Research On Fourier Ptychographic Imaging Algorithm With Position Correction Capability

Fourier ptychography reconstruct high-resolution images by stitch a series of low-resolution images,breaking the constraints between high resolution and large field in traditional imaging system.In recent years,the technique of Fourier ptychography has been rapidly developed.However,in the actual imaging system,there are various system errors and noise which effect the recovered image quality.In this paper,a variety of optimization algorithms are used to correct the systematical errors in Fourier ptychography imaging as well as improving the quality of image reconstruction.The main research contents are as follows:Firstly,using the sparsity of the sparse image itself,this paper uses the₁l-norm to measure the prior knowledge and as a regular term,which is combined with the fidelity term.In addition,the natural image has gradient sparsity.Combined with the total variation model,a Fourier ptychography algorithm based on total variation is proposed,which is solved by Split Bregman algorithm.It is verified by experiments that the two algorithms are better than those without multiple prior existence.Secondly,the momentum factor increases the convergence rate of the stochastic gradient descent method by adding the cumulative velocity step in the decreasing direction of the objective function.At the same time,the addition of momentum avoids the generation of local optimal solutions.This paper combines the momentum and gradient descent to solve the Fourier ptychography imaging problem.The results of simulation experiments in Gauss noise environment of this algorithm show that the image quality of the algorithm is improved obviously.In addition,the actual data experiment results show the effectiveness of the algorithm.Thirdly,genetic algorithms achieve the goal of optimization by simulating natural evolutionary processes.It has strong adaptability and effectiveness,and can effectively avoid the production of local optimal value.The genetic algorithm is used to correct the position of multiple LEDs in the Fourier ptychography imaging system to find the actual position of the LED light source,which can effectively improve the reconstructed image quality.The experimental results show that multiple LED light sources in the bright field can be effectively corrected,which significantly improves the quality of the resconstructed image and has a certain noise immunity.