无先验信息的低轨空间目标航迹起始方法  

Track Initiation for Robotic Tracking of Low Earth Orbit Objects without Prior Information

作　　者：许占伟[1,2] 王歆[1,2] 

机构地区：[1]中国科学院紫金山天文台,南京210008 [2]中国科学院空间目标与碎片观测重点实验室,南京210008

出　　处：《飞行器测控学报》2015年第6期540-544,共5页Journal of Spacecraft TT&C Technology

基　　金：中国科学院国防科技创新基金(No.CXJJ-15S129);国家自然科学基金(No.11373072)

摘　　要：针对光学观测中缺乏预报信息的低轨道空间目标自动识别捕获问题,提出一种基于修正Hough变换的空间目标航迹起始方法,从而实现空间目标的自动识别。首先利用背景恒星运动规律对观测数据进行预处理,剔除大部分背景恒星。再利用观测数据的时序信息、定向参数进行Hough变换。本方法并没有采用传统的Hough变换对所有参数空间中的数据点进行投票,而是仅对参数空间中部分点投票,可以大幅减少运算量和存储量。试验表明:该方法能够有效地抑制背景恒星产生的虚假航迹,并快速准确地完成无先验信息的低轨道空间目标的航迹起始;可以在光学观测中实现空间目标的自动识别捕获,以及实时发现新目标,这为低轨道空间目标光学观测的无人值守和联网观测奠定了技术基础。Aimed at robotic identification and capture problem of LEO （Low Earth Orbit） objects without prior information in optical observation, a track initiation method based on the modified Hough transform is proposed for realization of automatic recognition of space objects. First, the method rips most of background stars according to sidereal motion and reduces the number of candidate moving objects. Then, timing and orientation is used in Hough transformation to decrease the candidate track. In Hough transformation, unlike traditional algorithms voting on all the data points, the method chooses only votes on the part of data points in parameter space. The method greatly re- duces the amount of calculation and storage. Simulation shows that the method can effectively suppress false tracks produced by the background stars and accomplish track initiation of LEO objects fast and accurately with no prior information. The method realizes automatic recognition and capture of space objects as well as realtime detection of new objects. This lays foundation for fully robotic and networking observation of LEO objects in optical observation.

关 键 词：光学观测 低地球轨道 空间目标 航迹起始 HOUGH变换 自动识别 

分 类 号：V556.5[航空宇航科学与技术—人机与环境工程]