Fine Guidance Sensor Processing And Optical Closed-loop Semi-physical Simulation In Space Telescope

As an important experimental observation instrument for studying frontier physics problems such as the origin of the universe,stellar evolution and dark matter,space astronomical telescopes need to maintain long-term gaze on the target to obtain highquality astronomical observations.Because of the complex working environment of telescopes,and the stability of the line of sight(LOS)is affected by various internal and external disturbance factors.Therefore,a precision image stabilization system is designed to keep the telescope's LOS stable.The principle of detecting the LOS disturbance of the precision image stabilization system is to separately image the star by adding a Fine Guidance Sensor(FGS)to the edge field of the telescope.Using the principle that the stellar celestial position is constant in a short period of time,the relative attitude change of the LOS is calculated by the deviation of the star point extracted from different frames.At the same time,Long focal length primary optical system makes the FGS have the advantage of ultra-high angular resolution,and can calculate the high-precision absolute attitude of the telescope's LOS pointing through the star identification.Based on this background,this paper mainly studies the key techniques of Star centroiding,LOS relative attitude solution and star identification,and builds an experimental verification system to verify the proposed algorithm.The specific research work is as follows:On the basis of investigating the precision image system of foreign space telescope,According to the research objectives and contents formulated,the performance parameters and key technologies of the FGS are analyzed,and the basic functional components of the guide star detection system are proposed.The influence of system error and noise on the precision of the star centroiding of the FGS is studied,and the limit precision of the attitude measurement of the fine star guide is estimated.Aiming at the problem that the accuracy of star centroiding in dynamic environment is affected by star spot degradation,a two-step restoration method is proposed to eliminate star angle blur caused by carrier angular motion and small amplitude random disturbance.Firstly,the fuzzy parameters are detected from the dual Fourier transform amplitude spectrum of the degraded star spot for fast coarse restoration,and then the coarse restoration result is refined according to the image gradient prior of the clear star spot.The simulation results show that the peak signal-to-noise ratio(PSNR)of the restored star spot is greatly improved compared with the traditional restoration method.An system error correction algorithm for star centroiding based on Gradient Boosting Decision Tree(GBDT)model is proposed.The numerical simulation method is used to generate the sample for the training of GBDT model,and the system error of star centroiding is obtained as a function of the filling rate of the detector,the size of the sampling window,the radius of the star spot,and the calculated coordinates of the star point centroid.Based on this functional relationship,the systematic error can be corrected by the calculated value of the coordinates of the star point centroid.A star identification algorithm based on Smith-Waterman local feature matching is proposed.In order to meet the requirements of star identification for the number of stars in the field of FGS,a navigation star library matching the field size and detection ability of FGS is established based on the UCAC4 star library.Considering the observation star magnitude of the FGS and the limited capacity of the navigation star library,a star pattern based on local sequence alignment is proposed for the problem that the "false star" that is not included in the star library causes the recognition rate to decrease.The feature matching algorithm can effectively filter out the real star targets in the star pattern feature sequence containing "false stars".Simulation experiments show that the proposed algorithm is more robust than the grid algorithm and the improved grid algorithm in the presence of multiple false stars.The semi-physical simulation experiment platform of the FGS system was built,and the proposed algorithm was verified by the system error correction experiment and dynamic star centroiding experiment respectively.The accuracy of star centroiding in the static and dynamic environment of the FGS system is tested,and the expected experimental results of the algorithm are obtained.The FGS detection technology is applied to the closed-loop experimental system of precision image stabilization.Under the simulated low-frequency disturbance environment of the spacecraft platform,optical path compensation is based on the LOS deviation calculated by the FGS in real time,After the optical path compensation of LOS deviation,it is detected that the degree of disturbance of the image star point in the x and y directions is reduced by 71.81% and 73.82%,respectively.An exploratory experiment of energy concentration in a long exposure CCD camera with simulated astronomical observations was carried out,the energy concentration of the compensated star spot is increased by 53.3% compared to the diffuse star spot before uncompensated.The experimental results verify the ability of the FGS's detection technology to detect LOS deviation in the precision image stabilization system,Solved the problem of ground experimental verification of space telescope precision image stabilization system.