中国空间科学技术

中国空间科学技术 ›› 2025, Vol. 45 ›› Issue (5): 75-90.doi: 10.16708/j.cnki.1000-758X.2025.0098

• 论文 • 上一篇    下一篇

小天体表面重建技术及在自主光学导航中的应用

田启航1，刘一武2,*，王立1，闫斯畅1，林大泳1，华宝成1，梁潇1   

  1. 1.北京控制工程研究所 空间光电测量与智能感知实验室，北京100191
    2.北京控制工程研究所 空间智能控制技术国家级重点实验室，北京100191
  • 收稿日期:2024-06-02 修回日期:2024-08-26 录用日期:2024-09-13 发布日期:2025-09-17 出版日期:2025-10-01

Surface reconstruction techniques for asteroid missions and the applications in autonomous optical navigation

TIAN Qihang1，LIU Yiwu2,*，WANG Li1，YAN Sichang1，LIN Dayong1，HUA Baocheng1，LIANG Xiao1   

  1. 1.Beijing Institute of Control Engineering, Space Optoelectronic Measurement and Perception Lab，Beijing 100191,China
    2.Beijing Institute of Control Engineering, National Key Laboratory of Space Intelligent Control，Beijing 100191,China
  • Received:2024-06-02 Revision received:2024-08-26 Accepted:2024-09-13 Online:2025-09-17 Published:2025-10-01

PDF (PC)

21

摘要： 自主光学导航是飞行器实时定轨中的一项关键技术。对于深空小天体任务，由于通信时延的限制，地面可干预的制导方法能力上明显受限。在上行链路带宽有限及目标物理特性未知的情况下，该问题将会变得更加棘手。因此，星上自主的相对导航系统具有不言自明的意义。针对未来小天体探测任务中的自主光学导航问题进行了研究，提出了一种基于视觉成像数据的测量框架。该测量框架由地面支持系统及星上特征识别系统两部分组成。地面支持系统集成了运动恢复结构及立体光度法两种主要测量方式，用于不同图像分辨率下导航特征的制备工作，主要包括小天体全球模型及局部区域陆标地图。星上特征识别系统主要利用制导、导航与控制系统（GNC）提供的先验知识及地面上注的导航特征，实时生成预测图像，并与实拍图像进行配准。修正后的方位测量值可作为星上导航滤波器的输入，用于更新飞行器状态。首次提出使用ICQ模型进行在轨面目标视线跟踪，所需上注数据量仅为常规模型的20%，可实现3%目标像素直径的指向测量精度。首次应用频域掩膜NCC技术实现在轨特征识别，相比于传统标准NCC，在相同测量精度条件下，计算效率可提升约1个数量级。以实际在轨数据及仿真验证为例，对该框架的测量原理及性能进行了论述。所提出的框架可实现自主及高效的飞行器定位。

关键词: 运动恢复结构, 立体光度法, 自主光学导航, 小天体, 制导、导航与控制

Abstract: Autonomous optical navigation (OpNav) is expected to play a crucial role to aid in the in-situ determination of spacecraft trajectory. For asteroid missions, the high communication latency with Earth makes feedback guidance intractable and limited in capabilities. The situation becomes worse in the case of constrained uplink bandwidth and unknown physical properties of the target body. Therefore, the significance of an onboard target-relative navigation system is axiomatic. In this work, a novel and feasible framework of visual imagery-based measurements is proposed in support of the upcoming planetary missions. The framework consists of two segments: ground support system (GSS) and onboard feature recognition system (OFRS). In GSS, structure from motion and stereophotoclinometry are integrated for developing topographic models of both entire body and surface areas as navigation features at different image resolutions. OFRS works by rendering the expected appearance of the uploaded feature catalog, which is registered to the onboard collected image. The corrected bearing measurement can then be fed to navigation filter to update the onboard spacecraft knowledge. The application of ICQ model is first proposed for on-orbit extended-body centroid tracking. The amount of data uploaded is only 20% of the conventional model, and the pointing measurement accuracy of 3% of the target diameter in pixels can be achieved. The masked NCC technique was first applied to implement on-orbit feature recognition. It is demonstrated that the computation of masked NCC registration is as accurate as that of the standard NCC registration, while the efficiency can be improved by about 1 order of magnitude. This work describes the principle and performance of the framework, with examples from the previous small body asteroid mission and simulations. The proposed framework enables autonomous and efficient localization of spacecraft. 

Key words: structure from motion, stereophotoclinometry, autonomous optical navigation, asteroid, guidance, navigation and control