A Robust Local Magnitude Fitting Method for Star Identification

• Published in: IEEE sensors journal Vol. 25; no. 1; pp. 824 - 834
• Main Authors: Xie, Junfeng, Li, Xiang, Wang, Xiao, Zeng, Guoqiang, Mo, Fan
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.01.2025; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Attitude determination; Classification algorithms; Feature extraction; magnitude; Navigation; Noise; Normal distribution; Parameter identification; Position measurement; relative installation; Remote sensing; Robustness; Satellite imagery; Satellite tracking; Satellites; Sensors; star identification; Star trackers; Stars; Storage capacity
• ISSN: 1530-437X; 1558-1748
• DOI: 10.1109/JSEN.2024.3487580

Star identification is the most important part of satellite attitude determination. Existing star image identification algorithms show lower robustness with an increase in the number of stars. This study proposes a method for star identification based on local magnitude fitting. First, the similarity of neighboring star images is used for denoising. Then, the Gaussian distribution is used to determine the star point range and calculate the real grayscale cumulative value (RGCV). Finally, the star magnitude fitting range is obtained using the star tracker parameters and the fitting parameters between the RGCV and the star magnitude are determined. This method is used to optimize the rotation invariant additive vector sequence algorithm in this article. The results show that this method can reduce the storage capacity by 96%, enhance the efficiency of the algorithm and achieve a recognition rate of above 98% in real-situations. Furthermore, this method can also be applied to other star identification algorithms.