空间碎片跟踪图像序列的目标识别方法  

Object Recognition Method of Space Debris Tracking Image Sequence

作　　者：张晨[1,2] 平一鼎[1,3] 

机构地区：[1]中国科学院紫金山天文台,南京210008 [2]中国科学院大学,北京100049 [3]中国科学院空间目标与碎片观测重点实验室,南京210008

出　　处：《天文学报》2015年第5期506-515,共10页Acta Astronomica Sinica

基　　金：国家自然科学基金项目(11373071)资助

摘　　要：自动化观测已成为目前空间碎片光学观测的发展趋势,相应的完全无人工干预的目标自动识别成为迫切需要研究的课题.开环跟踪依据历史根数引导望远镜进行碎片的跟踪观测,是一种实现简单并且很稳健的观测方法.针对开环跟踪获取的观测图像,开展了目标在像素空间内的点列特征分析,提出使用像点簇识别算法来进行目标的自动识别,并对3种不同的簇识别算法实现进行了比较.In order to strengthen the capability of the space debris researches, auto- mated optical observation becomes more and more popular. Thus, the fully unattended automated object recognition framework is urgently needed to be studied. On the oth- er hand, the open loop tracking which guides the telescope only with historical orbital elements is a simple and robust way to track space debris. According to the analysis of point distribution characteristics in pixel domain of object＇s open loop tracking image sequence, the Cluster Identification Method is introduced into automated space debris recognition method. With the comparison of three algorithm implements, it is shown that this method is totally available in actual research work.

关 键 词：天体测量 技术 图像处理 望远镜 方法 观测 数据分析 

分 类 号：P123[天文地球—天体测量]