Multispectral light field imaging system, three-dimensional temperature field measurement method and measurement system

Abstract

The application discloses a kind of multispectral light field imaging system, three-dimensional temperature field measurement method and measurement system, multispectral light field imaging system includes: main lens, microlens array, first relay lens, dichroic mirror, second relay lens, third relay lens, first image sensor, first filter, second image sensor and second filter;The microlens array is located at the image surface of main lens;First relay lens and second relay lens, third relay lens respectively constitute the 1:1 relay lens group imaging system on spectral band one light path and spectral band two light path;Dichroic mirror is 45 with optical axis °;First image sensor is located at the focal plane of second relay lens, and second image sensor is located at the focal plane of third relay lens.The application realizes the collection of multispectral light field data, the light field data volume of each spectrum is invariable, and the measurement of flow field instantaneous three-dimensional temperature field can be realized.

Description

Technical Field

[0001] This invention relates to the field of fluid measurement technology, specifically to high-precision measurement of the three-dimensional temperature field of a fluid. Background Technology

[0002] High-precision measurement of the three-dimensional temperature field of high-temperature combustion gas in narrow turbine channels is of great significance for the optimization of turbine blade structure and performance improvement in aero-engines. However, the complex mechanical structure of turbine blades severely limits the application of traditional contact flow field measurement techniques. Among non-contact flow field temperature measurement techniques, laser-induced phosphorescence temperature measurement is a non-contact optical temperature measurement technique based on the phosphorescence thermal quenching effect. It has advantages such as low optical window requirements and no interference with the flow field structure, making it highly promising for measuring complex three-dimensional flow fields with limited optical windows. Currently, relatively mature laser-induced phosphorescence temperature measurement methods include the absolute intensity method, the lifetime decay method, and the intensity ratio method. Among these, the intensity ratio method measures temperature based on the relationship between the phosphorescence spectral intensity ratio and temperature. Compared to the absolute intensity method and the lifetime decay method, it is less affected by factors such as uneven distribution of phosphorescent particles, making it suitable for measuring the temperature field of dynamic flow fields. The position and spectral intensity ratio of phosphorescent tracer particles are used for the inversion calculation of the flow field temperature field. Therefore, multispectral imaging of the phosphorescent particle field and the calculation of the spectral intensity ratio are key core aspects of flow field temperature measurement.

[0003] For multispectral imaging of phosphorescent particle fields, laser-induced phosphorescent flow field temperature measurement systems primarily employ traditional imaging modes, which can only acquire multispectral images of phosphorescent particles within the focal plane. Therefore, they can only achieve instantaneous measurement of the two-dimensional temperature field of the flow field, and cannot acquire the instantaneous three-dimensional temperature field. In computational optics imaging technology, spectral light field imaging technology, through the coupling of spectral and light field imaging principles, can acquire multispectral light field information of three-dimensional objects. Combined with tomographic reconstruction algorithms, it can reconstruct the three-dimensional spectral intensity ratio of the target object. Therefore, combining spectral light field imaging technology with laser-induced phosphorescent intensity ratio temperature measurement technology holds promise for achieving the measurement of the instantaneous three-dimensional temperature field of the flow field.

[0004] Currently, there are two main types of spectral light field camera system structures: one is a spectral light field imaging system with a filter array coupled to the aperture of the main lens of the light field camera, and the other is a spectral light field imaging system with an RGB camera as the image sensor. The spectral light field imaging system with a coupled filter array has the following three problems: ① Because the aperture is evenly divided by the spectral filters in the filter array, the amount of light field data for each spectrum is inversely proportional to the number of spectra; ② Due to limitations in the manufacturing process of the filter array, multispectral data aliasing occurs at the boundaries of the sub-aperture images of the spectral light field camera, causing the spectral light field data at these boundaries to become invalid; ③ The greater the distance of the imaging region from the optical axis, the greater the aberration of the imaging system, and the size of the spectral sub-aperture image changes with the position of the imaging region. The spectral light field imaging system with a coupled color RGB camera has the following problems: the color RGB camera is coupled with a Bayer filter, which is only suitable for acquiring multispectral light field data of specific objects, thus limiting its applicability.

[0005] A search of existing technologies revealed Chinese invention patent number CN201910463400.9, entitled "Multispectral Temperature Measurement Method and System for High-Temperature Components Based on Light Field Camera." This invention employs a spectral light field imaging system using a light field camera's main lens aperture coupled with a filter array. This system structure suffers from the following four problems: ① While increasing the number of spectra, it reduces the amount of single-spectral light field data; ② Due to the aliasing of multispectral data at the boundaries of spectral sub-images, some image data becomes invalid; ③ Due to aberrations, the size of the spectral sub-aperture image changes with the position of the imaging area; ④ Because multispectral light field information is imaged on the same image sensor, the spectral data processing method is complex and the data processing flow is cumbersome.

[0006] For calculating the phosphorescence spectral intensity ratio, the measurement of the two-dimensional temperature field of the laser-induced phosphorescent flow field only requires taking the ratio pixel by pixel from the multispectral image of the phosphorescent particles to calculate the phosphorescence spectral intensity ratio of the two-dimensional plane. To calculate the spectral intensity ratio of the three-dimensional phosphorescent particle field based on the spectral light field image, a tomographic reconstruction algorithm is first used to reconstruct the spectral intensity distribution of the three-dimensional phosphorescent particle field. Then, the spectral intensity ratio of the three-dimensional phosphorescent particle field is calculated by taking the ratio pixel by pixel.

[0007] A search of existing technologies revealed Chinese invention patent number CN201710562875.4, entitled "A Three-Dimensional Flow Field Testing Method Based on a Dual Light Field Camera." This invention reduces the stretching effect of the reconstruction results by extracting the intersection of solutions. However, this invention requires an additional light field camera system, leading to increased optical window requirements. Therefore, it is not suitable for complex three-dimensional flow field measurements where the optical window is limited. Summary of the Invention

[0008] The first technical problem to be solved by the present invention is to provide a multispectral light field imaging system in which the amount of light field data in each spectrum remains unchanged.

[0009] The second technical problem to be solved by the present invention is to provide a method and system for measuring a three-dimensional temperature field with small intensity reconstruction error and high reconstruction quality.

[0010] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0011] This invention first provides a multispectral light field imaging system, comprising: a main lens, a microlens array, a first relay mirror, a dichroic mirror, a second relay mirror, a third relay mirror, a first image sensor, a first filter, a second image sensor, and a second filter; the microlens array is located at the image plane of the main lens; the distance between the first relay mirror and the equivalent plane of the microlens array is f+F, where f is the focal length of the microlens array and F is the focal length of the first relay mirror; the first, second, and third relay mirrors respectively form a 1:1 relay mirror image system in the first and second optical paths of the spectral bands; the dichroic mirror is at 45° to the optical axis; the first image sensor is located at the focal plane of the second relay mirror, and the second image sensor is located at the focal plane of the third relay mirror.

[0012] This invention also provides a method for measuring the three-dimensional temperature field of a laser-induced phosphorescent flow field, comprising:

[0013] Acquire the light field images captured by the first image sensor and the second image sensor respectively;

[0014] Based on the acquired light field image, the three-dimensional spectral intensity distribution corresponding to the light field image is obtained;

[0015] Based on the obtained three-dimensional spectral intensity distribution, the three-dimensional temperature distribution is obtained using the intensity ratio method.

[0016] Based on the acquired light field image, the corresponding three-dimensional spectral intensity distribution is obtained, including:

[0017] Based on the acquired light field image, establish the correspondence between the target object's position, the target object's spectral intensity, and the light field image:

[0018]

[0019] In the formula, λ represents the spectral wavelength, (m,n) are the pixel coordinates, and (i,j,k) are the voxel coordinates. This represents the grayscale value of the (m,n) pixel in the λ-spectral light field image. This represents the λ spectral intensity value of voxels (i,j,k). The weighting coefficient is the ratio of the grayscale contribution of voxel (i,j,k) to pixel (m,n) to the λ spectral intensity value of voxel (i,j,k).

[0020] Rearrange the three-dimensional discrete voxels of the control volume into a voxel column vector according to the order of "row first, then column last". Then, any original discrete voxel (i,j,k) is numbered (k-1)·I·J+(j-1)·I+i in the rearranged column vector. Rearrange the two-dimensional discrete pixels of the light field image into a pixel column vector according to the order of "row first, then column". Then, any original pixel (m,n) is numbered (n-1)·M+m in the rearranged pixel column vector. Let r = (n-1)·M+m, c = (k-1)·I·J+(j-1)·I+i, R = M·N, C = I·J·K. The correspondence between the control volume voxel column vector and the light field image pixel column vector is expressed as:

[0021]

[0022] The weighting coefficient matrix is ​​A λ The column vector of pixels in the light field image is B. λ The control volume voxel column vector is X λ Then equation (2) can be expressed as A λ X λ =B λ ;

[0023] Based on the characteristics of the tangent circular imaging region of the light field camera, the weighting coefficient matrix and A are realized by eliminating invalid pixels in the light field image and their corresponding weighting coefficients. λ X λ =B λ The equations are simplified; the criterion for invalid pixels is: pixels outside the tangent circle of the light field image are invalid pixels; by removing B... λ Invalid pixels and A λ The invalid weight coefficients corresponding to invalid pixels are used to achieve A. λ X λ =B λ The equations are simplified, and the preprocessed optical field tomography reconstruction equations are transformed into A. λ 2X λ =B λ 2;

[0024] The weighting coefficient matrix is ​​calculated using the ray tracing method. The formula for the ray tracing method is as follows:

[0025]

[0026] Where (x,y) are the coordinates of the light source, and (x',y') are the coordinates of the pixel where the light is incident. Let be the solid angle of the emitted ray. Let f be the solid angle of the incident ray; f be the focal length of the microlens unit; F be the focal length of relay mirror 1; S1 be the object distance; S2 be the image distance; (S x ,S y () represents the displacement of the center of the microlens array from the optical axis of the imaging system;

[0027] Chromatographic reconstruction of the three-dimensional spectral intensity of phosphorescent particles:

[0028] Preliminary reconstruction result X was obtained through hierarchical reconstruction using the SART algorithm. λ 1. The iterative calculation formula is:

[0029]

[0030] Where t is the iteration number, μ is the relaxation factor, p is the voxel number, b is the pixel gray value, and o is the pixel number. This is the ratio of the grayscale contribution of the p-voxel to the o-pixel to the spectral intensity value of the voxel. This represents the sum of the forward projections of the λ spectral intensity values ​​of all non-zero voxels within the line of sight of pixel o. This represents the sum of the weight coefficients of non-zero voxels within the line of sight of pixel o to pixel o. This represents the difference between the gray value of pixel o and the sum of the forward projections of the spectral intensity values ​​of all non-zero voxels within the pixel's line of sight.

[0031] The initial reconstruction result X1 was binarized using the Otsu's method, and the voxel intensity values ​​below the binarization threshold were set to zero to eliminate the overlap effect of the tracer particles.

[0032] Using the Otsu's method, for X λ The method for binarization is as follows:

[0033] Assuming the intensity threshold is τ, then X λ The voxel spectral intensity will be divided into two categories, A and B, where A is greater than τ and B is less than τ; traversing X λ 1. Voxel spectral intensity distribution: When the variances of classes A and B are maximized, the intensity threshold τ is X. λ The binarization threshold of 1.

[0034] The specific calculation method is as follows: First, use the frequency statistics function to calculate X. λ A probability table of voxel spectral intensity data, assuming the intensity is arranged in ascending order as E1, E2, E3…E n The probabilities of each intensity are p1, p2, p3...p n Set the intensity threshold to E. k (1≤k≤n), the probability of the spectral intensity of class A voxels is p A(k), the average voxel spectral intensity is m A (k), the probability of the spectral intensity of class B voxels is p B (k), the average voxel spectral intensity is m B (k), X λ The average intensity of the voxel spectrum is m G Then the variances of classes A and B are expressed as:

[0035]

[0036] Go through E1, E2, E3...E one by one n When the value of the expression shown in equation (5) is maximized, E k That is, X λ Binarization threshold of 1;

[0037] X λ 1. Intensity lower than E k The voxel spectral intensity is set to zero to obtain X-rays without particle overlap. λ 2;

[0038] Based on the correspondence between voxel column vector numbers and voxel 3D coordinates, X... λ 2. Restored to a three-dimensional voxel matrix X λ 3. Use Gaussian fitting to obtain X. λ 3. The voxel where the center of gravity of the three-dimensional connected mass is located.

[0039] X λ Any three-dimensional connected volume X within 3 λ The method for calculating the voxel containing the centroid of 3(i:i+Δi,j:j+Δj,k:k+Δk) is as follows:

[0040] Slicing the three-dimensional connected volume along a plane perpendicular to the Z-axis yields the X-axis. λ 3(i:i+Δi,j:j+Δj,n):

[0041] X3 λ (i:i+Δi,j:j+Δj,k:k+Δk)→X3 λ (i:i+Δi,j:j+Δj,n) (6)

[0042] Where, k <n<k+Δk;

[0043] Identify X using a two-dimensional Gaussian fitting function λ The centroid of 3(i:i+Δi,j:j+Δj,n):

[0044]

[0045] Among them, i <x<i+Δi,j<y<j+Δj,E λLet σ be the amplitude of the Gaussian distribution of voxel spectral intensity, (x0, y0, n) be the voxel coordinates of the centroid of the two-dimensional slice, and σ be the amplitude of the Gaussian distribution of voxel spectral intensity. x σ y These represent the full width at half maximum (FWHM) of the Gaussian distribution of the intensity of two-dimensional slice voxel spectra.

[0046] A two-dimensional Gaussian fitting function is applied to each slice of the three-dimensional connected volume to obtain the centroids of all slices; a one-dimensional Gaussian fitting is then performed on the voxel spectral intensities of all centroids to determine the centroid of the three-dimensional connected volume.

[0047]

[0048] Where (x0, y0, z0) is a three-dimensional connected volume X λ 3(i:i+Δi,j:j+Δj,k:k+Δk) Voxel coordinates of the centroid, σ z The half-maximum width of the Gaussian distribution of voxel spectral intensity;

[0049] Calculate X based on the centroid of the three-dimensional connected solid. λ Mark the 3D coordinates of all three-dimensional connected voxels containing the center of gravity, and remove X-axis voxels. λ X is obtained from all voxels other than the voxel where the center of gravity is located in 3. λ 4. Place X λ 4. Restore to a one-dimensional voxel column vector X λ 5. At the same time, A λ 2 will be simplified to A λ 3. The optical field tomography reconstruction equation is transformed into A λ 3X λ 5 = B λ 2;

[0050] For A λ 3X λ 5 = B λ 2. Perform a second SART tomography reconstruction to obtain X. λ 6;

[0051] Finally, based on the correspondence between the three-dimensional coordinates of the centroid voxel and the voxel column vector numbers, X... λ 6. Restored to the three-dimensional flow field control volume X λ 7.

[0052] The spectral light field imaging system proposed in this invention employs a relay mirror group coupled with a dichroic mirror in a cage-type light field camera system. This allows light field data of different spectral bands to be imaged onto different image sensors, achieving spectral data separation. This avoids the loss of single-spectral light field data, the aliasing of multispectral light field data, and the inconsistency in spectral sub-aperture image size caused by aberrations as the imaging area changes. Furthermore, when the required spectral wavelength changes, only the dichroic mirror and the corresponding filter need to be replaced, resulting in high system flexibility and convenient data processing.

[0053] Compared with the prior art, the present invention has the following advantages:

[0054] (1) The amount of light field data for each spectrum does not decrease with the increase of the number of spectra. Existing spectral light field cameras acquire multispectral light field data by adding a filter array to the aperture of the main lens of the light field camera. Since the aperture is evenly divided by the spectral filters of the filter array, the amount of light field data for each spectrum is inversely proportional to the number of spectra. This invention acquires multispectral light field data by adding a dichroic mirror inside the 1:1 relay lens group of the cage-type light field camera, while keeping the amount of light field data for each spectrum constant.

[0055] (2) There is no aliasing of multispectral light field data; Due to the limitations of the filter array process, existing spectral light field cameras have aliasing of multispectral light field data at the junction of each spectral sub-aperture image, which causes the spectral data at the junction of the sub-aperture images to fail; However, the light field data of different spectra in this invention are imaged on different image sensors, and there is no aliasing of multispectral light field data.

[0056] (3) There is no problem of inconsistent size of sub-aperture images of the spectrum due to aberrations; existing spectral light field cameras image multispectral light field data on the same image sensor. The greater the distance of the photographed area from the optical axis, the greater the aberration of the imaging system, and the size of each spectral sub-aperture image will become inconsistent; the light field data of different spectra of the present invention are imaged on different image sensors, and there is no problem of inconsistent size of each spectral sub-aperture image.

[0057] (4) The three-dimensional spectral intensity reconstruction process is more convenient; existing spectral light field cameras image multispectral light field data onto the same image sensor, and the light field data of each spectrum needs to be obtained by extracting sub-aperture images. In this invention, the light field data of different spectra are imaged onto different image sensors, and the light field data on each image sensor is the light field data of the same spectrum, which does not require the extraction of sub-aperture images, making the reconstruction calculation more convenient.

[0058] (5) Higher quality reconstruction results; The position and spectral intensity ratio of tracer particles are used for the inversion calculation of the flow field temperature field. Therefore, the accurate reconstruction of the three-dimensional spectral intensity ratio of the phosphorescent particle field is a key core link in flow field temperature measurement. When the SART algorithm reconstructs the three-dimensional particle field of optical field imaging in tomographic reconstruction, the reconstruction results are severely stretched along the depth direction of the control volume, resulting in a large error in the intensity reconstruction results, which seriously affects the accuracy of the flow field measurement results. This invention utilizes the characteristic that the particle intensity reconstruction results of the SART algorithm are approximately Gaussian distributed along the stretching direction. Based on the principle of Gaussian fitting to find the center of gravity of three-dimensional connected particles, a three-dimensional particle field reconstruction method combining Gaussian fitting positioning technology and SART algorithm is proposed. The initial three-dimensional particle field is reconstructed using the SART algorithm. Then, the overlap effect between tracer particles in the particle field is eliminated by the maximum inter-class variance method. Next, the voxel where the center of gravity of the stretched connected particles is located is located by the Gaussian fitting principle. The intensity value of the voxel where the center of gravity is located is reconstructed again by the SART algorithm, thereby reducing the intensity reconstruction error and improving the reconstruction quality.

[0059] (6) Realizing the measurement of the instantaneous three-dimensional temperature field of the flow field; Existing laser-induced phosphorescence flow field temperature measurement systems mainly adopt traditional imaging modes, which can only acquire multispectral images of phosphorescent particles within the focal plane. Therefore, they can only achieve instantaneous measurement of the two-dimensional temperature field of the flow field and cannot acquire the instantaneous three-dimensional temperature field of the flow field. This invention proposes a method for measuring the instantaneous three-dimensional temperature field of the flow field by combining spectral light field imaging technology with laser-induced phosphorescence intensity ratio method temperature measurement technology. Attached Figure Description

[0060] Figure 1 A method and system for measuring the three-dimensional temperature field of laser-induced phosphorescent flow based on a spectroscopic optical field camera;

[0061] Figure 2 Schematic diagram of a spectroscopic optical field imaging system;

[0062] Figure 3 Schematic diagram of the tangent circle imaging region of a light field camera;

[0063] Figure 4 Flowchart of a method for three-dimensional spectral intensity reconstruction by combining Gaussian fitting localization technique with SART algorithm for spectral light field imaging. Detailed Implementation

[0064] The present invention will now be described in detail with reference to the accompanying drawings:

[0065] The technical solution of this invention is as follows Figure 1As shown, a 355nm laser beam is expanded into a volume laser by a laser beam expander to excite phosphorescent particles in the flow field under test. Two image sensors of a spectroscopic laser-induced phosphorescence spectral imaging system simultaneously acquire phosphorescence field images of phosphorescent particles in two different bands within the same flow field region under test. Using the spectral field imaging three-dimensional spectral intensity reconstruction algorithm combining the Gaussian fitting localization technique and the SART algorithm proposed in this invention, the phosphorescence field images of the two bands are solved simultaneously to reconstruct the spectral intensity distribution of the three-dimensional phosphorescent particle field. The intensity ratio of the three-dimensional phosphorescent particle field is calculated using the intensity ratio method. Based on the phosphorescence spectral intensity ratio temperature calibration results, the three-dimensional temperature field distribution of the flow field under test is solved.

[0066] Based on the above technical solutions, the present invention can be divided into three parts, namely (1) the structure of the spectroscopic spectral light field imaging system; (2) the three-dimensional spectral intensity reconstruction algorithm of spectral light field imaging that combines Gaussian fitting positioning technology with SART algorithm; and (3) the reconstruction of the three-dimensional temperature field of the flow field based on the intensity ratio method.

[0067] Structure of a Spectroscopic Light Field Imaging System

[0068] The structure of a spectroscopic light field imaging system is as follows: Figure 2 As shown, the system includes a main lens, a microlens array, a relay mirror 1, a dichroic mirror, relay mirrors 2-1 and 2-2, an image sensor 1 and filter 1, and an image sensor 2 and filter 2. The microlens array is located at the image plane of the main lens; the focal length of the microlens unit is assumed to be f, and the focal length of relay mirror 1 is assumed to be F; the distance between relay mirror 1 and the equivalent plane of the microlens array is f+F; relay mirror 1, relay mirrors 2-1, and relay mirror 2-2 form a 1:1 relay mirror image system in the optical paths of spectral band 1 and spectral band 2, respectively; the dichroic mirror is at 45° to the optical axis; and image sensors 1 and 2 are located at the focal planes of relay mirrors 2-1 and 2-2, respectively.

[0069] A three-dimensional spectral intensity reconstruction algorithm combining Gaussian fitting localization technique and SART algorithm for spectral light field imaging.

[0070] The position and spectral intensity ratio of phosphorescent tracer particles are used for the inversion calculation of the flow field temperature field. Therefore, the calculation of the three-dimensional spectral intensity ratio of phosphorescent particles is a key step in flow field temperature measurement. This invention first constructs a spectral light field tomography reconstruction model based on the relationship between the position of phosphorescent tracer particles, the spectral intensity of phosphorescent tracer particles, and the light field image. The weighting coefficient matrix of the spectral light field tomography reconstruction model is calculated based on ray tracing. Then, the spectral light field tomography reconstruction model is simplified by combining the characteristics of the tangent circle imaging region of the light field camera. Finally, the stretching effect of the three-dimensional spectral intensity reconstruction result of phosphorescent particles is reduced based on Gaussian fitting localization technology.

[0071] (1) Spectral optical field tomography reconstruction model

[0072] like Figure 2 As shown, light rays emitted from the target object in different directions are converged by the main lens, subjected to secondary imaging by the microlens array, expanded by the relay mirror 1, and split by the dichroic mirror. Light rays of wavelength λ1 are converged by the relay mirror 2-1, filtered by the filter 1, and then incident on different pixels of the image sensor 1, completing photoelectric conversion. Similarly, light rays of wavelength λ2 are converged by the relay mirror 2-2, filtered by the filter 2, and then incident on different pixels of the image sensor 2, completing photoelectric conversion. The pixel position information and grayscale information of light field image 1 and light field image 2 correspond to the direction information, wavelength information, and spectral intensity information of the light rays emitted by the target object, respectively. Based on this correspondence, a relationship can be established between the target object's position, the target object's spectral intensity, and the light field image.

[0073]

[0074] In the formula, λ represents the spectral wavelength, (m,n) are the pixel coordinates, and (i,j,k) are the voxel coordinates. This represents the grayscale value of the (m,n) pixel in the λ-spectral light field image. This represents the λ spectral intensity value of voxels (i,j,k). The weighting coefficient is the ratio of the grayscale contribution of voxel (i,j,k) to pixel (m,n) to the λ spectral intensity value of voxel (i,j,k).

[0075] For ease of calculation, the three-dimensional discrete voxels of the control volume are rearranged into voxel column vectors in the order of "row first, then column, then column last". Thus, any original discrete voxel (i,j,k) is numbered (k-1)·I·J+(j-1)·I+i in the rearranged column vector. Simultaneously, the two-dimensional discrete pixels of the light field image are rearranged into pixel column vectors in the order of "row first, then column". Thus, any original pixel (m,n) is numbered (n-1)·M+m in the rearranged pixel column vector. Let r = (n-1)·M+m, c = (k-1)·I·J+(j-1)·I+i, R = M·N, and C = I·J·K. The correspondence between the control volume voxel column vector and the light field image pixel column vector can be expressed as follows:

[0076]

[0077] The weighting coefficient matrix is ​​A λ The column vector of pixels in the light field image is B. λ The control volume voxel column vector is X λ Then equation (2) can be expressed as A λ X λ =B λ Through the pixel grayscale information B of the light field image λReconstructing the voxel positions and spectral intensity information of tracer particles in the control volume is an inverse problem that is typically solved iteratively using the SART algorithm, which offers greater stability and parallelism. However, the SART algorithm requires repeated calls to the weighting coefficient matrix during computation, and the optical field weighting coefficient matrix occupies a large amount of memory, making the reconstruction process relatively time-consuming.

[0078] To reduce memory usage without repeatedly calculating the weighting coefficient matrix, this paper considers the characteristics of the tangent circular imaging region of the light field camera (see...). Figure 3 By removing invalid pixels from the light field image and their corresponding weight coefficients, the weight coefficient matrix and A are realized. λ X λ =B λ The equations are simplified to reduce the time required for tomography calculations. The criterion for invalid pixels is that, regardless of whether the subject is in focus or not, the image formed by the light field camera will not exceed [the specified range]. Figure 3 The tangent circle region is shown; therefore, pixels outside the tangent circle in the light field image are invalid pixels. By removing B... λ Invalid pixels and A λ The invalid weight coefficients corresponding to invalid pixels can achieve A. λ X λ =B λ The simplification of the equations reduces the computational load of tomography, thereby reducing the time required for tomography reconstruction. The preprocessed optical field tomography reconstruction equations are transformed into A λ 2X λ =B λ 2.

[0079] (2) Obtaining the weight coefficient matrix

[0080] The weight coefficient matrix is ​​calculated using ray tracing. The formula for ray tracing is as follows:

[0081]

[0082] Where (x,y) are the coordinates of the light source, and (x',y') are the coordinates of the pixel where the light is incident. Let be the solid angle of the emitted ray. Let f be the solid angle of the incident ray; f be the focal length of the microlens unit; F be the focal length of relay mirror 1; S1 be the object distance; S2 be the image distance; (S x ,S y ) represents the displacement of the center of the microlens array from the optical axis of the imaging system.

[0083] (4) Three-dimensional spectral intensity tomography reconstruction of phosphorescent particles

[0084] To address the stretching problem in 3D particle field reconstruction using the SART algorithm in spectral light field imaging, this paper optimizes the SART algorithm's reconstruction results by using Gaussian fitting to find the 3D connected mass center, and by removing X... λ Ineffective voxels and A λ In section 2, the invalidity weight coefficients corresponding to invalid voxels are used to achieve A. λ 2X λ =B λ Further simplification of the equations eliminates the stretching phenomenon and improves the accuracy of reconstructing the three-dimensional position and intensity of the tracer particles.

[0085] Due to X λ Since X is an unknown quantity, it needs to be reconstructed hierarchically using the SART algorithm to obtain a preliminary reconstruction result. λ 1. Its iterative calculation formula is as follows:

[0086]

[0087] Where t is the iteration number, μ is the relaxation factor, p is the voxel number, b is the pixel gray value, and o is the pixel number. This is the ratio of the grayscale contribution of the p-voxel to the o-pixel to the spectral intensity value of the voxel. This represents the sum of the forward projections of the λ spectral intensity values ​​of all non-zero voxels within the line of sight of pixel o. This represents the sum of the weight coefficients of non-zero voxels within the line of sight of pixel o to pixel o. This represents the difference between the gray value of pixel o and the sum of the forward projections of the spectral intensity values ​​of all non-zero voxels within the pixel's line of sight.

[0088] Because the SART algorithm severely stretches the reconstruction results along the depth direction, multi-particle field reconstruction results are prone to particle stretching and overlap, meaning that two or more tracer particles become the same three-dimensional connected body due to the stretching effect. Therefore, before extracting the three-dimensional connected body centroid through Gaussian fitting, this paper uses the Otsu method (also known as the maximum inter-class variance method) to perform X... λ 1. Binarization is performed, setting the voxel intensity values ​​below the binarization threshold to zero to eliminate the overlap effect of tracer particles. X is calculated based on the OTSU algorithm. λ The binarization thresholding method is as follows: assuming the intensity threshold is τ, then X λ The voxel spectral intensity will be divided into two categories, A (greater than τ) and B (less than τ), and the X-ray diffraction will be iterated. λ 1. Voxel spectral intensity distribution: When the variances of classes A and B are maximized, the intensity threshold τ is X. λ The binarization threshold is 1. The specific calculation method is as follows: First, use the frequency statistics function to statistically analyze X. λProbability table of voxel spectral intensity data of 1, assuming that the intensities are arranged in ascending order as E1, E2, E3…E n , and the probabilities of each intensity are p1, p2, p3…p n ; set the intensity threshold as E k (1≤k≤n), the probability of voxel spectral intensity of class A is p A (k), and the average value of voxel spectral intensity is m A (k), the probability of voxel spectral intensity of class B is p B (k), and the average value of voxel spectral intensity is m B (k), X λ The average value of voxel spectral intensity of 1 is m G ; then the variances of the two classes A and B can be expressed as,

[0089]

[0090] Traverse E1, E2, E3…E one by one n , when the value of the expression shown in formula (5) is the largest, E k is X λ The binarization threshold of 1.

[0091] Set the intensities of the voxels in X λ 1 that are lower than E k to zero, and then the X without particle overlap can be obtained λ 2.

[0092] According to the corresponding relationship between the voxel column vector number and the voxel three-dimensional coordinates, restore X λ 2 to the three-dimensional matrix X​​​​​​​​​​​​​​​​​​​​​​​​​Let σ be the amplitude of the Gaussian distribution of voxel spectral intensity, (x0, y0, n) be the voxel coordinates of the centroid of the two-dimensional slice, and σ be the amplitude of the Gaussian distribution of voxel spectral intensity. x σ y Let be the half-maximum widths (WHMs) of the Gaussian distribution of the voxel spectral intensities in the two-dimensional slices. A two-dimensional Gaussian fitting function is applied to each slice of the three-dimensional connected body to obtain the centroids of all slices. Finally, a one-dimensional Gaussian fitting is performed on the voxel spectral intensities of all centroids to determine the centroid of the three-dimensional connected body; the calculation expression is as follows:

[0097]

[0098] Where (x0, y0, z0) is a three-dimensional connected volume X λ 3(i:i+Δi,j:j+Δj,k:k+Δk) Voxel coordinates of the centroid, σ z The half-maximum width of the voxel spectral intensity Gaussian distribution.

[0099] Based on the method for calculating the center of gravity of a three-dimensional connected mass, calculate X. λ Mark the 3D coordinates of all three-dimensional connected voxels containing the center of gravity, and remove X-axis voxels. λ X is obtained from all voxels other than the voxel where the center of gravity is located in 3. λ 4. Place X λ 4. Restore to a one-dimensional voxel column vector X λ 5. At the same time, A λ 2 will be simplified to A λ 3. Therefore, the optical field tomography reconstruction equation is transformed into A λ 3X λ 5 = B λ 2. Regarding A λ 3X λ 5 = B λ 2. Perform a second SART tomography reconstruction to obtain X. λ 6. Finally, based on the correspondence between the three-dimensional coordinates of the centroid voxel and the voxel column vector numbers, X... λ 6. Restored to the three-dimensional flow field control volume X λ 7.

[0100] In summary, the flowchart of the three-dimensional spectral intensity reconstruction method combining Gaussian fitting localization technology and SART algorithm for spectral light field imaging is as follows: Figure 4 As shown,

[0101] Reconstructing the three-dimensional temperature field of the flow field based on the intensity ratio method

[0102] Based on the three-dimensional spectral intensity reconstruction process, the light field images of phosphorescence spectra λ1 and λ2 are solved separately to reconstruct the three-dimensional spectral intensity X of the phosphorescence particle field λ1 and λ2. λ1 7 and X λ2 7. Regarding X λ1 7 and Xλ2 7. Taking the ratio of each non-zero voxel, the spectral intensity ratio of the three-dimensional phosphorescent particle field is calculated using the following formula:

[0103]

[0104] In the formula, (i,j,k) are the three-dimensional coordinates of the voxel, and R (i,j,k) The phosphorescence spectral intensity ratio of (i,j,k) voxels;

[0105] The relationship between phosphorescence spectral intensity ratio and temperature was obtained through temperature calibration experiments.

[0106] T (i,j,k) =f(R) (i,j,k) (10)

[0107] In the formula, T (i,j,k) Represents the temperature of voxels (i,j,k);

[0108] Based on equation (10), the spectral intensity ratio R of the three-dimensional phosphorescent particle field is obtained. (i,j,k) Solve for the instantaneous three-dimensional temperature field of the flow field.

Claims

1. A method for measuring the three-dimensional temperature field of a laser-induced phosphorescent flow field based on a multispectral light field imaging system, wherein the multispectral light field imaging system includes: The system comprises a main lens, a microlens array, a first relay lens, a dichroic mirror, a second relay lens, a third relay lens, a first image sensor, a first filter, a second image sensor, and a second filter; the microlens array is located at the image plane of the main lens; the distance between the first relay lens and the equivalent plane of the microlens array is... f+F , f The focal length of the microlens array. F The focal length of the first relay mirror; the first, second, and third relay mirrors form a 1:1 relay mirror image system in the optical path of spectral band one and the optical path of spectral band two, respectively; the dichroic mirror is at 45° to the optical axis; the first image sensor is located at the focal plane of the second relay mirror, and the second image sensor is located at the focal plane of the third relay mirror. The method for measuring the three-dimensional temperature field of a laser-induced phosphorescent flow field is characterized by: Acquire the light field images captured by the first image sensor and the second image sensor respectively; Based on the acquired light field image, the three-dimensional spectral intensity distribution corresponding to the light field image is obtained; Based on the obtained three-dimensional spectral intensity distribution, the three-dimensional temperature distribution is obtained using the intensity ratio method; Based on the acquired light field image, the corresponding three-dimensional spectral intensity distribution is obtained, including: Based on the acquired light field image, establish a set of equations relating the target object's position, the target object's spectral intensity, and the light field image. Based on the characteristics of the tangent circular imaging region of the light field camera, the system of correspondence equations established in the previous step is simplified by eliminating invalid pixels in the light field image and their corresponding weight coefficients. Initial reconstruction was performed using the first SART algorithm; The initial reconstruction results are binarized using the Otsu's method, and the voxel spectral intensity values ​​below the binarization threshold are set to zero to obtain voxel column vectors without particle overlap. Based on the correspondence between the voxel column vector number and the three-dimensional coordinates of the voxel, the voxel column vector is restored to a three-dimensional voxel matrix, and Gaussian fitting is used to find the voxel where the center of gravity of the three-dimensional connected mass is located in the three-dimensional voxel matrix. Based on the centroid of the three-dimensional connected body, calculate the voxels containing the centroids of all three-dimensional connected bodies within the three-dimensional voxel matrix, mark the three-dimensional coordinates of the voxels containing the centroids, remove all voxels except those containing the centroids in the three-dimensional voxel matrix to obtain a simplified three-dimensional voxel matrix, restore the simplified three-dimensional voxel matrix to a simplified voxel column vector, and simultaneously convert the weight coefficient matrix to a simplified weight coefficient matrix to obtain a simplified light field tomography reconstruction equation. A second SART tomographic reconstruction is performed on the simplified tomographic reconstruction equation to obtain the target voxel column vector; Based on the correspondence between the three-dimensional coordinates of the centroid voxel and the voxel column vector number, the target voxel column vector is restored to the three-dimensional voxel matrix of the target flow field control volume.

2. The method for measuring the three-dimensional temperature field of a laser-induced phosphorescent flow field according to claim 1, characterized in that, The method for binarizing the initial reconstruction results of the SART algorithm using the Otsu's method is as follows: Assuming the intensity threshold is τ The voxel spectral intensities of the preliminary reconstruction results will then be divided into... A , B Two categories, among which A Greater than τ , B Less than τ ; Traverse the voxel spectral intensity distribution of the preliminary reconstruction results, when A, B When the variance of the two classes is at its maximum, the intensity threshold is... τ The threshold for binarization of the initial reconstruction results.

3. The method for measuring the three-dimensional temperature field of a laser-induced phosphorescent flow field according to claim 1, characterized in that, The established set of equations relating the target object's position, spectral intensity, and light field image is as follows: （1） In the formula, express λ Spectral light field image ( m,n The grayscale value of a pixel. λ Indicates the spectral wavelength, ( m,n () represents pixel coordinates; express( i,j,k voxels λ Spectral intensity value, ( i,j,k () represents voxel coordinates; express( i,j,k Voxel to ( m,n The grayscale contribution value of a pixel and ( i,j,k voxels λ The ratio of spectral intensity values ​​is the weighting coefficient; I, J, and K are respectively... i,j,k The upper limit represents the number of voxels; If the three-dimensional discrete voxels of the control volume are rearranged into a voxel column vector in the order of "row first, column second, column last", then any original discrete voxel ( i,j,k The columns in the rearranged column vector are numbered as follows: k- 1)· I · J +( j -1)· I + i Rearrange the two-dimensional discrete pixels of the light field image into a pixel column vector according to the "row-first, column-second" order, then any original pixel ( m,n The pixel column vector is numbered as ( n -1)· M + m ;Pick r =( n -1)· M + m , c =( k- 1)· I · J +( j -1)· I + i , R = M · N , C = I · J · K , M and N They are respectively m and n The upper limit represents the number of pixels in the light field image; therefore, the correspondence between the control volume voxel column vector and the light field image pixel column vector is expressed as: （2） The weighting coefficient matrix is A λ The column vector of pixels in the light field image is B λ The control volume voxel column vector is X λ Then equation (2) can be expressed as A λ X λ =B λ ; Based on the characteristics of the tangent circular imaging region of the light field camera, the simplification of the correspondence equations established in the previous step is achieved by eliminating invalid pixels and their corresponding weight coefficients in the light field image. The criterion for judging invalid pixels is as follows: Pixels outside the tangent circle of the light field image are invalid pixels; this is achieved by removing the pixel column vectors of the light field image. B λ Invalid pixels and weight coefficient matrix A λ The invalid weight coefficients corresponding to invalid pixels are used to achieve... A λ X λ =B λ The equations are simplified, and the preprocessed optical field tomography reconstruction equations are transformed into... A λ 2 X λ =B λ 2 .

4. The method for measuring the three-dimensional temperature field of a laser-induced phosphorescent flow field according to claim 3, characterized in that, The weighting coefficient matrix is ​​calculated using the ray tracing method. The formula for the ray tracing method is as follows: （3） in,( x,y ) represents the coordinates of the light source. x ， ,y ， ) represents the coordinates of the pixel where the ray is incident. ( ) is the solid angle of the emitted ray. ( ) is the solid angle of the incident ray; f The focal length of the microlens unit. F The focal length of the first relay lens. S 1 For object distance, S 2 For image distance, ( S x , S y ) represents the displacement of the center of the microlens array from the optical axis of the imaging system.

5. The method for measuring the three-dimensional temperature field of a laser-induced phosphorescent flow field according to claim 4, characterized in that, The initial reconstruction is performed using the first SART algorithm, including: Preliminary reconstruction results were obtained through tomographic reconstruction using the SART algorithm. X λ 1 The iterative calculation formula is: （4） in, t For the number of iterations, μ As a relaxation factor, p Number the voxels. b For pixel grayscale values, o Number the pixels. for p voxel pairs o The grayscale contribution value of a pixel and the voxel λ The ratio of spectral intensity values express o All non-zero voxels within the pixel's line of sight λ The sum of the forward projections of the spectral intensity values. express o Non-zero voxel pairs within the pixel's line of sight o The sum of the weighting coefficients of the pixels. express o Pixel grayscale value and all non-zero voxels within the pixel's field of view λ The difference between the sum of the forward projections of the spectral intensity values.

6. The method for measuring the three-dimensional temperature field of a laser-induced phosphorescent flow field according to claim 5, characterized in that, Will X λ 1 Medium strength or below X λ 1 Binarization threshold E k Setting the voxel intensity to zero yields a result where particle overlap is non-existent. X λ 2 Based on the correspondence between voxel column vector numbers and voxel 3D coordinates, X λ 2 Restored to a three-dimensional matrix X λ 3 ; X λ 3 Any three-dimensional connected body X λ 3 ( i : i+ i , j : j+ j , k : k+ k The method for calculating the voxel where the center of gravity is located is as follows: Slicing the three-dimensional connected volume along a plane perpendicular to the Z-axis yields... X λ 3 ( i : i+ i , j : j+ j , n ): （6） in, k < n < k+ k ; Identification using a two-dimensional Gaussian fitting function X λ 3 ( i : i+ i , j : j+ j , n The center of mass of ) （7） in, i < x < i+ i , j < y < j+ j , E λ voxels λ The amplitude of the Gaussian distribution of spectral intensity, x 0 , y 0 , n () represents the voxel coordinates of the centroid of the 2D slice. σ x , σ y Two-dimensional slice voxels λ The full width at half maximum (FWHM) of the Gaussian distribution of spectral intensity; Applying a two-dimensional Gaussian fitting function to each slice of the three-dimensional connected volume yields the centroids of all slices; for each voxel containing a centroid... λ The centroid of a three-dimensional connected volume is determined by one-dimensional Gaussian fitting of spectral intensity. （8） in,( x 0 , y 0 , z 0 () is a three-dimensional connected body X λ 3 ( i : i+ i , j : j+ j , k : k+ k The coordinates of the voxel where the center of gravity is located. σ z voxels λ The spectral intensity has a Gaussian distribution and a full width at half maximum (FWHM).

7. The method for measuring the three-dimensional temperature field of a laser-induced phosphorescent flow field according to claim 6, characterized in that, Based on the obtained three-dimensional spectral intensity distribution, the three-dimensional temperature distribution is obtained using the intensity ratio method, including: right λ1 , λ2 Three-dimensional spectral intensity of phosphorescence X λ1 7 and X λ2 7 By taking the ratio of each non-zero voxel, the spectral intensity ratio of the three-dimensional phosphorescent particle field is calculated: （9） In the formula, ( i,j,k ( ) represents the three-dimensional coordinates of a voxel. R （i,j,k） for( i,j,k The ratio of phosphorescent spectral intensity of voxels; The relationship between phosphorescence spectral intensity ratio and temperature was obtained through temperature calibration experiments. （10） In the formula, T （i,j,k） express( i,j,k The temperature of the voxel; Based on equation (10), the spectral intensity ratio of the three-dimensional phosphorescent particle field is obtained. R （i,j,k） Solve for the instantaneous three-dimensional temperature field of the flow field.

8. A three-dimensional temperature field measurement system for laser-induced phosphorescent flow field, comprising a pulsed laser, a laser beam expander, a spectral light field image acquisition system, and an image processing system, characterized in that, The image processing system calculates the instantaneous three-dimensional temperature field of the flow field based on the image acquired by the spectral light field image acquisition system, following the steps of the laser-induced phosphorescent flow field three-dimensional temperature field measurement method according to any one of claims 1-7.

Images