Optimization Based On Aperture Modulation And Aberration Correction In Microscopy

To realize the imaging of biological tissue,fluorescence microscopy uses fluorescence probes to mark the imaging targets,then collects and detects the excited fluorescence.Due to the advantages of high signal-to-noise ratio,highly specific,slight sample damage and the ability of multi-color staining,fluorescence microscopy technology has become one of the important tools in the field of biological imaging.However,due to the limitations of the conventional fluorescence microscopy system,it needs to be optimized when imaging into deep tissue.In confocal microscopy,there is a pinhole to block the defocused light,which is set to be conjugate with the point source and scanning point.Confocal microscopy is widely used because of its high resolution and optical sectioning capability.However,the ability of the pinhole to suppress background noise is always restricted,and then the imaging depth is limited.So here,based on the divided aperture modulation technique,the divided cosine-shaped apertures are proposed to be combined with the confocal microscopy.By adjusting the distribution frequency of cosine-shaped apertures,better transverse resolution,axial resolution and signal-to-background ratio can be acquired.In order to observe the biological tissue clearly,it is necessary to enhance the spatial resolution and break the optical diffraction limit.Thus,optical super-resolution imaging systems have gotten the rapid growth,which can be divided into point-scanning microscopy and wide-field microscopy.However,when light passes through thick scattering medium,the inhomogeneous refractive index and optical aberrations will affect the phase and polarization of the optical field,thus deteriorating the performance of the system.Currently,adaptive optics has been a valuable tool to compensate the optical aberrations,taking advantages of an active phase element and the relative algorithms.The application of adaptive optics in super-resolution fluorescence microscopy can overcome the aberrations and scatterings caused by the deep tissue.For the vortex beam in stimulated emission depletion microscopy based on point scanning,the iterative optimization algorithm based on correlation coefficient and the parallel optimization algorithm based on coherent optical adaptive technique are proposed.Then through the scattering medium,the doughnut shape focus of vortex beam can be recovered well,realizing high-efficiency focusing.Compared with the pupil adaptive optics system,the conjugate adaptive optics system can keep a high-quality focal spot among broader areas to achieve high-speed wavefront correction.Experimental systems are also built to further investigate the aberration correction process.What's more,in conjugate adaptive optics system,automatic optimal multiple guide stars selection algorithm based on point scanning is investigated to obtain a larger corrected field of view and correct the aberrations effectively.In wide-field structured illumination microscopy,deep learning is used to achieve the fast correction of optical aberrations.Optical aberrations can be decomposed into the combination of Zemike polynomials with finite terms.Then the mapping between the coefficients of Zernike modes and their corresponding distorted structured patterns is established to train the convolutional neural network.With the precise prediction of optical aberrations,distortion compensation for structured illumination microscopy can be achieved through the image reconstruction.In summary,this thesis studies the optimization of fluorescence microscopy imaging technique.For confocal microscopy,the divided cosine-shaped apertures are used to modulate and optimize the system.For stimulated emission depletion microscopy based on point scanning and structured illumination microscopy based on wide-field imaging,adaptive optics is taken for aberration correction.The techinuqes proposed here are of great significance for biological tissue imaging with high resolution and fast speed,which may have a broad application prospect in the field of biomedical research.