Star sensor is a high precision spacial attitude measurement device

precision calibration is important for high precision measurement. Wide field of view star sensor calibration system is built up based on analysis on the error factors that affect measuring error of star sensor

aiming at shortcomings of traditional calibration methods such as polynomial distortion models method and direct mapping method

propose BP neural network algorithm

an optimized genetic algorithm

to improve the BP neural network algorithm. Experimental results show that the angle measurement error of single star estimation decreases from 0. 14° to 0. 0053°

the method effectively improves single star estimation precision with good stability.