一种基于三角形几何结构的星图识别算法

提出了一种新的以星三角形为基本识别单元的几何结构星图识别算法。通过对观测误差的预处理,将该算法星三角形的匹配过程转换为一个没有误差的三角形识别问题。引入一个能部分描述三角形结构的特征量,将导航三角形的识别所需的特征匹配数量降低,而多三角形的组合有效地提高了正确识别率。仿真实验表明,该算法所需导航星库容量小,识别效率高,误识率低,实时性和鲁棒性优于传统的三角形星图识别算法。

A novel geometric structure-based autonomous star pattern identification algorithm

A novel geometric structure-based autonomous star pattern identification algorithm is presented. The matching procedure of the triplet between measured stars in the spacecraft body frame and corresponding reference stars known in the inertial frame is transformed partially into a geometric structure feature value matching problem by preprocessing the measurement error ahead of the matching procedure, since the star triplet removed the measurement error can be changed into integer triplet. The number of feature needed in the star pattern identification procedure are reduced by the introduction of the geometric structure feature. Even if the error or redundant matches exist in the triangle identification procedure, the correct identification rate is effectively improved by the combination of multi-triplets. Compared to the traditional triplet approach, the geometric structure-based star pattern recognition algorithm has a couple of advantages, including the smaller database of guide star pattern, higher rate of correct star pattern recognition, lower mismatches probability, and better real-time adaptability and robustness.