一种结合角距特征的改进栅格星图识别算法

针对栅格算法易受邻域星点影响导致误匹配的问题，提出了一种结合角距特征的改进栅格星图识别算法。首先介绍了改进栅格算法的原理，其次设计了融合栅格识别模式和星角距识别模式的算法实现流程，最后开展了基于不同视角下的大视场仿真星图的算法试验验证和性能分析。结果表明，由于角距特征具有旋转不变性和不易受观测星邻域星点分布影响等特点，结合角距特征对失效观测星进行再匹配的改进栅格算法，在兼顾存储量需求小、运行速度快等优势的同时，识别率和鲁棒性也得到了提升，最高识别率可达98.88%，在位置噪声干扰以及缺失星干扰下，改进算法的识别率仍可保持在95%，说明算法鲁棒性强，具有较好的应用前景。

An improved grid algorithm based on angular distance feature for star identification

In order to solve the problem that grid algorithm is easily affected by stars in the neighborhood, which leads to the mismatching problem, an improved grid algorithm based on angular distance feature is proposed. Firstly, the principle of the improved grid algorithm is introduced. Then the algorithm flow of combining grid recognition pattern with star angular distance recognition pattern is designed. Finally, the improved algorithm is tested and verified on the simulation star map of the sensor with large field of view (FOV) in different directions. The results show that the angular distance feature is rotational invariance and not easily affected by the distribution of stars in the vicinity of the observed star, the improved grid algorithm combined with angular distance feature to make mismatching stars match again, which can improve the recognition rate and robustness while taking into account the advantages of small memory requirement and fast running speed, the highest recognition rate is 98.88%, and the recognition rate is still 95% under the interference of position noise and missing stars, indicating that the algorithm is robust and has good application prospects.