A High-Accuracy Star Centroid Extraction Method Based on Kalman Filter for Multi-Exposure Imaging Star Sensors

Area of Science:

• Aerospace Engineering
• Image Processing
• Navigation Systems

Background:

• Multi-exposure imaging increases star sensor attitude update rates but introduces significant noise.
• Accurate star centroid extraction is crucial for reliable attitude determination.

Purpose of the Study:

• To develop a novel star centroid extraction method to mitigate noise in multi-exposure star images.
• To enhance the accuracy and reliability of star sensors for attitude updates.

Main Methods:

• A Kalman Filter-based method is proposed for star centroid extraction.
• Star point prediction windows are generated using a kinematic model.
• Coarse centroids are calculated within prediction windows and then filtered by individual Kalman Filters.

Main Results:

• Simulations showed coordinate errors reduced by up to 37.5% under varying noise levels.
• Experimental results indicated prediction windows captured only 0.95% of total objects, demonstrating efficiency.
• The Kalman Filter effectively filtered image noise, yielding fine centroids.

Conclusions:

• The proposed Kalman Filter-based method is feasible and effective for reducing noise in star centroid extraction.
• The method significantly improves the accuracy of star sensors, especially in noisy conditions.
• The star point prediction windows enhance the efficiency of the centroid extraction process.