基于最优路径的多视场全天自主星图识别  

All-sky autonomous star recognition for multi-FOV star sensors based on optimal path

作　　者：姜笛[1,2,3] 张科 DEBEIR Olivier[2] JIANG Di;ZHANG Ke(College of Astronautics,Northwestern Polytechnical University,Xi’an 710072,China;Laboratories of Image,Signal Processing and Acoustics at UniversitéLibre de Bruxelles, Brussels1050,Belgium;National Key Laboratory of Aerospace Flight Dynamics at Northwestern Polytechnical University,Xi’an 710072,China)

机构地区：[1]西北工业大学航天学院,陕西西安710072 [2]法语布鲁塞尔自由大学图像信号处理及声学实验室 [3]航天飞行动力学国家级重点实验室,陕西西安710072

出　　处：《系统工程与电子技术》2019年第1期148-153,共6页Systems Engineering and Electronics

基　　金：海装预研项目资助课题

摘　　要：为提高低成本小视场星敏感器在全天自主星图识别模式下的识别成功率,提出一种基于最优路径特征的多视场星敏感器星图识别方法。通过多视场观测星图的星图间识别,得到融合后的星图,设置随机分布在星图中主星邻域内各星的蚁群,以迭代的方法得到遍历各星的最优路径,以路径的几何特征作为与导航星库进行匹配的依据。与现有方法相比,该方法表现出高识别率以及针对位置和亮度噪声良好的鲁棒性。Based on optimal path features and multiple fields of view(FOVs),a star recognition algorithm is developed to improve the recognition rate of star sensors with low cost and small FOV under all-sky autonomous recognition pattern.First,observed star images in multiple FOVs are fused through identification between images.Second,randomly distributed ant colonies are set at the positions of all the stars in the neighborhood of the main star.Then an optimal traversal path combining all these stars is acquired by iteration.The geometrical features of the path are applied for matching with guide star catalogue.Compared with several existing methods,the proposed algorithm provides high recognition rate and robustness against position and magnitude noise.

关 键 词：最优路径 星图识别 多视场 全天自主 

分 类 号：V448[航空宇航科学与技术—飞行器设计]