived from the transform matrix. Numerical simulations are studied to verify the method under different input errors and calibration errors. The results show that this method is effective.

Key words: Monocular camera, Line structured light, Non-cooperative large target, Pose determination, Space robot