High-Accuracy Dynamic Star Spot Positioning Method Using Multistar Joint Processing for Star Sensors

In highly maneuverable spacecraft missions, star spot images suffer from reduced signal-to-noise ratio (SNR) and trailing effects, both of which interfere with accurate star spot positioning. This study presents a high-accuracy dynamic star spot positioning method based on multi-star joint processing. According to the pixel distribution characteristics within star spots and the spatial correlations among multiple stars, we innovatively integrate the positioning processes of all star spots into a unified optimization framework, which enhances the centroid positioning accuracy for weak star spots. Furthermore, a novel algorithm is proposed to transform the centroid positioning of star spots into a one-dimensional signal edge detection problem. This approach effectively addresses the limitations of the traditional grayscale centroid algorithm and reduces the influence of trailing length on centroid positioning. Simulation results show that the proposed method maintains the centroid positioning error within 0.25 pixels at SNR = 1. Ground experiments demonstrate that, at 10°/s and 30°/s, the angular distance error is reduced by 39.65% and 70.47% compared to the mainstream method. These results verify the effectiveness of the proposed method for improving star spot positioning accuracy under highly dynamic conditions.