论文部分内容阅读
快速可靠的星图识别算法是星敏感器确定姿态的关键部分。综合当前多种星图识别算法,设计了实用快速的全天星图识别算法。根据星敏感器视场筛选导航星,构建最小星库;选择星对角距为识别特征,根据多星结构重复性选择星形识别算法;考虑误差,将星对角距离散化成数组,特征匹配无需搜索,直接查找表定位。实验证明本算法是一种速度快(首次捕获<1s)、存储空间小(<1 MByte)、识别正确率高(>99.99%)、抗假目标干扰强的全天星图识别算法,方便在嵌入式芯片上运行。
A fast and reliable algorithm for star image recognition is a key part of the star sensor’s attitude determination. Combining with the current many kinds of star pattern recognition algorithms, a practical and fast all-day star pattern recognition algorithm is designed. According to the star sensor field screening navigation star, the smallest star bank was constructed; the star diagonal distance was selected as the recognition feature, the star recognition algorithm was selected according to the repetitiveness of the multi-star structure; the error was considered, the star diagonal distance was dispersed into an array, No need to search, locate the table directly. Experiments show that this algorithm is a fast all-day star recognition algorithm (<1s), less storage space (<1 MByte), high recognition accuracy (> 99.99% On-chip operation.