The Research Of Phase Retrieval Technique Based On Complex Algorithm

Phase retrieval technique reconstructs phase information by acquiring the intensity information of defocus plane,and it is widely used in optical fields such as optical element detection,adaptive optics,and active optics.In recent years,Phase retrieval technology continues to develop along multiple dimensions such as changing iterative update conditions,adding modulation,axial multi-plane measurement and radial splicing.The traditional single-image iterative phase retrieval algorithm optimizes the iterative update conditions,although accuracy is improved,but the robustness to noise is poor,and the convergence speed needs to be improved.Based on the above problems,this paper first combines phase retrieval algorithms optimized in different dimensions,and designs a single-image iterative hybrid algorithm and two multi-image iterative hybrid algorithms.The single-image iterative hybrid algorithm mixes the over-sampling smoothness algorithm with the error reduction algorithm,optimizes the filter parameters of the algorithm,and further improves the accuracy of the algorithm.The multi-image iterative hybrid algorithm mixes axial multi-image algorithm and the update conditions of single-image algorithm to improve the accuracy of the algorithm while improving the convergence speed and noise resistance of the algorithm.At the same time,the article analyzes the possible axial distance error,radial alignment error,random disturbance error,and detector quantization error that may occur in experimental measurement based on the hybrid phase retrieval algorithm,and gives correction methods for axial distance error,radial alignment error and random disturbance error.Finally,a phase retrieval detection platform is built,a convenient and editable phase recovery algorithm integration software is designed,and experiments are used to verify the effectiveness of the hybrid algorithm and error correction method.