Diffraction tomography three-dimensional imaging method based on phase shift method and automatic focusing algorithm

Abstract

The invention discloses a diffraction tomography three-dimensional imaging method based on a phase shift method and an automatic focusing algorithm. The method comprises the following steps: reconstructing complex field distribution at a detector by adopting a multi-step phase shift technology; based on an automatic focusing algorithm, calculating acutance values of a phase diagram under different propagation distances, and accurately determining an optimal propagation distance and complex amplitude distribution of an object plane; and high-precision three-dimensional refractive index distribution reconstruction is realized by combining an optical diffraction tomography reconstruction algorithm and a non-negative constraint GS iterative algorithm. The method effectively improves the reconstruction precision of the particle refractive index distribution in optical diffraction tomography, and can be applied to the fields of industrial detection and scientific research such as particle analysis and flow field measurement. The method has the characteristics of high reconstruction efficiency and good precision, and has important application value in the field of micro-nano-scale three-dimensional imaging.

Description

Technical Field

[0001] The present invention relates to the field of particle optical diffraction tomography three-dimensional imaging, and in particular to a diffraction tomography three-dimensional imaging method based on a phase shift method and an automatic focusing algorithm. Background Art

[0002] In microfluidic particle imaging, optical diffraction tomography (ODT) is often used to reconstruct the three-dimensional morphological distribution of particles, thereby obtaining information such as the particle's position, shape, size, and refractive index. Due to the axial thickness of the microfluidic device, the random motion of particles within the device, and inevitable measurement errors in the experiment, the distance from the object plane to the detector plane in ODT is often difficult to determine. If the measured distance cannot be accurately obtained, the object's complex amplitude cannot be reconstructed with high precision, which in turn affects the final quality of the ODT reconstruction. Although many studies have proposed methods to improve the quality of ODT reconstruction, they often focus on the ODT reconstruction process and ignore the optimization of the first step of solving the object's complex field. Low-precision complex field solutions can lead to deviations in the transmission of the original input information of the ODT, thereby reducing the final three-dimensional imaging quality.

[0003] To this end, the present invention proposes a diffraction tomography three-dimensional imaging method based on a phase shift method and an autofocus algorithm, which restores the precise complex field through the phase shift method and the autofocus algorithm, thereby improving the final optical diffraction tomography three-dimensional reconstruction quality. Summary of the Invention

[0004] In order to overcome the shortcomings and deficiencies of the prior art, the present invention provides a diffraction tomography three-dimensional imaging method based on a phase shift method and an automatic focusing algorithm.

[0005] The technical solution adopted by the present invention is a diffraction tomography three-dimensional imaging method based on a phase shift method and an autofocus algorithm, the method comprising: Step S1: using the principle of double-beam interference, changing the reference light angle and recording the intensity graph at different angles; Step S2: using a phase shift method to obtain different interference intensity maps at each reference light angle; Step S3: Using the obtained intensity map, restore the complex light field on the detector plane according to the principle of phase shift method; Step S4: using an autofocus algorithm to determine the precise propagation distance and thereby obtain the object plane complex amplitude; Step S5: reconstruct the three-dimensional refractive index distribution of the object by using the ODT algorithm based on the complex amplitude of the object plane; Step S6: Use the GP iterative algorithm for multiple iterations to ultimately improve the quality of 3D reconstruction.

[0006] Furthermore, the dual-beam interference principle is adopted to change the reference light angle and record the intensity diagram at different angles. The interference pattern is recorded by utilizing the reference light and the object light after reflection to form interference fringes on the detector plane.

[0007] Furthermore, the phase shift method is used to obtain different interference intensity diagrams at each reference light angle by changing the phase of the reference light I_0, and after four phase shifts of 0, π / 2, π, and 3π / 2, it interferes with the object light to obtain different interference intensity diagrams I_n.

[0008] Furthermore, the intensity map obtained is used to restore the complex light field on the detector plane according to the principle of phase shift method. The four-step phase shift method is used to restore the complex light field on the detector plane according to the different interference intensity maps I_n obtained. The complex field u_0 at the detector is obtained, where A 0 is the reference light amplitude, I 0, I 1, I 2, I 3 are the interference intensity diagrams obtained by four reference light phase shifts, j Is an imaginary unit.

[0009] Furthermore, the method of using the autofocus algorithm to determine the precise propagation distance and then obtain the object plane complex amplitude is to propagate the complex field at the detector back to the set distance range using angular spectrum diffraction to obtain a series of complex fields, and then the phase map of each complex field and the Laplace operator convolution kernel are used. Convolution is performed to evaluate the sharpness of the phase image. The distance corresponding to the phase image with the smallest sharpness value is selected as the optimal propagation distance obtained by autofocusing, and the corresponding object plane complex amplitude u(r) is determined.

[0010] Furthermore, the object plane complex amplitude is used to reconstruct the three-dimensional refractive index distribution of the object using the ODT algorithm, which is to reconstruct the multi-angle complex amplitude Make a Rytov approximation: ,in, is the multi-angle complex amplitude, is the incident plane wave, is the complex phase function, is the spatial distribution, and the complex phase function of the scattered field is obtained After that, do a two-dimensional Fourier transform to get , then according to Fourier's diffraction theorem ,in, is the scattering potential The three-dimensional Fourier transform of are the wave vector x, y, z components of the scattering potential, are the x, y, and z components of the incident light wave vector, is the lighting angle The Fourier transform of the scattered field under It is an imaginary unit, which maps the data of each angle to the frequency domain K space. Then perform a three-dimensional inverse Fourier transform to obtain the spatial distribution of the scattering potential. , and then defined according to the scattering potential ,in is the spatial distribution of the scattering potential, is the free-space wave number, is the known background refractive index value, is the spatial distribution of the refractive index, It is the spatial distribution, and the three-dimensional distribution of the refractive index is solved .

[0011] Furthermore, the use of the GP iterative algorithm and multiple iterations to ultimately improve the three-dimensional reconstruction quality is based on the prior condition that the refractive index of the sample will not be less than the known background refractive index. The spectrum is iteratively filled. In each iteration, the spectrum is first subjected to a three-dimensional inverse Fourier transform to obtain a refractive index distribution. Then, the part below the known background refractive index is replaced with the background refractive index value. Then, a three-dimensional Fourier transform is performed to obtain a new spectrum. The part of the original K space that is not filled with data is filled with the new spectrum data. Through multiple iterations, the spectrum space is completely filled to improve the quality of reconstruction.

[0012] Beneficial effects: The present invention proposes a diffraction tomography three-dimensional imaging method based on phase-shifting interference and automatic focusing. In order to solve the problem of inaccurate complex amplitude reconstruction caused by propagation distance error in ODT, which in turn leads to a decrease in imaging quality, an integrated high-precision complex amplitude recovery and imaging solution is proposed. By introducing a four-step phase-shifting interference method in the imaging system, the complex light field information can be stably and accurately extracted from the four phase-shifted holograms without relying on prior information, significantly improving the accuracy of amplitude and phase recovery; furthermore, combined with the automatic focusing algorithm based on the spatial sharpness evaluation of the phase image, the propagation distance at different angles is optimized one by one to ensure that each complex amplitude image is accurately focused on the object plane, effectively overcoming the influence of random fluctuations in the axial position of particles in the three-dimensional imaging scene. The present invention is superior to traditional ODT methods based on single interference patterns or fixed propagation distances in both complex amplitude recovery accuracy and propagation distance estimation accuracy, and can significantly improve the spatial resolution, boundary clarity and numerical stability of subsequent three-dimensional refractive index reconstruction results. Especially under complex conditions such as dense particle distribution and large position deviation, it can still maintain high imaging quality and has good robustness and adaptability. This technology can be widely used in fields such as microfluidic chips, biomedical testing, and environmental particle analysis that require high three-dimensional imaging accuracy, and has important engineering value and industrialization prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1The figure is a flow chart of the method steps of the present invention. DETAILED DESCRIPTION

[0014] It should be noted that, unless there is a conflict, the embodiments in this application and the features described in the embodiments can be combined with each other. The application is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0015] like Figure 1 As shown, a diffraction tomography three-dimensional imaging method based on a phase shift method and an automatic focusing algorithm comprises: Step S1: using the principle of double-beam interference, changing the reference light angle and recording the intensity graph at different angles; Specifically, the core of step S1 is to utilize the principle of dual-beam interference. By varying the reference light angle, the intensity maps at different angles are systematically recorded, forming a data foundation for subsequent imaging. In implementation, this step uses an optical device to cause the reference light and the object light, after being reflected or transmitted by the object, to meet and interfere at the detector plane. The reference light and object light differ in frequency, phase, and other characteristics. When they overlap at the detector plane, they form alternating light and dark interference fringes. These fringes are not simply patterns but contain important information about the object's surface topography and optical properties. By precisely controlling the reference light angle, the detector captures a new interference intensity map with each angle change. The accuracy and range of angle variation are important technical parameters. The higher the accuracy and the wider the range, the more comprehensive the object information captured. Generally, the angle variation range is determined by the object size and imaging requirements. A common range is between 0° and 360°, with an accuracy of 0.1° or higher. Continuously recording intensity maps at different angles is like scanning the object from multiple directions, capturing optical information from various perspectives. This information serves as the foundation for subsequent analysis and imaging.

[0016] Changes in the reference light angle directly affect the distribution and characteristics of the interference fringes. Different angle settings change the interference conditions between the object light and the reference light, which in turn leads to differences in the spacing, shape, and intensity of the interference fringes. These differences reflect the changes in the optical properties of the object in different directions. By recording intensity maps at multiple angles, the loss of information caused by single-angle observation can be avoided. Multi-angle information acquisition can not only obtain details on the surface of the object, but also reflect the spatial structure information of the object to a certain extent, providing a strong guarantee for the subsequent accurate restoration of the three-dimensional information of the object. This step occupies a fundamental position in the entire imaging method. The quality and richness of the data collected directly determine the accuracy and completeness of the subsequent imaging, and is a prerequisite for the smooth progress of the entire imaging process. If the data acquisition in this step is insufficient or there are errors, it will be difficult to obtain accurate results in subsequent imaging operations, so its importance is self-evident.

[0017] Step S2: using a phase shift method to obtain different interference intensity maps at each reference light angle; Specifically, step S2 utilizes a phase shifting method to obtain different interference intensity maps at each reference light angle, providing sufficient data support for recovering the complex light field on the detector plane. In practice, this step focuses on changing the phase of the reference light I_0. Four phase shifts are performed sequentially according to a preset phase change pattern, with the phase shift values set to 0, π / 2, π, and 3π / 2, respectively. After each phase shift, the reference light interferes with the object light, forming a different interference intensity map I_n on the detector. This process requires a precise optical control device to ensure that the reference light phase accurately changes according to the set value. The accuracy of the phase change has a direct impact on the quality of the interference intensity map, and the phase change accuracy is generally required to achieve an extremely small error range to ensure the accuracy and reliability of the interference intensity map at different phases. Each phase change changes the interference conditions between the reference light and the object light, resulting in a change in the intensity distribution of the interference fringes. By recording these four interference intensity maps at different phases, more detailed information about the optical properties of the object can be obtained, which is important data for subsequent complex light field recovery.

[0018] The four preset phase shift values are not arbitrarily set, but rather derived based on the rigorous principles of the phase-shift method. The theoretical basis of the phase-shift method dictates that these four preset phase shifts meet the data requirements for subsequent complex light field recovery. Based on mathematical and optical principles, the interference intensity patterns generated by these four phase shifts exhibit a predetermined relationship. By analyzing and processing these relationships, the amplitude and phase information of the light can be extracted. Obtaining a distinct interference intensity pattern at each reference light angle provides sufficient data support for the subsequent phase-shift method-based recovery of the complex light field at the detector plane. These intensity patterns complement and validate each other, and their comprehensive analysis and calculation enable a more accurate restoration of the true state of the light field, an essential step in achieving precise imaging. If this step fails to obtain sufficient intensity patterns according to the preset phase shifts, the subsequent complex light field recovery will be inaccurate, compromising the accuracy and quality of the overall imaging.

[0019] Step S3: Using the obtained intensity map, restore the complex light field on the detector plane according to the principle of phase shift method; Specifically, the main task of step S3 is to accurately restore the complex light field on the detector plane using the intensity map obtained in the previous step based on the principle of the phase shift method. This step plays a role in data conversion and an important transition in the entire imaging process. During the specific implementation process, the different interference intensity maps I_n obtained in step S2 are processed strictly according to the calculation rules of the four-step phase shift method. The four-step phase shift method is a precise calculation method established based on the phase shift theory. It converts the light information contained in the interference intensity map into a complex light field representation by performing preset mathematical operations on the intensity maps at four different phases. This conversion process requires the application of complex mathematical calculations and optical principles to carefully analyze and process the data in each intensity map. The mathematical operations involve addition, subtraction, multiplication, division, and trigonometric functions of the intensity values. Through these operations, the amplitude and phase information of the light are gradually extracted, and the complex field u_0 at the detector is ultimately obtained. Compared with simple intensity maps, complex light fields can describe the characteristics of light more comprehensively and accurately. They not only contain light intensity information, but also accurately record light phase information, which plays a vital role in accurately presenting the optical properties of objects.

[0020] The computational relationships in the four-step phase shift method are derived through rigorous theoretical derivation and extensive experimental validation. They ensure accurate extraction of complex light field information from intensity maps. Each step in the computational process adheres to strict parameter settings and operational rules; deviations in any step can lead to inaccurate results. Recovering the complex light field at the detector plane is a crucial step in the imaging process, transforming the intensity data collected earlier through interferometry and phase shifting into more valuable and meaningful light field information. This light field information lays a solid foundation for the subsequent determination of the complex amplitude in the object plane and reconstruction of the object's three-dimensional information, directly impacting the accuracy and reliability of the final image. Errors in the computational process or improper data processing during this step will prevent the accurate complex light field at the detector plane from being obtained. Subsequent imaging operations will lose accurate data, leading to deviations or even errors in the overall imaging result. Therefore, the accuracy and reliability of this step crucially impacts image quality.

[0021] Step S4: using an autofocus algorithm to determine the precise propagation distance and thereby obtain the object plane complex amplitude; Specifically, step S4 aims to use the autofocus algorithm to accurately determine the precise propagation distance and then obtain the complex amplitude of the object plane. This step is of great significance for accurately restoring the optical properties of the object. During the implementation process, the complex field at the detector obtained in step S3 is first back-propagated to a pre-set distance range using the angular spectrum diffraction method. The set distance range is determined based on factors such as the approximate distance between the object and the detector and the imaging accuracy requirements. It generally covers the possible object position range to ensure that the accurate propagation distance can be found. Through the back propagation of angular spectrum diffraction, a series of complex fields are obtained at different distances. Then, for the phase map of each complex field, a convolution operation is performed using a preset Laplace operator convolution kernel. The parameters of the Laplace operator convolution kernel are carefully designed and optimized, and the selection of parameters such as its size and coefficient will affect the accuracy of the sharpness evaluation of the phase map. By performing a convolution operation on the phase map, the edge and detail information in the phase map can be enhanced, thereby facilitating the evaluation of the sharpness of the phase map. When evaluating phase image sharpness, a preset calculation method and standard are used to compare the sharpness values of the phase images at different distances. The distance corresponding to the phase image with the minimum sharpness value is selected as the optimal propagation distance for autofocus, and the corresponding object plane complex amplitude u(r) is then determined. This series of operations, through in-depth analysis and processing of the complex field at different distances, gradually selects the propagation distance and complex amplitude that most accurately reflect the actual object.

[0022] Technical parameters such as the set distance range and the parameters of the Laplacian convolution kernel play an essential role in the autofocus process. The appropriateness of the distance range directly impacts the ability to accurately determine the propagation distance. If the range is too small, the true object position may be missed; if the range is too large, the computational effort and processing time will increase. The parameters of the Laplacian convolution kernel determine the accuracy of the phase image sharpness assessment. Inappropriate parameters can lead to incorrect sharpness judgments, thus affecting the determination of the optimal propagation distance. Accurately determining the propagation distance and object plane complex amplitude accurately restores the light field information at the detector to the object plane, effectively avoiding imaging errors caused by inaccurate propagation distance. An accurate object plane complex amplitude more closely resembles the object's true optical properties, providing accurate and reliable data for the subsequent reconstruction of the object's 3D refractive index distribution and crucial for achieving high-precision 3D imaging. Problems in the parameter setting or calculation process during this step will prevent the accurate determination of the propagation distance and object plane complex amplitude, severely impacting the subsequent 3D refractive index distribution reconstruction and causing the imaging results to fail to truly reflect the object's actual condition.

[0023] Step S5: reconstruct the three-dimensional refractive index distribution of the object by using the ODT algorithm based on the complex amplitude of the object plane; Specifically, the core of step S5 is to reconstruct the three-dimensional refractive index distribution of the object using the ODT algorithm from the complex amplitude of the object plane. This is an important step in achieving three-dimensional imaging of the object. In the specific implementation, the reconstructed multi-angle complex amplitude is first subjected to Rytov approximation. Rytov approximation is an approximation method based on preset optical conditions and assumptions. It can convert the complex amplitude into a form that is easier to handle under certain conditions, thereby obtaining the complex phase function of the scattered field. When performing Rytov approximation, certain conditions need to be met, such as the optical properties of the object, the propagation conditions of light, etc. Only when these conditions are met can the approximation result have a high accuracy. After obtaining the complex phase function of the scattered field, a two-dimensional Fourier transform is performed on it. Through this mathematical transformation, the information in the spatial domain is converted to the frequency domain to obtain relevant data. The Fourier transform can decompose complex spatial information into a superposition of different frequency components, which is convenient for subsequent analysis and processing. Then, according to the Fourier diffraction theorem, the data at each angle is mapped to the frequency domain K space. The Fourier diffraction theorem establishes a connection between the spatial and frequency domains, providing a theoretical basis for data conversion and processing. In K-space, a three-dimensional inverse Fourier transform is performed on the data to obtain the spatial distribution of the scattering potential. Finally, based on the definition of the scattering potential, the three-dimensional distribution of the refractive index, n(r), is calculated through inverse solution. This series of operations is closely linked, and through the combined application of mathematical transformations and physical principles, the information in the complex amplitude of the object plane is gradually converted into the three-dimensional refractive index distribution of the object.

[0024] The related theories and transformation methods, such as the Rytov approximation, Fourier transform, and Fourier diffraction theorem, as well as the various parameters involved, are all based on profound optical and mathematical principles. The applicable conditions for the Rytov approximation, the parameter settings for the Fourier transform, and the scope of application of the Fourier diffraction theorem must all be determined in strict accordance with relevant theories and practical situations. Reconstructing the object's three-dimensional refractive index distribution is one of the core goals of this imaging method. This step enables in-depth exploration of the object's internal optical properties from light field information. The object's refractive index distribution reflects important information such as its internal material composition and structure, and is crucial for analyzing its physical and chemical properties. An accurate three-dimensional refractive index distribution provides crucial physical parameters for 3D imaging and is fundamental to achieving high-quality 3D imaging. Improper application of the relevant theories and methods, or inappropriate parameter settings during this step, will result in an inability to accurately reconstruct the object's 3D refractive index distribution. Consequently, the final 3D imaging result will not truly reflect the object's internal structure and optical properties, compromising the accuracy and practicality of the imaging.

[0025] Step S6: Use the GP iterative algorithm for multiple iterations to ultimately improve the quality of 3D reconstruction.

[0026] Specifically, step S6 utilizes the GP iterative algorithm and, through multiple iterations, is dedicated to improving the quality of three-dimensional reconstruction, making the final imaging result closer to the actual condition of the object. In the specific implementation process, this step iteratively fills the spectrum based on the a priori condition that the refractive index of the sample will not be less than the known background refractive index. At the beginning of each iteration, the spectrum is first subjected to a three-dimensional inverse Fourier transform, converting the frequency domain information to the spatial domain to obtain a refractive index distribution. During this conversion process, the calculation rules of the Fourier transform must be accurately implemented to ensure the accuracy of the conversion result. After obtaining the refractive index distribution, it is analyzed and the portion below the known background refractive index is replaced with the background refractive index value. This operation corrects unreasonable data based on the a priori condition to ensure the rationality of the data. Next, the corrected refractive index distribution is subjected to a three-dimensional Fourier transform and converted back to the frequency domain to obtain a new spectrum. Finally, the portion of the original K space that was not filled with data is filled with the new spectrum data. By repeating this series of operations, as the number of iterations increases, the spectrum space is gradually completely filled. The number of iterations is a critical technical parameter that needs to be appropriately set based on the required imaging accuracy and the complexity of the data. Generally speaking, a higher number of iterations results in a more complete spectrum, but this also increases computing time and resource consumption. Through multiple iterations, the spectrum data is continuously optimized, improving reconstruction quality.

[0027] The iterative rules and operation steps based on prior conditions are the core of the GP iterative algorithm to improve the reconstruction quality. The prior conditions provide a basis for the correction and filling of data, ensuring the rationality and reliability of the data. Each operation in the iterative process optimizes the spectral data, and by continuously filling the spectral space, the error caused by missing data is reduced. In the three-dimensional reconstruction process, the integrity of the data is crucial to the resolution and accuracy of imaging. As the spectral space is continuously filled, the details of the reconstructed image are richer and can more accurately reflect the true structure and optical properties of the object. Through multiple iterations, the final three-dimensional reconstruction result is more practical and reliable, and can meet the needs of high-precision imaging of objects. If the iterative rules are set unreasonably in this step, or the number of iterations is insufficient, the spectral space will not be fully filled, resulting in blurring, distortion and other problems in the reconstructed image, and the ideal imaging effect cannot be achieved, affecting the accurate analysis and judgment of the object.

[0028] Preferably, the dual-beam interference principle is adopted to change the reference light angle and record the intensity diagram at different angles by utilizing the reference light and the object light after reflection to form interference fringes on the detector plane and record the interference diagram.

[0029] Specifically, by precisely varying the phase of the reference light I_0, setting it to four preset phase values—0, π / 2, π, and 3π / 2—the reference light interferes with the object light, forming distinct interference intensity patterns I_n on the detector. During implementation, precise optical control is required to ensure high precision in phase variation, as this precision directly impacts the accuracy and reliability of the interference intensity pattern. These four preset phase values are derived from the theoretical basis of the phase shift method. By acquiring intensity patterns at different phases, they capture more detailed information about the object's optical properties, providing sufficient data support for subsequent reconstruction of the complex light field on the detector plane based on the phase shift principle. This is a crucial data acquisition step for achieving precise imaging, and the quality of the acquired data plays a crucial role in determining subsequent imaging accuracy.

[0030] Preferably, the phase shift method is used to obtain different interference intensity diagrams at each reference light angle by changing the phase of the reference light I_0, and interfering with the object light after four phase shifts of 0, π / 2, π, and 3π / 2 to obtain different interference intensity diagrams I_n.

[0031] Specifically, the different interference intensity maps I_n are processed according to the calculation rules of the four-step phase shift method. This precise calculation method, based on the phase shift theory, utilizes complex mathematical calculations and optical principles to meticulously analyze the data in each intensity map. By performing preset operations on the intensity values, the light amplitude and phase information are gradually extracted, ultimately obtaining the complex field u_0 at the detector. The calculation process requires strict parameter settings and operation rules for each step. The calculation relationships are rigorously derived theoretically and verified through extensive experimental validation. This ensures accurate extraction of complex light field information from the intensity map, transforming the previously collected intensity data into more valuable light field information. This lays the foundation for the subsequent determination of the complex amplitude in the object plane and reconstruction of the object's three-dimensional information. Its accuracy and reliability directly impact imaging quality.

[0032] Preferably, the intensity map obtained is used to restore the complex light field on the detector plane according to the principle of phase shift method, based on the different interference intensity maps I_n obtained, according to the four-step phase shift method, The complex field u_0 at the detector is obtained, where A 0 is the reference light amplitude, I 0, I 1, I 2, I 3 are the interference intensity diagrams obtained by four reference light phase shifts, j Is an imaginary unit.

[0033] Specifically, the complex field at the detector is first back-propagated to a pre-set distance range using angular spectrum diffraction. This distance range is determined based on the approximate distance between the object and the detector and the imaging accuracy requirements to ensure coverage of possible object positions. After back-propagation, a series of complex fields at different distances are obtained. For the phase map of each complex field, a convolution operation is performed using a preset Laplace operator convolution kernel. The sharpness is evaluated by enhancing the edges and detail information of the phase map. The distance corresponding to the phase map with the smallest sharpness value is selected as the optimal propagation distance, and the complex amplitude u(r) in the object plane is then determined. Technical parameters such as the set distance range and Laplace operator convolution kernel parameters are crucial for autofocus. Accurately determining the propagation distance and object plane complex amplitude can accurately restore the detector light field information to the object plane, avoid imaging deviation, and provide reliable data for the subsequent reconstruction of the three-dimensional refractive index distribution of the object. This is an important guarantee for achieving high-precision three-dimensional imaging.

[0034] Preferably, the method of using the autofocus algorithm to determine the precise propagation distance and then obtain the object plane complex amplitude is to propagate the complex field at the detector back to the set distance range using angular spectrum diffraction to obtain a series of complex fields, and then the phase map of each complex field and the Laplace operator convolution kernel are used. Convolution is performed to evaluate the sharpness of the phase image. The distance corresponding to the phase image with the smallest sharpness value is selected as the optimal propagation distance obtained by autofocusing, and the corresponding object plane complex amplitude u(r) is determined.

[0035] Specifically, the multi-angle complex amplitudes are first subjected to a Rytov approximation. This approximation must satisfy pre-determined conditions such as the object's optical properties and light propagation conditions to obtain the complex phase function of the scattered field. A two-dimensional Fourier transform is then performed on the complex phase function to convert the spatial domain information into the frequency domain. Based on the Fourier diffraction theorem, the data at each angle is mapped into the frequency domain K space. A three-dimensional inverse Fourier transform is then performed in K space to obtain the spatial distribution of the scattering potential. Finally, the three-dimensional distribution of the refractive index n(r) is inversely solved based on the definition of the scattering potential. The Rytov approximation, Fourier transform, Fourier diffraction theorem, and other related theories and their parameters are determined based on optical and mathematical principles. This step, through a series of mathematical transformations and the application of physical principles, extracts the distribution of the internal optical properties of the object from the light field information, providing important physical parameters for 3D imaging of the object. This is a core step in achieving high-quality 3D imaging. The application of the theory and methods, as well as the accuracy of the parameter settings, directly impacts the reconstruction results.

[0036] Preferably, the object plane complex amplitude is used to reconstruct the three-dimensional refractive index distribution of the object using the ODT algorithm, which is to reconstruct the multi-angle complex amplitude Make a Rytov approximation: ,in, is the multi-angle complex amplitude, is the incident plane wave, is the complex phase function, is the spatial distribution, and the complex phase function of the scattered field is obtained After that, do a two-dimensional Fourier transform to get , then according to Fourier's diffraction theorem ,in, is the scattering potential The three-dimensional Fourier transform of are the wave vector x, y, z components of the scattering potential, are the x, y, and z components of the incident light wave vector, is the lighting angle The Fourier transform of the scattered field under It is an imaginary unit, which maps the data of each angle to the frequency domain K space. Then perform a three-dimensional inverse Fourier transform to obtain the spatial distribution of the scattering potential. , and then defined according to the scattering potential ,in is the spatial distribution of the scattering potential, is the free-space wave number, is the known background refractive index value, is the spatial distribution of the refractive index, is the spatial coordinate distribution, and the three-dimensional distribution of the refractive index is solved .

[0037] Specifically, based on the prior condition that the sample refractive index is not less than the known background refractive index, the spectrum is iteratively filled. In each iteration, the spectrum is first three-dimensionally inverse Fourier transformed to obtain the refractive index distribution, and the part below the background refractive index is replaced with the background refractive index value. Then, a three-dimensional Fourier transform is performed to obtain a new spectrum, and the new spectrum is used to fill the unfilled part of the K space. The number of iterations is an important technical parameter and needs to be reasonably set according to the imaging accuracy and data complexity. Multiple iterations optimize the spectrum data, reduce data missing errors, fill the spectrum space to make the reconstructed image richer in details, accurately reflect the real structure and optical properties of the object, and improve the quality of three-dimensional reconstruction. If the iteration rules or number of times are unreasonable, it will affect the reconstructed image effect and accurate analysis of the object.

[0038] Preferably, the use of the GP iterative algorithm and multiple iterations to ultimately improve the three-dimensional reconstruction quality is based on the prior condition that the refractive index of the sample will not be less than the known background refractive index. The spectrum is iteratively filled. In each iteration, the spectrum is first subjected to a three-dimensional inverse Fourier transform to obtain a refractive index distribution, and then the part below the known background refractive index is replaced with the background refractive index value. Then, a three-dimensional Fourier transform is performed to obtain a new spectrum, and the part of the original K space that is not filled with data is filled with the new spectrum data. Through multiple iterations, the spectrum space is completely filled to improve the quality of reconstruction.

[0039] Specifically, the iterative process is started. The first step is to perform a three-dimensional inverse Fourier transform on the spectrum. This operation aims to convert from the spectral domain to the spatial domain to obtain the corresponding refractive index distribution, which carries the spatial information of the refractive index inside the sample; then, based on the prior conditions, the part of the refractive index distribution that is lower than the background refractive index is identified and replaced with the background refractive index value. This step is to use the prior knowledge to constrain the refractive index distribution in the spatial domain to ensure that it conforms to physical reality; then, a three-dimensional Fourier transform is performed on the corrected refractive index distribution, and the new spectrum is generated back to the spectral domain. The new spectrum is more in line with the real sample characteristics due to the previous correction; then, the new spectrum data is used to fill the original The unfilled data area in this K-space (spectral space) and the degree of K-space filling are related to the integrity of subsequent reconstruction. Each iteration supplements the K-space information. By repeatedly repeating the iterative cycle of "inverse Fourier transform-refractive index correction-Fourier transform-K-space filling", the spectral space content is continuously refined, allowing the spectral information to continuously approach the true spectrum of the sample, and ultimately achieving full filling of the spectral space, providing a more complete and accurate spectral basis for 3D reconstruction, thereby improving the quality of 3D reconstruction. The entire process relies on algorithm iteration and prior condition constraints to gradually optimize the spectral data required for reconstruction, promoting the improvement of 3D imaging quality from the perspective of spectral processing.

[0040] While embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A diffraction tomography three-dimensional imaging method based on phase shift method and autofocus algorithm, characterized in that: The method includes: Step S1: using the principle of double-beam interference, changing the reference light angle and recording the intensity graph at different angles; Step S2: using a phase shift method to obtain different interference intensity maps at each reference light angle; Step S3: Using the obtained intensity map, restore the complex light field on the detector plane according to the principle of phase shift method; Step S4: using an autofocus algorithm to determine the precise propagation distance and thereby obtain the object plane complex amplitude; Step S5: reconstruct the three-dimensional refractive index distribution of the object by using the ODT algorithm based on the complex amplitude of the object plane; Step S6: Use the GP iterative algorithm for multiple iterations to ultimately improve the quality of 3D reconstruction.

2. The diffraction tomography three-dimensional imaging method based on phase shift method and autofocus algorithm according to claim 1, characterized in that: The dual-beam interference principle is adopted to change the reference light angle and record the intensity diagram at different angles. The interference pattern is recorded by utilizing the reference light and the object light reflected from the object to form interference fringes on the detector plane.

3. The diffraction tomography three-dimensional imaging method based on phase shift method and autofocus algorithm according to claim 1, characterized in that: The phase shift method is used to obtain different interference intensity maps at each reference light angle by changing the phase of the reference light I_0, and after four phase shifts of 0, π / 2, π, and 3π / 2, it interferes with the object light to obtain different interference intensity maps I_n.

4. The diffraction tomography three-dimensional imaging method based on phase shift method and autofocus algorithm according to claim 1, characterized in that: The intensity map obtained is used to restore the complex light field on the detector plane according to the principle of phase shift method. The complex light field on the detector plane is restored according to the obtained different interference intensity maps I_n according to the four-step phase shift method. The complex field u_0 at the detector is obtained, where A 0 is the reference light amplitude, I 0, I 1, I 2, I 3 are the interference intensity diagrams obtained by four reference light phase shifts, j Is an imaginary unit.

5. The diffraction tomography three-dimensional imaging method based on phase shift method and autofocus algorithm according to claim 1, characterized in that: The method of using the autofocus algorithm to determine the precise propagation distance and then obtain the object plane complex amplitude is to propagate the complex field at the detector back to the set distance range using angular spectrum diffraction to obtain a series of complex fields. Then, the phase map of each complex field and the Laplace operator convolution kernel are used. Convolution is performed to evaluate the sharpness of the phase image. The distance corresponding to the phase image with the smallest sharpness value is selected as the optimal propagation distance obtained by autofocusing, and the corresponding object plane complex amplitude u(r) is determined.

6. The diffraction tomography three-dimensional imaging method based on phase shift method and autofocus algorithm according to claim 1, characterized in that: The object plane complex amplitude is used to reconstruct the three-dimensional refractive index distribution of the object using the ODT algorithm. The reconstructed multi-angle complex amplitude Make a Rytov approximation: ,in, is the multi-angle complex amplitude, is the incident plane wave, is the complex phase function, is the spatial distribution, and the complex phase function of the scattered field is obtained After that, do a two-dimensional Fourier transform to get , then according to Fourier's diffraction theorem ,in, is the scattering potential The three-dimensional Fourier transform of are the wave vector x, y, z components of the scattering potential, are the x, y, and z components of the incident light wave vector, is the lighting angle The Fourier transform of the scattered field under It is an imaginary unit, mapping the data at each angle to the frequency domain K space; then performing a three-dimensional inverse Fourier transform to obtain the spatial distribution of the scattering potential , and then defined according to the scattering potential ,in is the spatial distribution of the scattering potential, is the free-space wave number, is the known background refractive index value, is the spatial distribution of the refractive index, is the spatial coordinate distribution, and the three-dimensional distribution of the refractive index is solved .

7. The diffraction tomography three-dimensional imaging method based on phase shift method and autofocus algorithm according to claim 1, characterized in that: The use of the GP iterative algorithm and multiple iterations to ultimately improve the three-dimensional reconstruction quality is based on the prior condition that the refractive index of the sample will not be less than the known background refractive index. The spectrum is iteratively filled. In each iteration, the spectrum is first subjected to a three-dimensional inverse Fourier transform to obtain a refractive index distribution. Then, the part below the known background refractive index is replaced with the background refractive index value. Then, a three-dimensional Fourier transform is performed to obtain a new spectrum. The part of the original K space that is not filled with data is filled with the new spectrum data. Through multiple iterations, the spectrum space is completely filled.

Images