Improving Performance Of Star Trackers: Brightness Prediction And Star Centroid Accuracy

This research presents two strategies to improve the performance of star trackers for nanosatellite applications. The first strategy is the development of a brightness prediction model with the aim of improving the star catalogue selection during the early stages of star tracker development. This brightness prediction model reduces the need for high number of calibration images, and relies on calculation of fractional responses in the Johnson-Cousins U,B,V,R,I passbands. This method shows an improvement in the photometric predictions by a brightness magnitude order of 0.4 compared to the standard visible magnitudes from reference catalogues. The second presented strategy focuses on the star centroid detection. A centroid refinement method is implemented with aim of improving the centroid accuracy while reducing the effect of random noise. The effectiveness of this method was investigated by using a set of simulated and real images, and centroid errors were found to be lower than 0.5 pixels.