摘要
在导航星表的建立过程中 ,由于恒星的数量太多 ,往往要进行筛选 ,通常这种选择复杂费时 ,而结果往往并不是最优的。本文引入了动态星等阈值分布函数 ,将传统星等阈值过滤算法中的静态阈值用动态星等阈值代替 ,建立了一种新的动态星等阈值过滤选择模式。而基于统计学习理论的支持向量机方法为求解高维非线性动态星等阈值分布函数提供了新的途径。本文讨论了这种基于支持向量机的导航星自动选择算法——回归选取算法。实验表明 ,用该算法所选取的导航星表 ,导航星数量少、分布均匀性好。同时它还能适应多种任务的导航星选取要求 ,具有很强的通用性。
The selection of the guide star is a crucial part of an advanced star tracker design because the performance and reliability of star pattern recognition and attitude determination depend on the guide star selection. To generate a desirable mission catalog, a novel method of the automatic selection of guide star, the regression selection algorithm, was presented in this paper. With the creation of distribution function of the dynamic visual magnitude threshold (VMT), the static VMT in the traditional visual magnitude filtering method was replaced by the dynamic VMT. The high dimensional and nonlinear dynamic VMT distribution function was easily solved by the support vector machines (SVM), which was based on the statistical learning theory. The experiment results demonstrate that the guide star catalog identified by the proposed algorithm has a lot of advantages including fewer total numbers, smaller catalog size and better distribution uniformity. As a powerful method, it also flexibly meets the different tasks.
出处
《宇航学报》
EI
CAS
CSCD
北大核心
2004年第1期35-40,共6页
Journal of Astronautics
基金
"十五"民用航天项目的资助 (2 0 0 2 0 112 )
关键词
导航星星库
动态星等阈值
星等阈值分布函数
回归选取算法
支持向量机
Guide star catalog
Dynamic visual magnitude threshold
Visual magnitude threshold function
Regression selection algorithm
Support vector machines
