The Study Of Fourier Ptychographic Microscopy Based On Vision Assisted Positioning

With the ongoing microscopic world exploration,the microscopic imaging technology hasbeen developed rapidly.Fourier ptychographic microscopy(FPM)which developed in recent years is a novel approach for improving the imaging resolution property.By illuminating sample from multiple angles,the large field of view(FOV)and high-resolution imaging can be achieved through synthetic aperture and phase retrieval algorithms.It can break the contradictory between FOV and resolution in traditional imaging methods,and the experimental equipment with a set of LED array is easy to implement relatively.There are great practical potentials in the fields of cell imaging,fluorescence microscopy,phase microscopy,digital holography,etc.This dissertation focuses on the improvement of FPM and the main works includes:Firstly,we made a detailed analysis about basic principles and imaging performance of FPM from several aspects including iterative sequence,overlapping areas of adjacent subapertures,and the change of light intensity.Then the optimizing parameters can be estimated to build the FPM system with high resolution imaging in large FOV.Secondly,considering the existing position error of LED array,the improved FPM method was proposed with a vision system to assist in achieving accurate LED positions.In this method,the three-dimensional locations of LED array can be achieved by using the extra vision assisted system.However,there are two actual problems: the FOV mismatch between the microscopic camera and vision assisted camera,the LED coordinates in camera coordinate system can not be directly converted to the object plane coordinate system.To target these problems,a standard resolution board with large circular mark points attached additionally was used to establish the relationship between LED array and object plane coordinate system,where the locations of center circular points and the original scale mark on resolution board were calibrated previously.Therefore,the accurate position and orientation parameters from each LED to the object plane can be determined in FPM,and there is no need for the tedious manual alignment or aligning error compensation algorithm.The experimental results of resolution plate and human blood cell demonstrated the validity and reliability for imaging quality improvement in FPM.