Research On Blind Deconvolution Algorithms For Solar Adaptive Optics Images

Ground-based optical telescopes are important tools for astronomical observations.However,atmospheric turbulence causes wavefront distortions in light waves,resulting in severe degradation of the imaging resolution of optical telescopes.Adaptive optics technology can detect and correct wavefront aberrations in real time,and is widely used in astronomical ground-based optical telescopes to improve the imaging quality of the telescopes.However,due to the limitations of the adaptive optics system's own hardware performance,the correction is partial and insufficient,and the adaptive optics images still have residual aberrations and require post-processing.Living things and the environment on earth depend on the energy provided by the sun,and the use of ground-based optical telescopes to obtain high-resolution solar images is of great significance for the solar observation and research.Aiming at the degraded solar adaptive optics images,this thesis conducts the blind deconvolution algorithms research to realize the reconstruction of solar adaptive optics images.The research work of this thesis includes the following contents:Firstly,atmospheric turbulence,adaptive optics,blind deconvolution,image restoration quality evaluation and other related theories are introduced,which provide a theoretical basis for the subsequent research work in this thesis.Secondly,theoretical analysis and experiments demonstrate that multi-frame blind deconvolution has better robustness than single-frame blind deconvolution.And two regularization-based multi-frame blind deconvolution models are proposed,one based on the L0 gradient prior and one based on the second-order total generalized variation.The models are solved by alternating minimization and half-quadratic splitting methods,and the algorithm processes are given respectively.Thirdly,according to the regularization-based multi-frame blind deconvolution algorithms and the characteristics of the solar adaptive optics images,the specific process and methods for the solar adaptive optics images reconstruction are given.And the reconstruction experiments of the solar images observed by NVST are carried out.Through the comparison of subjective visual effects and objective indexes as well as analysis of the reconstruction results,the effectiveness of the multi-frame blind deconvolution algorithms based on the L0 gradient prior and the second-order total generalized variation proposed in this thesis are demonstrated.Finally,the phase diversity is regarded as a special blind deconvolution under a unified framework,the related theory of the classical phase diversity is introduced,and an improved phase diversity model based on the low rank prior is proposed.Alternating minimization and half-quadratic splitting methods are used to solve the model,and the specific algorithm process for image reconstruction by the phase diversity is given.And reconstruction experiments are carried out on the simulated focused and defocused solar images.Through the comparison of subjective visual effects and objective indexes as well as analysis of the reconstruction results,it is demonstrated that the phase diversity based on low-rank prior proposed in this thesis can improve the stability of image reconstruction.