凸多边形星图识别算法

为解决星敏感器中较大视场快速、可靠的星图识别,提出了以凸多边形为基元、完全不依赖于星等的星图识别算法。对给定的视场,挑选其中较亮的恒星,依其坐标排序,然后采用由平面上的点生成凸多边形的算法,就能得到唯一的、以恒星为顶点的凸多边形。为验证星图识别算法的有效性,建立了导航星数据库,其储存单元为凸多边形的边和相邻边的夹角,共有3832个边数不等的凸多边形。在CPU为33MHz 的PC104上仿真结果表明:在任意视场中,生成凸多边形的时间小于5ms,基于凸多边形的星图识别成功率高于99%,并具有较强的鲁棒性。

A convex polygon star pattern identification algorithm for star sensor

In order to solve the problem of star pattern identification quickly and reliably in a larger field for star sensor, star pattern identification algorithm that using the minimal convex polygon vertexes as the original star pattern identification and independent of magnitude entirely is proposed. By selecting the brightest stars in the given field, sequencing according to their coordinates and using an algorithm of generating convex polygon with the points on a plane, a unique convex polygon with stars as its vertexes can be obtained. Altogether 3832 convex polygons with unequal segment can be obtained in the whole celestial sphere. For demonstrating the effectiveness of the star pattern identification algorithm, a star map database for navigation is established and its storage datum is the line segments and interior angles moving clockwise from the shortest segment. The simulation results carried out on PC104 computer with a CPU of 33MHz show that the generation time of convex polygon in any field is less than 5ms. The star pattern identification rate based on convex polygon is more than 99% with stronger robustness.