Retina hemodynamic analysis and measurement system based on optical fundus image

Abstract

The invention discloses a retina hemodynamic analysis and measurement system based on an optical fundus image. The system comprises an acquisition module; a modeling module and a simulation analysis module. According to the method, bidirectional fluid-solid coupling simulation analysis is carried out on the three-dimensional model of the fundus blood vessel through Fluent and Transient Structural in the finite element software ANSYS, blood flow dynamic data closely related to clinic such as blood vessel section pressure, flow and WSS can be rapidly and accurately obtained, whether parameters are within a normal value range or not is judged, and therefore the fundus blood vessel lesion can be found in time, and the fundus blood vessel lesion can be found in time. The patient is helped to enter deep examination as soon as possible and take a corresponding subsequent treatment scheme.

Description

Technical Field

[0001] The invention belongs to the field of ophthalmic medicine, and in particular relates to a retinal hemodynamics analysis and measurement system based on optical fundus images. Background Art

[0002] The retina is an important sensory organ. Retinal function is closely related to human vision, and regulating retinal blood and nutrition is the key to maintaining retinal function and vision. At present, although significant progress has been made in the study of retinal biology, there is little information about retinal biomechanics. Studying blood flow in the retina is very important for characterizing health and disease conditions. The retina is composed of a complex array of nerve cells, which are affected by stress and strain from blood vessels, and blood flow is related to nutrient delivery and waste removal. In the retina, as in the brain, there is a close correlation between cessation or reduction of blood perfusion and neuronal death. When there are early problems with blood vessels, such as transient blockage or ischemia caused by retinal branch vein occlusion, small amounts of bleeding in diabetic retinopathy, etc., vision is usually not affected. When the disease progresses and vision loss is discovered, it is usually too late, and visual function has been irreversibly damaged, so early analysis of early fundus vascular lesions is very important.

[0003] The retinal vascular imaging system based on AL-SLO can directly measure the vessel diameter, wall thickness, and blood flow velocity. However, there is currently no system that can obtain accurate and more retinal vascular hemodynamic parameters, such as intravascular cross-sectional pressure, flow, WSS and other hemodynamic parameters. Therefore, it is necessary to design a method to calculate the flow of blood in the blood vessels through simulation to obtain hemodynamic parameters such as pressure, flow and WSS, so as to provide a basis for future research on the correlation between these parameters and fundus diseases. Summary of the invention

[0004] The purpose of the present invention is to provide a retinal hemodynamic analysis and measurement system based on optical fundus images in view of the defects of the prior art.

[0005] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: A retinal hemodynamic analysis and measurement system based on optical fundus imaging, comprising:

[0006] An acquisition module is used to acquire the retinal fundus image to be tested and the retinal blood vessels and blood flow parameters, and to extract the contour curves of the blood vessel wall and lumen;

[0007] A modeling module, used to establish a three-dimensional model of the blood vessel and lumen of the corresponding blood vessel segment according to the extracted contour curve data of the blood vessel wall and lumen;

[0008] The simulation analysis module is used to perform two-way fluid-solid coupling simulation analysis on the established three-dimensional model of blood vessels and lumens to obtain a three-dimensional model of blood vessels and lumens with data interaction, thereby obtaining the hemodynamic data of the corresponding blood vessel segment.

[0009] Furthermore, the parameter data of retinal blood vessels include vessel wall thickness, lumen diameter, vessel diameter, wall-to-lumen ratio, and vessel wall cross-sectional area.

[0010] Furthermore, the blood flow parameters are the average blood flow velocity and blood flow rate, and their measurement adopts the following steps: use a high-frequency galvanometer to quickly scan the fundus blood vessels, capture the movement trajectory of blood cells in the blood vessels, process the collected blood flow signals, analyze the displacement and velocity of the blood cells, and obtain the average blood flow velocity and blood flow data of the blood vessels.

[0011] Furthermore, the modeling module uses the modeling software Unigraphics to establish a three-dimensional model of blood vessels and lumens through a sweeping function.

[0012] Furthermore, the simulation analysis module performs bidirectional fluid-solid coupling simulation analysis on the three-dimensional model of blood vessels and luminal cavities, including the following steps:

[0013] (1) Importing 3D models of blood vessels and lumen;

[0014] (2) Establish a vascular wall structure model and mesh the vascular wall;

[0015] (3) Meshing the blood and setting boundary conditions;

[0016] (4) Through the laminar flow model of blood non-Newtonian fluid, the velocity change curves of the blood inlet and outlet are obtained, and a dynamic grid is set at the fluid-solid interface to perform smoothing and recasting functions. After fluid-solid coupling, the displacement and pressure data transmission between the blood and the blood vessel wall is realized, thereby obtaining the hemodynamic data of the corresponding blood vessel segment.

[0017] Furthermore, unstructured grids are used to mesh the blood vessel wall, and three-layer boundary layer grids are divided at the fluid-solid interface.

[0018] Furthermore, tetrahedral meshes are used for blood mesh division, and three-layer boundary layer meshes are set at the fluid-solid interface, and mesh encryption is performed at the inlet and outlet.

[0019] Furthermore, the laminar flow model adopts the Lamina laminar flow model, and the control equation is formula 1:

[0020] (1)

[0021] Among them: ∇ is an operator, is the shear stress, ∇p is the pressure, ∂ is the partial derivative, velocity field u, time t;

[0022] The inlet boundary condition of the model was set as velocity inlet, and the user-defined function was compiled using C language to describe the curve of blood inlet velocity changing with time. The outlet boundary condition was set as pressure outlet, and the pressure value was the average value of retinal vascular pressure.

[0023] Furthermore, the hemodynamic data include pressure, blood flow velocity and wall shear stress at the cross section of the blood vessel.

[0024] The present invention uses an adaptive optical fundus impact system to measure retinal blood flow and vascular parameters. Based on the fluid-solid coupling theory and taking into account the elasticity of the vascular wall, a hemodynamic simulation analysis of the blood and blood vessels is performed to obtain retinal hemodynamic parameters that are difficult to directly observe with instruments.

[0025] Compared with the prior art, the present invention has the following beneficial effects:

[0026] The present invention uses Fluent and Transient Structural in the finite element software ANSYS to perform bidirectional fluid-solid coupling simulation analysis on the three-dimensional model of fundus blood vessels, which can quickly and accurately obtain hemodynamic data closely related to clinical practice, such as vascular cross-sectional pressure, flow, WSS, etc., and judge whether the parameters are within the normal value range, so as to timely discover fundus vascular lesions and help patients enter in-depth examinations as soon as possible and adopt corresponding follow-up treatment plans. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 The present invention is a flowchart of a retinal hemodynamics analysis and measurement system based on an adaptive optical fundus imaging system according to an embodiment.

[0028] Figure 2 is an embodiment retinal fundus image.

[0029] Figure 3 is a blood vessel image of the embodiment for measuring blood flow velocity.

[0030] Figure 4 is a blood vessel signal that changes continuously in the embodiment.

[0031] Figure 5 Schematic diagram of blood flow signals before and after registration in the embodiment.

[0032] Figure 6 : is the blood flow signal S obtained by subtracting the image after filtering from the image before filtering in the embodiment.

[0033] Figure 7 Schematic diagram of stripe angles in the embodiment.

[0034] Figure 8 is the movement trajectory of a single cell in the example.

[0035] Fig. 9 Graph showing the blood inlet velocity versus time in the embodiment.

[0036] Fig.10 3D blood vessel model established in the embodiment (a is a schematic diagram of the whole blood portion mesh (the left side is the inlet, the right side is the outlet); b is a magnified schematic diagram of the boundary layer mesh at the blood portion inlet; c is a schematic diagram of the whole blood vessel wall portion mesh (the left side is the inlet, the right side is the outlet); d is a magnified schematic diagram of the boundary layer mesh at the blood vessel wall portion inlet). DETAILED DESCRIPTION

[0037] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0038] It should be noted that the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or apparatus comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or apparatuses.

[0039] A retinal hemodynamic analysis and measurement system based on an adaptive optical fundus imaging system, the system adopts the following steps:

[0040] The parameters of the retinal blood vessels and blood flow to be measured are obtained based on the commercial Mona Ⅱa adaptive optical fundus imaging system (AO-SLO);

[0041] The contour curves of the vascular wall and lumen were extracted from the data of vascular wall thickness and retinal fundus images derived from the commercial Mona Ⅱa adaptive optical fundus imaging system;

[0042] The extracted vascular wall and lumen data were imported into the modeling software Unigraphics (UG) to establish the corresponding vascular model;

[0043] Based on Fluent in finite element software ANSYS, the measured blood flow velocity is used as the fluid inlet boundary condition through user defined functions (UDF) to perform hemodynamic simulation analysis;

[0044] Based on the System Coupling module, the Fluent and Transient Structural modules are coupled to achieve bidirectional fluid-solid coupling between blood and blood vessel walls;

[0045] To obtain data that are difficult to measure with instruments, such as pressure in the blood vessel section, WSS, and blood flow rate.

[0046] It also includes obtaining the length, diameter, and wall thickness of the blood vessel through manual measurement based on the image of the blood vessel to be measured;

[0047] According to the AO-SLO blood flow velocity measurement system, the blood flow velocity of each section in the blood vessel to be measured within a period of time is obtained: a high-frequency galvanometer is used to quickly scan the fundus blood vessels to capture the movement trajectory of blood cells in the blood vessels. The sampling time can last for several seconds. The algorithm processes the collected blood flow signals and can automatically analyze the displacement and velocity of millions of blood cells, the average blood flow velocity and blood flow of the sampled blood vessels, and other data;

[0048] Import the blood vessel model created in the 3D modeling software UG into ANSYS Workbench;

[0049] The curve of measured blood velocity changing with time is used as the inlet condition and the average pressure of retinal blood vessels is used as the outlet condition for simulation.

[0050] The steps of the AO-SLO blood flow velocity measurement method are as follows: use a high-frequency galvanometer to quickly scan the fundus blood vessels to capture the movement trajectory of blood cells in the blood vessels. The sampling time can last for several seconds. The algorithm processes the collected blood flow signals and can automatically analyze the displacement and speed of millions of blood cells, the average blood flow velocity and blood flow of the sampled blood vessels and other data.

[0051] Constructing a three-dimensional model of fundus blood vessels includes the following steps: obtaining images of retinal blood vessels to be measured by a commercial Mona Ⅱa AO-SLO system, and measuring retinal blood vessel parameters by marking the inner and outer diameters of the blood vessels, including wall thickness, lumen diameter, blood vessel diameter, wall-to-lumen ratio, and cross-sectional area of ​​the blood vessel wall; obtaining the blood flow velocity of each section in the blood vessel to be measured over a period of time according to the AO-SLO blood flow velocity measurement system; extracting the contour curves of the blood vessel wall and lumen through the derived data such as the blood vessel wall thickness, the angle of the movement trajectory of the collected blood cells, and the angle of the blood vessel shape and the retinal fundus image; and establishing the corresponding three-dimensional blood vessel model through scanning and modeling after denoising and quality inspection of the extracted blood vessel wall and lumen data.

[0052] The bidirectional fluid-structure interaction simulation between blood and the tube wall includes the following:

[0053] Due to the coupling between blood and blood vessel wall, the fluid domain (blood) does not blend or stack with the solid domain (blood vessel wall). The coupling between the two only occurs at the interface between blood and blood vessel wall, which is described by ALE algorithm.

[0054] The movement of blood (such as flow rate and pressure) will exert a certain fluid force on the blood vessel wall, thereby affecting the shape and movement of the blood vessel. The force generated by the deformation of the blood vessel can also affect the movement of blood. Therefore, the force generated by the blood in the fluid domain and the force generated by the deformation of the blood vessel in the solid domain are simultaneously applied to the fluid-solid coupling interface of the blood-blood vessel three-dimensional model, thereby obtaining a data-interactive blood-blood vessel three-dimensional model.

[0055] The characteristics of simulation analysis and simulation include:

[0056] After importing the established 3D blood vessel model, the simulation analysis module is used to perform bidirectional fluid-solid coupling simulation analysis on the blood vessel and blood to obtain a data-interactive 3D blood vessel model, thereby obtaining the pressure, flow velocity and wall shear stress of the cross section inside the blood vessel.

[0057] Import the blood-vessel model and perform fluid-solid coupling between the blood vessels and the blood. First, establish the vascular wall (solid) structure model, and mesh the vascular wall. Use unstructured meshes, and divide three layers of boundary layer meshes at the fluid-solid interface. Perform pre-processing such as meshing and setting boundary conditions on the blood (fluid). Use tetrahedral meshes, and in order to make the calculation results more accurate, set three layers of boundary layer meshes at the fluid-solid interface, and encrypt the meshes at the inlet and outlet. Set the blood as a viscous, incompressible non-Newtonian fluid. Through the laminar flow model of non-Newtonian fluid, obtain the velocity change curves of the blood inlet and outlet, set a dynamic mesh at the fluid-solid interface, and set the smoothing and recasting functions. After fluid-solid coupling, the displacement and pressure data transmission between the blood and the vascular wall is realized, so as to obtain the pressure of the cross section in the blood vessel, the blood flow velocity and the wall shear stress.

[0058] The specific steps are: taking the measured curve of blood velocity changing with time as the inlet condition and the average pressure of retinal blood vessels as the outlet condition to perform simulation;

[0059] First, pre-process the blood (fluid domain), import the blood-blood vessel model into the meshing software Meshing provided by Fluent, suppress the solid, mesh the blood, use tetrahedral mesh, and set three-layer boundary layer mesh at the fluid-solid interface to make the calculation results more accurate, and encrypt the mesh at the inlet and outlet;

[0060] Import the divided grid into Fluent and select the Lamina laminar flow model. Since blood is a non-Newtonian fluid and an incompressible non-Newtonian fluid, the control equation can be described as:

[0061]

[0062] Unlike Newtonian fluids, the viscosity coefficient of non-Newtonian fluids is no longer a constant, but changes with shear stress. The relationship between shear stress and shear rate is nonlinear.

[0063] The relationship between shear rate γ and shear stress ι can be described as:

[0064]

[0065] In the formula, μ α It is a quantity that changes with shear stress and approximately describes the viscosity of non-Newtonian fluids. It is called the apparent viscosity of non-Newtonian fluids.

[0066] The Carreau equation can be described as:

[0067]

[0068] Where n is the flow behavior index or non-Newtonian index, generally n<1; λ is the time constant, also known as relaxation time, which refers to the time required for the material to return to normal after the external force is released after deformation; η0 is the zero shear viscosity, also known as the first Newtonian viscosity. When the shear rate is very small, the polymer viscosity is large and usually a constant value, so it is often used as a standard for polymer viscosity; η ∞ It is the infinite shear viscosity, also known as the second Newtonian viscosity. When the shear rate is very large, the polymer macromolecular chain is prone to degradation, and sometimes it can be taken as 0;

[0069] The inlet boundary condition was set as velocity inlet, and the user-defined function (UDF) was compiled using C language to describe the curve of blood inlet velocity changing with time. The outlet boundary condition was set as pressure outlet, and the pressure value was the average value of retinal vascular pressure;

[0070] Set up a dynamic mesh at the fluid-solid interface and perform smoothing and recasting function settings;

[0071] Considering the viscous movement of blood, Womersley theory is used to describe it, which can be specifically expressed as follows:

[0072]

[0073] Where: R is the radius of the blood vessel, ω is the angular frequency of the fluid, ρ is the density of blood, and μ is the viscosity of the fluid;

[0074] Establish a vascular wall structure model, mesh the vascular wall, use unstructured meshes, and divide three layers of boundary layer meshes at the fluid-solid interface;

[0075] Based on the System Coupling module, the Fluent and Transient Structural modules are coupled to achieve data transmission of displacement and pressure between blood and blood vessel walls;

[0076] The required hemodynamic data, such as intravascular cross-sectional pressure, flow, WSS, etc., can be directly read out through CFD-post post-processing software. Among them, the essence of WSS is the friction between blood and vascular endothelium during blood flow, and its magnitude is directly related to blood viscosity, flow rate and vascular inner diameter. Its calculation formula is:

[0077]

[0078] Where: μ is viscosity, υ is velocity, and r is the vessel radius.

[0079] The curve of the measured blood flow velocity changing with time is used as the inlet condition, and the average pressure of the retinal blood vessels is used as the outlet condition for simulation.

[0080] The flow of blood in blood vessels is the same as that of ordinary fluids, and its movement laws conform to the general laws of fluid mechanics, namely the three conservation equations:

[0081] The mass conservation equation:

[0082]

[0083] In the formula, the speeds in the x, y, and z directions are represented by u x , u y , u z represents, t represents time, and ρ represents fluid density.

[0084] Momentum conservation equation:

[0085]

[0086] Where p represents the pressure on the fluid unit, τ xx , τ xy , τ xz It represents the component of viscous stress τ on the unit body, and fx, fy, fz represent the components of unit mass force in three directions.

[0087] Energy conservation equation:

[0088]

[0089] Where u, v, w represent the blood flow velocity in the x, y, and z directions, p is the pressure in the blood vessel, ρ is the blood density, and μ is the blood viscosity coefficient.

[0090] The flow of blood in blood vessels should satisfy the Navier-Stokes equation:

[0091]

[0092] In the formula, ρ f is the blood density, is the blood flow velocity, t is the time, p is the pressure, and μ is the viscoelastic parameter.

[0093] In addition to satisfying the above equations, the flow of blood in blood vessels also needs to satisfy the continuity equation:

[0094]

[0095] Fluent and Transient Structural software are used to perform fluid-structure coupling between blood vessels and blood, including the following steps:

[0096] Import the blood vessel model created in the 3D modeling software UG into ANSYS Workbench;

[0097] Use the Fluid Flow (Fluent) module to mesh the blood part (fluid), write user defined functions (UDFs) to set boundary conditions, and other pre-processing tasks;

[0098] The Transient Structural module is used to perform pre-processing work such as meshing the blood vessel wall (solid) and setting boundary conditions.

[0099] Finally, the System Coupling module is used to transfer the data of hemodynamics and structural mechanics in Fluent and Structural, and perform bidirectional fluid-structure coupling simulation analysis;

[0100] Example 1

[0101] like Figure 1 As shown, a retinal hemodynamic analysis and measurement system based on an adaptive optical fundus imaging system comprises the following steps:

[0102] Step 1: Obtain an image of the blood vessel to be tested through the Mona Ⅱa adaptive optical fundus imaging system (AO-SLO), obtain the blood vessel parameters through manual measurement, and obtain the blood flow velocity of each section in the blood vessel to be tested over a period of time through the AO-SLO blood flow velocity measurement system. Figure 2 Shown is a retinal fundus image.

[0103] Step 1.1: According to the different reflection brightness of the blood vessel wall, the inner and outer boundary positions of the blood vessel wall are marked using the measurement tool provided by the software, and the blood vessel parameters at the corresponding positions can be obtained, including: wall thickness, lumen diameter, blood vessel diameter, wall-to-lumen ratio, and blood vessel wall cross-sectional area. The specific data are shown in Table 1.

[0104] Table 1 Measured wall thickness, lumen diameter, etc. of sections 1 to 5

[0105]

[0106] Step 1.2: The calculation of blood flow velocity in blood vessels in line scan mode is divided into two parts: angle calculation of blood vessels and detection and angle measurement of blood cell movement trajectory in line scan data.

[0107] Step 1.2.1: Calculation of the angle of the blood vessel, such as Figure 3As shown, the blood vessels in the figure are the blood vessels whose blood flow velocity is to be measured. The blood flow data is collected based on the fundus blood vessel map displayed in real time by the camera. P is the manually selected line scan position when collecting data. This point is located in the blood vessel cavity. Line a is the fast scanning direction of the galvanometer, b is the straight line where the blood vessel is located, and Φ is the angle between lines ab. The line scan process records the signal of blood flow movement in the blood vessel at P in the scanning direction within a certain period of time. The steps for solving the blood vessel angle Φ are to crop the local blood vessel image S with P as the center. S rotates counterclockwise around P [0°, 179°], and solve the pixel value and of the row at each angle. The angle corresponding to the largest column is Φ or (180°-Φ). In terms of hardware, a galvanometer with a frequency of 16khz is used in the horizontal direction, and the galvanometer in the vertical direction is fixed. The scanning is continued for a certain period of time to obtain a continuously changing blood flow signal, such as Figure 4 shown.

[0108] Step 1.2.2: Detection of blood cell movement trajectory and angle measurement in line scan data. First, the quality of the blood flow data collected within a certain period of time is evaluated. The blood flow signal within a certain period of time is intercepted and calculated from top to bottom and from left to right in a sliding window manner. There is overlap between windows. The stripe consistency is judged, that is, the angle formed by the cell stripes calculated in each window and the horizontal direction is judged. Because the blood flow direction in the blood vessel is unidirectional and there is no reverse flow, if the β of the calculated M cell stripe directions are all in [0°, 90°) or (90°, 180°], it means that the blood flow signal quality in the blood vessel is good and the stripes are clear and consistent. In order to increase visibility, a stripe angle distribution diagram is drawn. If there is no high consistency, it is considered that the data quality is poor and feedback to the user is required to re-collect.

[0109] The blood flow data that meet the above quality evaluation are then registered, denoised and velocity calculated.

[0110] The registration process is based on the time slice of the original blood flow data. The window size is set to frequency / 100, and the blood flow data dataOneStrip on a time slice is taken out to perform data registration on dataOneStrip. Since the line scan mode will continue to scan the fundus signal for a certain period of time, there is a certain time difference, and the blood vessels will have a certain degree of displacement. In order to ensure the accuracy of subsequent operations, the blood flow signal needs to be registered to eliminate the impact of movement. The principle of registration is to align the static structure of the fundus. The ncc algorithm is used for registration. First, a reference frame is selected, and the middle (preferred) column of dataOneStrip is used as the reference frame. The displacement of other columns except the reference frame is calculated separately: the static structures located in the upper and lower parts of the blood vessels in the image are calculated by ncc calculations to obtain the amount of displacement required for each column on the time slice. The blood flow data is subjected to a cyclic displacement operation in the column dimension to obtain the blood flow signal after registration, such as Figure 5 As shown, it is a schematic diagram of blood flow signals before and after registration.

[0111] Denoising is to remove the static background. A one-dimensional mean filter is performed on dataOneStrip in a direction perpendicular to the line scan direction. The filter kernel size is 1*N to obtain the filtered image f. Then, the image before filtering is subtracted from the image after filtering and normalized to obtain the blood flow signal S, as shown in Figure 6 shown.

[0112]

[0113] For most capillary analysis, it is necessary to clip the blood flow source data, calculate the x-coordinate of the acquisition point P on the line scan data, extend a certain effective range to both sides, delete the line scan data outside the range, and perform non-maximum suppression in the direction perpendicular to the line scan. The formula is shown below, and the processing result is as follows: Figure 7 shown.

[0114]

[0115] Calculation using Radon transform Figure 7 The angle of the stripes in the middle is used to calculate the movement speed of the blood cells on the time slices. The process of Radon transform is similar to the calculation of the blood vessel angle Φ. The reverse rotation and the accumulation of row data are used to solve the maximum angle of the accumulated signal. The angle of the blood cell movement trajectory is searched clockwise based on the Time dimension, thereby solving the angle between the blood cell movement trajectory and the space dimension.

[0116] like Figure 8 As shown, the green oblique line is the motion trajectory of a single cell identified in this process. This part solves the speed of a single blood cell on the green oblique line of the line-scan blood flow signal.

[0117]

[0118] That is, after Δt time and Δx distance, the velocity of a single cell relative to the line scan direction a can be solved. Figure 3 , using formula (4), we can get the movement speed of a single blood cell in the blood vessel direction b.

[0119]

[0120] After calculating all the blood flow velocities, it is also necessary to determine the quadrant of the blood flow signal. Since the line scan is performed after the tracking is successful after the manual selection of the line scan position, and tracking cannot be performed during the line scan process, the angle of the initial several (2~5) frames of data will be calculated to obtain the quadrant mark of the blood flow signal in the blood vessel collected this time, the first quadrant or the second quadrant. Figure 8 The red oblique line in the middle is the noise identified as the cell movement trajectory, which does not conform to the movement trajectory of the quadrant marked by the blood flow signal or the ROI with a low signal-to-noise ratio. When calculating the average blood flow velocity, the results of such cases need to be eliminated to ensure the accuracy of the overall data.

[0121] The measured curve of the blood vessel entrance velocity changing with time is as follows Fig. 9 shown.

[0122] Step 2: Extract the contour curves of the vascular wall and lumen from the measured cross-sectional data and retinal fundus images, import the extracted vascular wall and lumen data into the modeling software Unigraphics (UG), and use the sweep function to establish the corresponding vascular model for subsequent calculations. The final vascular model is as follows: Fig.10 shown.

[0123] Step 3: Import the established three-dimensional blood vessel model into the finite element software ANSYS, and perform fluid-solid coupling between the blood vessel and blood through Fluent and TransientStructural software.

[0124] Step 3.1: Import the vascular model established in the 3D modeling software UG into ANSYS Workbench, and perform pre-processing such as meshing the blood part (fluid) and setting boundary conditions through the Fluid Flow (Fluent) module. Mesh the blood part (fluid) through the Mesh in the Fluid Flow (Fluent) module, using an unstructured grid with a minimum grid unit size of 3 μm, a boundary layer grid set to 3 layers, and a first layer grid height of 1 μm. The blood part model has a total of 1,086,442 unit grids.

[0125] Step 3.2: Use the Transient Structural module to perform pre-processing such as meshing the vascular wall part (solid) and setting boundary conditions. Use the Model in the Transient Structural module to mesh the vascular wall part (solid). Use an unstructured grid, with a minimum grid unit size of 3μm, a boundary layer grid set to 3 layers, and a first layer grid height of 1μm. The vascular wall model has a total of 625755 unit grids. Finally, output the data of these two modules to the System Coupling module for fluid-solid coupling calculations. The mesh diagram of the blood part and the vascular wall part is shown in the figure below. Fig.10 shown.

[0126] Step 4: Set the blood as a viscous, incompressible, non-Newtonian fluid and perform laminar flow in the blood vessels. Take the blood density as ρ = 1060 Kg / m 3 The viscosity of blood is calculated using the Carreau viscosity model. The blood vessel wall is assumed to be an isotropic, uniform, and incompressible elastic material, and the Young's modulus of the blood vessel is set to 0.5 MPa and the density is 1150 kg / m 3 , Poisson’s ratio is set to 0.45.

[0127] Step 5: Set the boundary conditions, set the inlet condition to velocity inlet, and the curve of inlet velocity changing with time is as follows: Fig. 9 As shown in the figure, the outlet condition is set to pressure outlet. Since the normal value of retinal artery systolic pressure is 60-80 mmHg and diastolic pressure is 30-40 mmHg, the outlet pressure is taken as the average value of 50 mmHg, which is 6666 Pa. The time step is set to 0.011 s, 60 steps are calculated, and the total time is 0.66 s.

[0128] Step 6: After the server calculation simulation, the pressure of the cross section in the blood vessel, the flow velocity and WSS (wall shear stress) in the blood vessel during the calculation period can be obtained. The pressure of the cross section in the blood vessel is shown in Table 2.

[0129]

[0130] The simulation results show that the wall shear stress generated by blood flow at 0.099s is in the range of 6-9Pa, and the wall shear stress generated by blood flow at 0.462s is in the range of 3-6Pa. The blood flow rate during the calculation time is 6.6432×10 -3 ml / min.

[0131] The analysis and measurement system of the present invention can be applied to the study of the correlation between retinal vascular diseases and hemodynamic parameters. The present invention can simulate the flow of blood in blood vessels through the finite element software ANSYS, and then obtain some hemodynamic parameters that are difficult to detect directly. Doctors can conduct research on the correlation and early warning of fundus vascular diseases based on these hemodynamic data.

[0132] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A retinal hemodynamic analysis and measurement system based on optical fundus imaging, characterized in that include: An acquisition module is used to acquire the retinal fundus image to be tested and the retinal blood vessels and blood flow parameters, and to extract the contour curves of the blood vessel wall and lumen; A modeling module, used to establish a three-dimensional model of the blood vessel and lumen of the corresponding blood vessel segment according to the extracted contour curve data of the blood vessel wall and lumen; The simulation analysis module is used to perform two-way fluid-solid coupling simulation analysis on the established three-dimensional model of blood vessels and lumens to obtain a three-dimensional model of blood vessels and lumens with data interaction, thereby obtaining the hemodynamic data of the corresponding blood vessel segment.

2. The retinal hemodynamics analysis and measurement system based on adaptive optical fundus imaging according to claim 1, characterized in that: The parameter data of the retinal blood vessels include vessel wall thickness, lumen diameter, vessel diameter, wall-to-lumen ratio, and vessel wall cross-sectional area.

3. The retinal hemodynamics analysis and measurement system based on adaptive optical fundus imaging according to claim 1, characterized in that: The blood flow parameters are average blood flow velocity and blood flow rate, and their measurement adopts the following steps: use a high-frequency galvanometer to quickly scan the fundus blood vessels, capture the movement trajectory of blood cells in the blood vessels, process the collected blood flow signals, analyze the displacement and velocity of the blood cells, and obtain the average blood flow velocity and blood flow data of the blood vessels.

4. The retinal hemodynamics analysis and measurement system based on adaptive optical fundus imaging according to claim 1, characterized in that: The modeling module uses the modeling software Unigraphics to establish a three-dimensional model of blood vessels and lumens through a sweeping function.

5. The retinal hemodynamics analysis and measurement system based on adaptive optical fundus imaging according to claim 1, characterized in that: The simulation analysis module performs bidirectional fluid-solid coupling simulation analysis on the three-dimensional model of blood vessels and luminal cavities, including the following steps: (1) Importing 3D models of blood vessels and lumen; (2) Establish a vascular wall structure model and mesh the vascular wall; (3) Meshing the blood and setting boundary conditions; (4) Through the laminar flow model of blood non-Newtonian fluid, the velocity change curves of the blood inlet and outlet are obtained, and a dynamic grid is set at the fluid-solid interface to perform smoothing and recasting functions. After fluid-solid coupling, the displacement and pressure data transmission between the blood and the blood vessel wall is realized, thereby obtaining the hemodynamic data of the corresponding blood vessel segment.

6. The retinal hemodynamics analysis and measurement system based on adaptive optical fundus imaging according to claim 5, characterized in that: The blood vessel wall is meshed using an unstructured grid, and three layers of boundary layer grids are divided at the fluid-solid interface.

7. The retinal hemodynamics analysis and measurement system based on adaptive optical fundus imaging according to claim 5, characterized in that: The blood is meshed using tetrahedral meshes, and three layers of boundary layer meshes are set at the fluid-solid interface, and mesh encryption is performed at the inlet and outlet.

8. The retinal hemodynamics analysis and measurement system based on adaptive optical fundus imaging according to claim 5, characterized in that: The laminar flow model adopts the Lamina laminar flow model, and the control equation is Formula 1: （1） Among them: ∇ is an operator, is the shear stress, ∇p is the pressure, ∂ is the partial derivative, velocity field u, time t; The inlet boundary condition of the model was set as velocity inlet, and the user-defined function was compiled using C language to describe the curve of blood inlet velocity changing with time. The outlet boundary condition was set as pressure outlet, and the pressure value was the average value of retinal vascular pressure.

9. The retinal hemodynamics analysis and measurement system based on adaptive optical fundus imaging according to claim 5, characterized in that: The hemodynamic data include pressure, blood flow velocity and wall shear stress in the cross section of the blood vessel.

Images