Phase Retrieval Method And 3D Morphological Reconstruction Technology Of Blood Cells Based On Non-orthogonal Phase Imaging

Blood cells have multiple functions such as oxygen transfusion,blood coagulation,and immunity,and they take over the most basic work of the body.The study of its number of types,internal structure and functional activities is the only way to master the essence of life and human health.Nowadays,microscopic super-resolution imaging of blood cells has developed to a new level.However,the combination of flow and mirror examination is still used in the clinical application of blood analysis,which cannot meet the development requirement of precision medicine.Our preliminary achievements show that its main reasons are both the method and technology bottlenecks in non-destructive,rapid and accurate imaging of batch of blood cells' substructures.Therefore,the need for cross-scale,cross-structure,multi-level,multi-view,multi-modal of non-destructive,high-efficiency and ultra-precision detection technology for cells is imminent.Quantitative phase imaging,as a non-invasive,no damage,high-efficiency,high-precision bio-optical imaging technology,mainly uses the interaction of light and phase objects such as cells to generate spatial phase delay to achieve clear imaging.It is worth mentioning that since it only needs to obtain the light wave phase shift caused by the different refractive index distributions corresponding to the cell substructure,it is a native imaging technology that does not require special preparation and processing for translucent samples such as biological cells.This feature has far-reaching significance for living cells.In the past 30 years,a large number of researchers at home and abroad have successively proposed many advanced quantitative phase imaging.This technology has achieved outstanding results in the fields of cell imaging,parameter measurement,dynamic behavior analysis,and substructure reconstruction,and has even been applied in clinical medicine,disease diagnosis,and efficacy evaluation.At present,the mainstream quantitative phase imaging includes processes such as information acquisition,phase restoration,and three-dimensional reconstruction.Among them,the phase retrieval method and the 3D reconstruction technology play a decisive role in imaging quality,speed and accuracy,so the research on these two aspects is extremely important.In this academic dissertation,the key theories and related new methods of rapid phase information retrieval and efficient 3D morphology reconstruction based on non-orthogonal with double channel are systematically studied and explored.It aims to solve the key problems of the poor anti-interference ability of the phase retrieval method under phase-shifting interference and the traditional tomographic reconstruction technology relies on large data volume,so as to build a complete detection theory and technology for label-free,non-destructive,accurate morphological and rapid 3D imaging of batch blood cells,provides nano-scale spatial resolution and millisecond-level time domain resolution,and realize a new model of one-stop accurate blood analysis.The main achievements are as follows:(1)The technical characteristics of various current typical phase imaging methods and the characteristics of typical phase retrieval methods under the technology are systematically analyzed.Aiming at the problems of traditional phase retrieval methods such as weak anti-interference ability,insufficient phase shift detection accuracy,and insufficient phase retrieval speed,a three-step blind phase-shifting phase retrieval method based on the Euclidean matrix M-P norm theory is proposed.This method is easy to implement and insensitive to noise.Only three interference patterns under any phase shift can be used to achieve phase retrieval,without any approximate processing of background light intensity,intensity modulation factor and phase.This method exhibits outstanding stability and accuracy in the phase retrieval calculations of various types of waveforms and typical blood cells.(2)Aiming at the retrieval of phase information in single-wavelength phase-shifting interferometry often requires complex unwrapping operations to deal with the problem of discontinuous phase,a dual-wavelength blind phase-shifting phase retrieval method based on iterative algorithm is proposed.This method eliminates the tedious and time-consuming calculation steps of the phase unwrapping algorithm,and has a longer wavelength range,a larger measurement scale and higher imaging quality.In addition,this method does not require other known physical quantities,only requires five-step blind phase shift to achieve phase retrieval,which can effectively avoid the adverse effects of phase shift errors and high-order harmonics,and ensure high-quality phase imaging.Simulations and experiments of a variety of measurement targets show the effectiveness,reliability and universality of the method.(3)The development of quantitative phase imaging in the 3D reconstruction of homogeneous cells(red blood cells)has matured,but there is still no perfect solution for heterogeneous cells(nucleated cells),especially binuclear and even multinucleated cells.Aiming at the problems of uneven refractive index distribution,multiple jumps and unclear edges of multi-medium cells,based on the spatial geometry theory,combined with image edge detection algorithms,a method of reconstructing the 3D spatial morphology of multi-medium cells under incomplete data is proposed.This method does not need to collect multiple images,scan and calculate refractive index or thickness information.It only needs to extract the edges of the two phase images in the orthogonal direction of the measured cell,and then realize the 3D reconstruction by rotating around the axis.Both the simulation and experiment of binuclear cell using this method have obtained good results.(4)Traditional tomographic reconstruction technology requires scanning to collect a large amount of data and computing,and the orthogonal phase imaging cannot be applied to the reconstruction of real cells due to the limitations of fixed angle and slow reconstruction speed.To solve the above problems,a maximum entropy tomographic reconstruction method under non-orthogonal phase imaging is proposed.This method only requires two phases in non-orthogonal directions as the original data,combined with the optimized maximum entropy tomography algorithm,to achieve fast and efficient 3D reconstruction.It is worth mentioning that the non-orthogonal phase imaging system combines flow cytometry and optical tweezers to achieve real-time and accurate acquisition of dual-channel phase information.The simulation and experimental results of a variety of typical blood cells show that this method is of great significance to the upgrading of clinical blood cell detection technology.(5)At present,clinical blood routine examinations mainly obtain the number of different types of blood cells as the basis for diagnosis and treatment,and the secondary microscopy of diseased cells requires the help of invasive fluorescence microscopy.In response,a non-invasive and non-destructive blood cells in batches classification and identification method based on the phase map analysis of the cells under non-orthogonal with double channel is proposed.This method includes five phase characteristic parameters that reflect the external morphology and substructure characteristics of blood cells.In addition,the non-orthogonal phase imaging requires only a single exposure to obtain two phases information for accurate blood cell identification.In the verification experiment,combined with the threshold segmentation method,the five subtypes of white blood cells(monocytes,lymphocytes,neutrophils,eosinophils and basophils)were accurately classified.Finally,based on this method,a blood cell classification and recognition software based on quantitative phase characteristic analysis is developed.It opens up a new path for clinical diagnosis and treatment of blood cell testing,and provides theoretical support.The research work of this dissertation has not only made some progress in theory,but also has been verified experimentally.It not only provides new ideas and methods for blood cell detection,but also provides prototypes for the replacement of sophisticated and high-end medical diagnostic instruments.