Fast and robust star detection algorithm based on the dyadic wavelet transform

Abstract

Abstract Star detection is an important part of the spacecraft attitude determination performed by star trackers. The current advanced star detection algorithms can effectively extract star points in complex backgrounds. However, there are still two problems: first, the parameters of the star detection algorithm are not adaptive, which limits its reliability and robustness; second, the high performance and low computational time of algorithms usually cannot be achieved simultaneously. In this work, a rapid star detection algorithm utilizing dyadic wavelet transform (SDDWT) was developed based on the à Trous algorithm. The authors constructed a suitable dyadic wavelet basis and applied the dyadic wavelet transform to the star image. The maximum wavelet coefficients at a certain scale were binarized according to a threshold value to separate stars from the background. The threshold was adaptively determined according to the white Gaussian noise (WGN) power. Finally, a method for separating individual stars was developed. Experiments indicated that the proposed algorithm was able to ensure a low false alarm rate and high detection rate by adaptively threshold value. Moreover, for a star image of 1024×1024, the computational time of the SDDWT algorithm was stable within 0.3 s, which was smaller than other algorithms.

Keywords

• dyadic wavelet transform
• image processing
• star detection
• star tracker