Realization And Processes Optimization In Astigmatism-based 3D Localization Microscopy

Since its invention, super-resolution localization microscopy has been widely used in biology because of its ultra-high spatial resolution(which breaks the diffraction limit) and system simplicity. The demand of three-dimensional spatial information at nanoscale in the study of life science leads to the development of 3D super-resolution localization microscopy. Currently, although there are different kinds of methods to realize 3D super-resolution localization imaging, the most popular one is optical astigmatism-based method due to its simplicity and high performance. However, the current astigmatism-based 3D super-resolution localization imaging method lacks quantitative analysis of the performance, thus fails to provide effective imaging optimization scheme under different biological application scenarios.To overcome the difficulties mentioned above, this thesis presents an optimized 3D super-resolution localization imaging method, basing on a quantitative analysis and optimization of the imaging system parameters and an optimized post-acquisition image processing algorithm. The proposed method can achieve high localization accuracy under a relatively deep image depth, and the optimization scheme in this method could provide useful guidance for 3D super-resolution localization microscopy imaging.The main works of the thesis include:(1) 3D super-resolution localization microscopy system optimization. We discuss the position and the focus length of cylindrical lens on the influence of point spread function in astigmatism-based 3D localization microscopy, and thus obtain an optimized point spread function.(2) Image processing optimization. Firstly, we compare the different raw image pre-processing methods used in astigmatism-based 3D localization microscopy and find a optimized method for raw image pre-processing. Secondly, we compare the localization results from different 3D single molecule localization algorithms, and quantify the best 3D single molecule localization algorithm within the imaging depth. Finally, we design and realize the GPU parallel calculation-based data processing of astigmatism-based 3D localization microscopy, which greatly enhances the data analysis efficiency.(3) We verify the performance of the optimized astigmatism-based 3D localization imaging system by simulation, fluorescent microsphere experiment, as well as biological experiments using cell microtubules and rat brain slices. The results indicate our optimized system is capable of imaging deeper and achieving higher localization precision.