Motion Blurred Star Image Restoration Based on MEMS Gyroscope Aid and Blur Kernel Correction

Area of Science:

• Aerospace Engineering
• Image Processing
• Sensor Technology

Background:

• Motion blur significantly degrades star image quality in star sensors.
• This degradation impacts star centroid extraction and identification, reducing overall sensor performance.

Purpose of the Study:

• To develop and validate a star image restoration algorithm to mitigate motion blur effects.
• To improve the accuracy of star centroid extraction and star identification for enhanced star sensor performance.

Main Methods:

• Utilized a MEMS gyroscope for motion trajectory deduction and initial blur kernel calculation.
• Employed Delaunay triangulation for star strip structure extraction and a preconditioned conjugate gradient interior point algorithm for blur kernel correction.
• Implemented a scaled gradient projection (SGP) method for accelerated star image reconstruction.

Main Results:

• Simulated experiments confirmed the effectiveness of both blur kernel determination and image reconstruction methods.
• Real star image experiments demonstrated increased accuracy in star centroid extraction and a higher number of identified stars post-restoration.

Conclusions:

• The proposed algorithm effectively reduces motion blur in star images.
• The developed method significantly enhances the performance of star sensors by improving star detection capabilities.