Wavefront Correction Of Model-based Wavefront Sensorless Adaptive Optics System

In the imaging process of optical system,there are static random aberrations inside and dynamic aberrations caused by atmospheric turbulence outside.The image sensor in the system is mixed with various types of noise.The existence of these aberrations and noise greatly reduces the image quality and makes the image effect blurred and distorted.The conventional adaptive optics(AO)system is an effective mean to correct wavefront aberrations in real time.However,wavefront measurement is very difficult in complex situations,so the correction effect is limited.The wavefront sensorless AO system eliminates the wavefront sensor,and the correction is still effective under adverse conditions.It has the characteristics of simple structure,high reliability and small size.In terms of control algorithms of wavefront sensorless AO system,the control algorithm based on geometrical optics principle needs to define basis functions beforehand,but it does not depend on the specific type of basis functions.It can directly use the measured influence functions as basis functions without removing the system aberrations.It is simple and convenient,and can be applied to large and small aberrations.Firstly,the wavefront sensorless AO system under noisy is established by using a 88-element deformable mirror and charge coupled device(CCD)and with an extended object as correction object.The linear relationship between the masked detector signal(MDS)and the mean square gradient(MSG)of the extended object imaging is verified under noise.The algorithm based on this linear relationship is used as the control algorithm of the wavefront sensorless AO system.Simulation is carried out to check the correction ability of the model wavefront sensorless AO system for extended object imaging under noise.Results showed the correction effect is very close to each other under different noise for the same turbulence condition.According to the order of turbulence conditions from small to large,compared with results of signal-to-noise ratio of 20 dB,averaged root mean square(RMS)relative errors of signal-to-noise ratio of 5dB after correction are 3.71%,3.37% and 2.32% respectively.Above results show that control algorithm based on this linear relationship has the good anti-noise capability.Secondly,the 37-unit deformable mirror optical experiment platform is built by using a piezoelectric deformable mirror,CCD camera,shack hartmann sensor and other components,taking point object as correction object.Control algorithm based on this linear relationship is implemented by Microsoft Visual Studio 2010(VS2010)software in the closed-loop module of MFC.Then,the calibration ability of control algorithm based on this linear relationship for point object imaging is explored by comparing with the experiment of stochastic parallel gradient descent(SPGD)control algorithm.Experiments show that the control algorithm based on this linear relationship has faster correction speed and is not prone to local convergence,compared with the SPGD control algorithm.