| At present,Adaptive Optics(AO)system has been successfully applied to ground-based astronomical telescope system in China and beyond,which improves the imaging quality of astronomical observation images by real-time correction of dynamic distorted wavefront.As the real-time detector of AO system,Shark-Hartmann Wavefront Sensor(SHWFS)is of great significance,and its detecting precision directly affects the correction performance and closed-loop stability of the whole system.The detecting precision of SHWFS mainly depends on the spot location precision and the wavefront reconstruction precision,both of which are related to the Signal to Noise Ratio(SNR)of the spot array images of SHWFS.The research object of this dissertation is the SHWFS spot array image with low SNR.SHWFS spot array image with low SNR enjoys particularity compared with general images.A frame of image is composed of multiple sub-aperture images,where each sub-aperture completes imaging separately and the spot shape of each sub-aperture is different.The spot and noise are staggered and distributed unevenly,and the spot edge is fuzzy.By contrast,the SHWFS spot array image with low SNR is formed because of strong background noise,limited beacon light brightness,limited observation target brightness,and relatively strong noise of photoelectric detectors.Spot array images with low SNR are divided into two categories,including spot array images with low overall SNR and spot array images with low SNR of partial sub-apertures.Therefore,the features of spot images of all sub-apertures should be fully considered during spot location and wavefront reconstruction analysis,so as to improve the reconstruction precision of wavefront to be detected.To take the various sub-apertures into full consideration,the wavefront reconstruction errors were adopted in this dissertation as the evaluation indicators of the spot location algorithm and wavefront reconstruction algorithm.The main contents of research are described below:1.In this dissertation,an SHWFS digital simulation platform was established according to wavefront simulation,SHWFS working principle,spot location and theoretical analysis of wavefront reconstruction,which consisted of three parts,including distorted wavefront by simulation,spot location and wavefront reconstruction.The functions of this platform are as follows:(1)Various static and dynamic wavefronts with single aberration or composite aberrations can be generated,and the aberration of all wavefronts and the parameters of distorted wavefronts due to atmospheric turbulence are adjustable.(2)Spot array images can be generated,and the structure parameters of the Hartman wavefront detector can be adjusted,including the number of pixels on the target surface of charge coupled device(CCD),the sub-aperture arrangement style,the number of sub-apertures,the number of pixels of sub-apertures,the number of pixels of diffraction spots,etc.The noise type can be selected at random,and the mean value and variance of the background and noise,as well as the light intensity,are adjustable.(3)Spot location and wavefront reconstruction can be carried out.The simulation platform has integrated multiple spot location algorithms and wavefront reconstruction algorithms,where the parameters involved in each algorithm can be adjusted as per actual needs of the system,and the analysis of spot location errors and wavefront reconstruction can be performed.2.Lack of light or weak light may occur in partial sub-apertures of SHWFS due to the interference of strong turbulence,the dimness of the observation targets or the blocking of the light path.Therefore,slope zero-setting reconstruction of sub-apertures and sub-aperture removal and reconstruction methods were applied to solve the problems in this research.The influences of lack of light of single sub-aperture,continuous multiple sub-apertures and non-continuous multiple sub-apertures on wavefront reconstruction errors were analyzed in detail,and the results showed that:(1)When the slope zero-setting method was applied in the case of lack of light of single sub-aperture,the wavefront reconstruction errors increased with the increase in the radius of the sub-aperture position relative to the center of the full aperture.(2)When the sub-aperture removal method was applied in the case of lack of light of single sub-aperture,the wavefront reconstruction errors were uniformly distributed on the full aperture and presented small value.(3)In the case of lack of light of continuous and non-continuous multiple sub-apertures,the removal and reconstruction method for sub-apertures lack of light required a large amount of calculation and occupied large memory spaces,but the wavefront reconstruction errors were smaller than that when sub-aperture slope zero-setting method was applied.Based on the coefficient of error transfer of the reconstruction matrix and its correlation with the approximate Airy disk width of the number of microlens arrays and spots,the influences of spatial sampling rate and spot width on wavefront reconstruction were analyzed.3.By means of numerical simulation analysis,the wavefront reconstruction errors of several spot location algorithms were compared under different SNRs.The priority of spot location algorithms under different SNRs was proposed,and the influences of the optimal parameters,application range and stability of each spot location algorithm,as well as windowed pre-processed images on the algorithms were presented.The parameter setting rules of the two image pre-processing algorithms(thresholding algorithm and windowed algorithm)after the combination with the spot location algorithm were as follows:When images were pre-processed by the thresholding algorithm only,the threshold should be increased appropriately with the decrease in SNR;when images were pre-processed by both the thresholding algorithm and the windowed algorithm,the threshold and window width should be decreased gradually with the decrease in SNR.Compared with commonly used Gaussian function,the weight function sinc~2 in weighted centroid algorithm had better anti-noise performance under low SNR.The selection orders of spot location algorithms under different SNRs were as follows:When the mean of SNR(SNR_m)was lower than 2,the cross-correlation(CORR)algorithm was the best,followed by the thresholding weighted center of gravity(T-WCOG)algorithm;When the SNR_m was between 2 and 10,the T-WCOG algorithm was the best,followed by the thresholding intensity weighted centroiding(T-IWC)algorithm and the Thresholding center of gravity(T-COG)algorithm;when the SNR_m is higher than 10,The T-COG algorithm was simple and effective.4.To validate the parameter setting rules in spot location algorithms and the selection orders of spot location algorithms under different SNRs,an experiment system was built.In the experiment,spot array images under different SNRs were obtained by adjusting the optic power,light path attenuation multiple and the exposure time of the laser.The correctness of the simulated spots and noise in the simulation platform was proved by analyzing the dark field images and spot array images in the experimental data.Moreover,the correctness of the parameter setting rules of spot location algorithms in the simulation and the selection orders of spot location algorithms under different SNRs was validated by analyzing the parameter optimization results of the spot location algorithms and comparative analysis of wavefront reconstruction errors under different SNRs in spot location algorithms. |
PDF Full Text Request