基于遥感图像的无固定标志点卫星自主定轨  

Autonomous Orbit Determination Based on Remote Sensing Image without Fixed Landmarks

作　　者：高有涛[1] 刘靖雨 王兆龙 李木子 孙俊 GAO Youtao;LIU Jingyu;WANG Zhaolong;LI Muzi;SUN Jun(School of Astronautics,Nanjing University of Aeronautics and Astronautics,Nanjing 210016,Jiangsu,China;Shanghai Institute of Aerospace Control Technology,Shanghai 201109,China;Shanghai Key Laboratory of Space Intelligent Control Technology,Shanghai 201109,China)

机构地区：[1]南京航空航天大学航天学院,江苏南京210016 [2]上海航天控制技术研究所,上海201109 [3]上海市空间智能控制技术重点实验室,上海201109

出　　处：《上海航天（中英文）》2021年第2期66-75,共10页Aerospace Shanghai（Chinese&English）

基　　金：国防基础科研项目(2016110C019);国家自然科学基金(61803027);上海市青年科技英才杨帆计划(19YF1420300);上海航天科技创新基金(SAST2017088)。

摘　　要：目前基于事先选取地面标志点的卫星自主导航可能会造成误匹配和漏匹配,并且不包含地标点的遥感图像也有可能存在导航信息。为了解决这两个问题,本文提出了一种不依赖事先标定地面标志点的自主导航方法,通过YOLOv3深度学习对获得的遥感图像进行目标识别缩小图像匹配区域,将识别后的目标区域进行图像匹配获得匹配特征点及其像平面坐标,最后利用成像模型计算特征点的地理坐标用于自主导航。仿真结果表明:本文方法相比于有控点精度更高,平均误差为31.9438 m,平均速度误差为0.0384 m/s。在引入测角信息与本文所提出的无地面标志点构成多源信息情况下,可以较为快速准确地实现航天器自主定轨。The satellite autonomous navigation based on pre-selected ground landmark points may cause mismatching and missing matching,and remote sensing images without landmark points may also contain navigation information.In order to solve these two problems,in this paper,an autonomous navigation method is proposed,which is independent of pre-calibrated ground landmarks.First,through YOLOv3 deep learning,the obtained remote sensing images are identified,which reduces the image matching area.Then,image matching is performed on the identified target area to obtain the feature points and their image plane coordinates.Finally,the imaging model is used to calculate the geographic coordinates of the feature points for autonomous orbit determination.The simulation results show that the proposed method has higher accuracy than the method with controlled points.The average error is 31.9438 m,and the average velocity error is 0.0384 m/s.With multi-source information including the angle measurement information and information without landmark points,the autonomous orbit determination of spacecraft can be achieved more quickly and accurately.

关 键 词：遥感图像 自主定轨 地标点 深度学习 非线性卡尔曼滤波 

分 类 号：P134[天文地球—天体力学]