Light field tomography system based on dual-camera fusion and flame three-dimensional temperature field reconstruction method

Through the dual-camera fusion light field tomography system, combined with interpolation and ray tracing technology, the problem of large error in flame three-dimensional temperature field reconstruction is solved, and the temperature field reconstruction with high time resolution and high spatial resolution is achieved, which significantly improves the reconstruction. Accuracy.

CN119023076APending Publication Date: 2024-11-26HARBIN INST OF TECH
View PDF 0 Cites 3 Cited by

Patent Information

Application Number
CN202411182647.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-27
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

The existing flame three-dimensional temperature field reconstruction method has problems such as large reconstruction error and low time resolution, and high-precision temperature distribution reconstruction cannot be achieved.

Method used

Using a light field tomography system based on dual camera fusion, using RGB camera and high-speed camera combination, through interpolation and ray tracing technology, a reconstruction equation system is established to solve the bold spectral radiation intensity of each microbody in flame, thereby achieving high time resolution Temperature field reconstruction with rate and high spatial resolution.

Benefits of technology

The time resolution of temperature field reconstruction is improved, the number of light samples is increased, the reconstruction error is reduced, and high-precision three-dimensional temperature distribution reconstruction is achieved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119023076A_ABST
    Figure CN119023076A_ABST
Patent Text Reader

Abstract

The invention discloses a light field tomography system based on dual-camera fusion and a flame three-dimensional temperature field reconstruction method, and belongs to the technical field of flame temperature measurement. The problems that an existing flame three-dimensional temperature field reconstruction method is large in reconstruction error and low in reconstruction time resolution are solved. According to the method, the RGB camera and the high-speed camera are used for collecting flame light field images respectively, interpolation is carried out on two adjacent frames of images collected by the RGB camera in combination with the characteristics of the RGB camera and the high-speed camera, high-resolution images corresponding to all collection time points of the high-speed camera are obtained, then light tracing is carried out in combination with the interpolated images and the images collected by the high-speed camera, and the high-resolution images are obtained. And establishing a reconstruction equation set according to a ray tracing result, finally solving the reconstruction equation set to obtain blackbody spectrum radiation intensity, and converting the blackbody spectrum radiation intensity into blackbody temperature to complete three-dimensional temperature field reconstruction. The method can be applied to flame three-dimensional temperature field reconstruction.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of flame temperature measurement, and in particular relates to a light field tomography system based on dual-camera fusion and a flame three-dimensional temperature field reconstruction method. Background Art

[0002] Combustion diagnostic technology evaluates the working status and efficiency of combustion equipment by detecting and analyzing relevant parameters in the combustion process. The characteristic parameters of the combustion process usually include temperature, flow rate, pressure and concentration, among which temperature is an important parameter reflecting the combustion process. The temperature field measurement of high-temperature flames is an important testing technology in the fields of aerospace, energy, chemical industry, materials, environment, and mechanical manufacturing. The accurate measurement of the temperature field plays an important guiding role in actual production processes such as pollutant emission control, design optimization of combustion equipment, and monitoring of the combustion process. Light field imaging technology can simultaneously record the intensity and direction of flame radiation light, and collect measurement information through camera image detectors. This technology only requires a small number of cameras to reconstruct the three-dimensional temperature field distribution of the flame based on the light field image information.

[0003] However, light field images are usually collected using ordinary RGB cameras, which have the disadvantage of long exposure time and cannot achieve high temporal resolution parameter field reconstruction. On the other hand, light field images are only projections of spatial three-dimensional distribution at a certain viewing angle. The number of sampling rays for obtaining reconstruction information in a single light field imaging system is limited and cannot pass through the entire measurement area. The coefficient matrix in the equation set for reconstruction calculation is usually a sparse matrix, and the solution of the equation set is seriously ill-conditioned, so the error of temperature field reconstruction is large. Therefore, it is necessary to propose a new flame three-dimensional temperature field reconstruction method to solve the problems of large reconstruction error and low temporal resolution of reconstruction in existing methods. Summary of the invention

[0004] The purpose of the present invention is to solve the problems of large reconstruction error and low reconstruction time resolution of the existing flame three-dimensional temperature field reconstruction method, and propose a light field tomography system based on dual-camera fusion and a flame three-dimensional temperature field reconstruction method.

[0005] The technical solution adopted by the present invention to solve the above technical problems is:

[0006] According to one aspect of the present invention, a light field tomography system based on dual-camera fusion is provided, wherein the light field tomography system comprises a first light field tomography component and a second light field tomography component, wherein:

[0007] The first optical field tomography assembly includes a main lens, a microlens array, a relay lens and an image detector; and the main lens, the microlens array, the relay lens and the image detector share a common optical axis, that is, the center of the main lens, the center of the microlens array, the center of the relay lens and the center of the image detector are located on the same straight line;

[0008] The structure of the second light field tomography component is the same as that of the first light field tomography component, and both the first light field tomography component and the second light field tomography component are placed obliquely above the flame to be measured in the horizontal direction.

[0009] Furthermore, the image detector of the first light field tomography component is an RGB camera.

[0010] Furthermore, the image detector of the second light field tomography component is a high-speed camera.

[0011] Furthermore, the relay lens is composed of two identical lenses.

[0012] Further, the light field tomography system includes a plurality of cage plates and cage rods;

[0013] The cage plate is provided with through holes at the center and around the edges, and the primary lens and the relay lens are mounted on the cage plate through the through holes at the center of the cage plate;

[0014] The cage rods pass through the through holes around each cage plate.

[0015] According to another aspect of the present invention, a method for reconstructing a three-dimensional temperature field of a flame is provided, and the method specifically comprises the following steps:

[0016] Step 1: adjusting the main lens so that the first light field tomography component is focused on the flame to be measured, and using the RGB camera of the first light field tomography component to shoot the light field image of the flame to be measured; adjusting the main lens so that the second light field tomography component is focused on the flame to be measured, and using the high-speed camera of the second light field tomography component to shoot the light field image of the flame to be measured;

[0017] Step 2: interpolate the measured flame light field image taken by the RGB camera to obtain the measured flame light field interpolation image corresponding to each image taken by the high-speed camera;

[0018] Step 3: The space area where the flame to be measured is located is taken as the measurement area, and the measurement area is divided into various micro-elements;

[0019] According to the relative position between the light field tomography system and the measurement area, flame radiation ray tracing is performed starting from each pixel point on the light field image taken by the high-speed camera and the interpolation image;

[0020] Step 4: Establish a reconstruction equation group based on the radiation transfer equation and the flame radiation ray tracing results, and obtain the blackbody spectral radiation intensity of each flame microelement at each moment by solving the reconstruction equation group;

[0021] Step 5: Convert the blackbody spectral radiation intensity of each flame microelement into the temperature of each microelement in the measurement area to complete the reconstruction of the three-dimensional temperature field of the flame at each moment.

[0022] Furthermore, the interpolation of the measured flame light field image taken by the RGB camera adopts a cubic spline interpolation method.

[0023] Furthermore, the radiation transfer equation is specifically:

[0024] Considering the absorption, emission and scattering of the radiation transmission path medium:

[0025]

[0026] Among them, I bλ (s,s i ′) is the blackbody radiation intensity of the flame at position s and in the direction s′, I bλ (s) is the blackbody radiation intensity of the flame at position s, β λ (s) is the attenuation coefficient, unit is m -1 , κ λ (s) is the absorption coefficient, in m -1 , σ sλ (s) is the scattering coefficient, in m -1 ,Φ(s′ i ,s′) is s′ i The scattering phase function of the flame radiation incident from the direction to the s′ direction, Ω i is the integration interval;

[0027]

[0028] Where C1 is the first radiation constant, C2 is the second radiation constant, λ is the wavelength of the flame radiation, T(s) is the temperature of the flame at position s, and e is the base of the natural logarithm;

[0029] According to formula (8), the steady-state radiation transfer equation of the microelement is:

[0030]

[0031] Among them, I λ is the spectral radiation intensity of the flame radiation, I bλ is the blackbody spectral radiation intensity of the flame microelement, τ is the optical thickness of the flame radiation light passing through the flame microelement;

[0032] τ=κ·δ (15)

[0033] Among them, κ is the absorption coefficient of the flame microelement through which the flame radiation passes, and δ is the geometric length of the flame microelement through which the flame radiation passes.

[0034] Furthermore, the reconstruction equations are established based on the radiation transfer equation and the flame radiation ray tracing results, specifically:

[0035] For any image in the RGB camera image and the interpolated image, the total number of flame micro-elements that the radiation light corresponding to any pixel point m on the image passes through is recorded as n, and the radiation transmission equation of the flame radiation light corresponding to the pixel point m is:

[0036]

[0037] Among them, I λ (m) is the spectral radiation intensity of the flame radiation corresponding to pixel m, I bλ (i) is the blackbody spectral radiation intensity of the i-th flame element, i = 1, 2, ..., n, A m,i is the calculation coefficient;

[0038] Based on the flame radiation ray tracing results of step 3, the radiation transfer equations of all flame radiation rays are combined into a reconstruction equation group;

[0039]

[0040] Among them, b h,i is a vector composed of the spectral radiation intensity of the flame radiation light corresponding to each pixel in the image taken by the high-speed camera, b r,i is the vector of the spectral radiation intensity of the flame radiation light corresponding to each pixel in the interpolation image, x is the vector of the blackbody spectral radiation intensity of each flame microelement, A r is the coefficient matrix of RGB camera detection, A h is the coefficient matrix detected by the high-speed camera.

[0041] Furthermore, the calculation coefficient A m,i for:

[0042]

[0043] Among them, τ j It is the optical thickness of the jth flame element through which the flame radiation light corresponding to pixel point m passes.

[0044] The beneficial effects of the present invention are:

[0045] The present invention uses an RGB camera and a high-speed camera to collect flame light field images respectively, combines the characteristics of the RGB camera and the high-speed camera, interpolates two adjacent frames of images collected by the RGB camera, obtains high-resolution images corresponding to each acquisition time point of the high-speed camera, and then combines the interpolated image and the image collected by the high-speed camera to perform ray tracing, establishes a reconstruction equation group according to the ray tracing results, and finally solves the reconstruction equation group to obtain the blackbody spectral radiation intensity, and then converts the blackbody spectral radiation intensity into the blackbody temperature. Compared with the existing light field tomography temperature field reconstruction method, the present invention uses a double-cage imaging system combining an ordinary RGB camera with a high-speed camera, which can improve the time resolution of temperature field reconstruction, while increasing the number of light sampling and reducing reconstruction errors. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 is a schematic diagram of the structure of the light field tomography component;

[0047] The light field tomography system of the present invention includes two groups of light field tomography components, each group of light field tomography components is composed of a main lens 3, a microlens array 2, a relay lens 6 and an image detector 1. The structures of the two groups of light field tomography components are substantially the same, and the only difference is that the image detector 1 of one group of light field tomography components adopts a high-speed camera, and the image detector 1 of the other group of light field tomography components adopts an RGB camera. The relay lens 6 is an optical system composed of two identical lenses, and its optical principle is to generate an upright and enlarged virtual image of a close object, that is, the actual measurement object, at a farther distance, and can perform secondary imaging on the image formed by the microlens, and finally image the image on the image detector. Compared with the traditional light field camera, a relay lens 6 is added between the microlens array and the image detector to separate the two from each other, and the positions of the two can be flexibly adjusted.

[0048] The cage system is mainly composed of a cage plate 5 and cage rods 7. The center of the cage plate 5 has a through hole for installing optical elements, and the four corners of the cage plate 5 have through holes for fixing and passing through the cage rods 7. The four sides of the cage plate 7 can be connected to other optical platforms by installing bolts and connecting rods and other fasteners to adjust the position of the cage imaging system. The cage plate can slide freely along the cage rods to adjust the position of the optical elements on the cage plate.

[0049] In order to collect as much light field data as possible to improve the spatial resolution of light field tomography reconstruction, a high-resolution large-size image sensor is used. In order to increase the sampling angle range of flame radiation, the aperture of the main lens is increased, that is, the F number is reduced. A main lens with adjustable focal length and F number is selected to adapt to different measurement requirements and achieve F number matching. In order to match the F number of the main lens with the microlens, the F number of the microlens should be equal to or slightly larger than the F number of the main lens. In order to reduce the influence of the relay lens on the final imaging quality, a lens with adjustable F number and small aberration is selected to match the size of the microlens and image sensor.

[0050] Figure 2 is a schematic diagram of using a light field tomography system to image a flame to be measured;

[0051] The two sets of light field tomography components are arranged at an angle of 90 degrees at an appropriate distance from the measurement area A where the measured flame is located, so that the main lens focuses on the center of the measured flame and the image size is close to the image detector size. The two sets of light field tomography components are respectively equipped with a high-speed camera and an RGB camera.

[0052] Figure 3 It is a schematic diagram of the optical path of the ray tracing process of light field tomography;

[0053] In order to establish the relationship between the flame radiation intensity value and the grayscale value of the image captured by the image detector, flame radiation ray tracing is performed starting from the pixels in the image captured by the image detector to determine the trajectory of the flame radiation light in the flame. The direction of the flame radiation beam received by the pixel is determined by the microlens center (number D in the figure) and the pixel center (number C in the figure). The sampling light is traced by utilizing the conjugate relationship between the microlens array and the object plane 4 with respect to the main lens 3, the thin lens imaging model and the geometric similarity theorem. Finally, the relationship between the internal radiation intensity of the flame and the radiation intensity of the light received by the image detector is established based on the flame radiation transmission equation. The ray tracing method uses the optical center of the main lens 3 as the coordinate origin and the normal direction of the image detector as the z-axis direction to establish a Cartesian right-handed coordinate system. The distance between the object plane 4 and the main lens 3 is denoted as l OE , the distance between the microlens array 2 and the image detector 1 is denoted as l AB , the distance between the microlens array 2 and the main lens 3 is recorded as l OB , the focal length of the microlens is denoted as f MLA , the focal length of the main lens is denoted as f. The coordinates of each pixel on the image captured by the image detector 1 are denoted as (x i ,y i ), the corresponding microlens center coordinates are marked as (x m ,y m ), the intersection point of the pixel flame radiation and the main lens is recorded as (x j ,y j ), the optical center coordinate of the main lens is marked as (x O ,y O ).

[0054] According to the geometric similarity theorem, we can get:

[0055]

[0056] The intersection coordinates can be obtained from the pixel coordinates and the microlens center coordinates:

[0057]

[0058] The microlens array surface and the object plane are conjugate with respect to the main lens. From the thin lens imaging model and the geometric similarity theorem, we can get:

[0059]

[0060] The object point coordinates can be obtained from the microlens center coordinates and the main lens center coordinates:

[0061] DETAILED DESCRIPTION

[0062] Specific implementation method 1: Combination Figure 1 This embodiment describes a light field tomography system based on dual-camera fusion, wherein the light field tomography system comprises a first light field tomography component and a second light field tomography component, wherein:

[0063] The first optical field tomography assembly includes a main lens 3, a microlens array 2, a relay lens 6 and an image detector 1; and the main lens 3, the microlens array 2, the relay lens 6 and the image detector 1 share a common optical axis, that is, the center of the main lens 3, the center of the microlens array 2, the center of the relay lens 6 and the center of the image detector 1 are located on the same straight line;

[0064] The structure of the second light field tomography assembly is the same as that of the first light field tomography assembly, and both the first light field tomography assembly and the second light field tomography assembly are placed obliquely above the flame to be measured in the horizontal direction (that is, in the measuring state, the line connecting the centers of the components in the first light field tomography assembly is parallel to the horizontal direction, and the line connecting the centers of the components in the second light field tomography assembly is parallel to the horizontal direction).

[0065] The relay lens 6 can generate an upright and enlarged virtual image of a close-up object, that is, the actual measurement object, at a farther distance, and perform secondary imaging on the image formed by the microlens array 2, and finally form an image on the image detector 1. In order to increase the sampling angle range of the flame radiation light, the aperture of the main lens is increased, that is, the F number is reduced. A main lens with adjustable focal length and F number is selected to adapt to different measurement requirements and achieve F number matching. In order to match the F number of the main lens with the microlens, the F number of the microlens should be equal to or slightly greater than the F number of the main lens. In order to reduce the influence of the relay lens on the final imaging quality, a lens with adjustable F number and small aberration is selected to match the size of the microlens and the image detector.

[0066] Specific implementation method 2: This implementation method is different from specific implementation method 1 in that the image detector of the first light field tomography component adopts an RGB camera.

[0067] The other steps and parameters are the same as those in the first embodiment.

[0068] Specific implementation method three: This implementation method is different from specific implementation methods one or two in that the image detector of the second light field tomography component adopts a high-speed camera.

[0069] The other steps and parameters are the same as those in the first or second embodiment.

[0070] The relay lens is an optical system composed of two identical lenses. It can generate an upright and enlarged virtual image of a close-up object, i.e. the actual measurement object, at a farther distance, and perform secondary imaging on the image formed by the microlens, and finally form an image on the image detector.

[0071] Specific implementation method 4: This implementation method is different from any one of specific implementation methods 1 to 3 in that the relay lens 6 is composed of two identical lenses.

[0072] The other steps and parameters are the same as those in Specific Embodiments 1 to 3.

[0073] In this embodiment, the several tiny lenses constituting the microlens array 2 are convex lenses, and the size of the tiny lenses can be 10 mm×10 mm, but is not limited to 10 mm×10 mm, and the specific number can be selected according to actual conditions. Moreover, the array direction of the microlens array 2 is perpendicular to the line connecting the centers of each component. In addition, the lenses included in the main lens 3 and the relay lens 6 are thin lenses (specifically convex lenses).

[0074] Specific embodiment 5: This embodiment is different from any one of specific embodiments 1 to 4 in that the light field tomography system includes a plurality of cage plates 5 and cage rods 7;

[0075] The cage plate 5 is provided with through holes at the center and around the edges, and the main lens 3 and the relay lens 6 are installed on the cage plate 5 through the through holes at the center of the cage plate 5;

[0076] The cage rods 7 pass through the through holes around each cage plate 5 .

[0077] The other steps and parameters are the same as those in Specific Embodiments 1 to 4.

[0078] The optical field tomography system can be fixed by a fixing device mainly composed of a cage plate and a cage rod. The center of the cage plate has a through hole for installing optical elements, and the four corners of the cage plate have through holes for fixing and passing through the cage rods. The four sides of the cage plate 7 can be connected to other optical platforms by installing bolts and connecting rods and other fasteners to adjust the position of the imaging system. The cage plate 5 can slide freely along the cage rod 7 to adjust the position of the optical element on the cage plate 5 to adapt to measurement objects of different distances and sizes. The cage structure can ensure that the various components of the imaging system share the same optical axis, and each component can be independently adjusted or replaced to adapt to measurement objects of different distances and sizes.

[0079] Specific implementation method six: Combination Figure 2 The present embodiment describes a flame three-dimensional temperature field reconstruction method, the method specifically comprising the following steps:

[0080] Step 1: adjusting the main lens so that the first light field tomography component is focused on the flame to be measured, and using the RGB camera of the first light field tomography component to shoot the light field image of the flame to be measured; adjusting the main lens so that the second light field tomography component is focused on the flame to be measured, and using the high-speed camera of the second light field tomography component to shoot the light field image of the flame to be measured;

[0081] The image acquisition frame rate of traditional RGB cameras is low, usually 10 2 Hz, but the image resolution is higher, usually up to 10 8 ; High-speed cameras have low image resolution, usually 10 7 But the image acquisition frame rate is higher, usually up to 10 3 Hz;

[0082] Step 2: interpolate the measured flame light field image taken by the RGB camera to obtain the measured flame light field interpolation image corresponding to each image taken by the high-speed camera;

[0083] Step 3: The space area where the flame to be measured is located is taken as the measurement area, and the measurement area is divided into various micro-elements;

[0084] According to the relative position between the light field tomography system and the measurement area, flame radiation ray tracing is performed starting from each pixel point on the light field image taken by the high-speed camera and the interpolation image;

[0085] Step 4: Establish a reconstruction equation group based on the radiation transfer equation and the flame radiation ray tracing results, and obtain the blackbody spectral radiation intensity of each flame microelement at each moment by solving the reconstruction equation group (the specific moment is the moment when the RGB camera takes the image and the moment corresponding to the interpolated image);

[0086] Step 5: According to Planck's law, the blackbody spectral radiation intensity of each flame microelement is converted into the temperature of each microelement in the measurement area to complete the reconstruction of the three-dimensional temperature field of the flame at each moment.

[0087] Step 5 is based on the relationship between the blackbody spectrum radiation intensity of the microelement and the temperature of the microelement. The method for obtaining the relationship between the blackbody spectrum radiation intensity of the microelement and the temperature of the microelement is:

[0088] Step 51. Determine the temperature range of the blackbody furnace calibration according to the temperature range of the flame to be measured (the calibrated temperature range should be slightly larger than the temperature range of the flame to be measured);

[0089] Step 52: Set the temperature interval, determine each temperature measurement point of the black body furnace calibration according to the temperature interval, and the temperature difference between adjacent temperature measurement points is equal to the set temperature interval;

[0090] Step 53: At each temperature measurement point, the first light field tomography component is used to capture a blackbody furnace light field image, a polynomial fitting is performed on the grayscale of the blackbody furnace light field image and the temperature of the blackbody furnace, and a relationship between the grayscale of the blackbody furnace light field image and the temperature of the blackbody furnace is established;

[0091] Step 54: Start ray tracing from the pixel points in the collected image, reconstruct the equation group based on the ray tracing results, calculate the blackbody spectral radiation intensity, and establish the relationship between the grayscale of the blackbody furnace light field image and the blackbody spectral radiation intensity;

[0092] Step 55: Then obtain the relationship between the blackbody spectrum radiation intensity of the microelement and the temperature of the microelement:

[0093] The present invention proposes a method for reconstructing the three-dimensional temperature field of flames based on light field tomography imaging with dual cameras fusion. High-speed imaging technology is widely used in many fields such as industrial process control, industrial detection, combustion diagnosis, mechanical motion detection, and trajectory analysis. In the field of combustion diagnosis, high-speed three-dimensional imaging can capture the instantaneous complex flow structure of turbulence and achieve three-dimensional reconstruction of flow fields with high temporal resolution. By combining a high-image resolution RGB camera with a high-speed camera with a high image acquisition frame rate, high spatial resolution and high temporal resolution information are fused during temperature field reconstruction calculation, which can not only achieve high temporal resolution reconstruction of the temperature field, but also increase the number of viewing angles for light field image acquisition, thereby achieving high-precision reconstruction of temperature distribution.

[0094] Specific implementation example 7: This implementation example is different from specific implementation example 6 in that the interpolation of the measured flame light field image taken by the RGB camera adopts a cubic spline interpolation method.

[0095] The other steps and parameters are the same as those in the sixth embodiment.

[0096] Since the time interval between image acquisitions by the RGB camera is relatively large, two adjacent frames of images acquired by the RGB camera can be interpolated to obtain a high-resolution image of the image acquisition time point corresponding to the high-speed camera within the time period of the two frames. A new reconstruction equation is constructed using the grayscale value corresponding to the interpolated image and the sampling light, and the reconstruction equation group corresponding to the original high-speed camera image is added to improve the spatial resolution of the reconstruction. The specific description of the process is as follows:

[0097] The image series collected by the RGB camera is denoted as b r,1 ,b r,2 ,…,b r,M , the image corresponding to the acquisition time is t r,1 ,t r,2 ,…,t r,M The interpolation calculation process takes cubic spline interpolation as an example. Cubic spline interpolation is a commonly used numerical analysis method that can fit a smooth continuous function curve through a given set of scattered data. The basic idea is to use low-order polynomials to approximate the data in a small interval, and use the connection conditions of these polynomials to ensure the smoothness of the entire curve. It is widely used in data approximation and interpolation problems in engineering, science, mathematics and other fields. For N known data points (t r,i ,b r,i ), i = 1, 2, ..., M, where t r,1 <t r,2 <… <t r,M , the cubic spline interpolation method can find a set of functions b r =f(t r ), and satisfy f(t r ) is a cubic polynomial in each small interval, f(t r ) has the same first-order and second-order derivatives at the connection between two adjacent small intervals and f(t r ) passes through all given data points.

[0098] The cubic spline interpolation method first fits a cubic polynomial in each small interval, then sets the boundary conditions to obtain a system of equations, and solves the system of equations to find the coefficients of each small interval; finally, the interpolation points are given and the corresponding function values ​​are found using the coefficients.

[0099] The interpolation function between each cell is expressed as f i (t r ), the natural boundary conditions are:

[0100] f1″(t r1 )=0 (6)

[0101] f N ″ -1 (t rN)=0 (7)

[0102] From the natural boundary conditions, we can get γ1=γ N =0, the interpolation formula is:

[0103]

[0104] Let δ i =t r,i+1 -t r,i , Δ i =f(t r,i+1 )-f(t r,i ), the above conditions can be used to obtain the iterative calculation formula:

[0105]

[0106] The two adjacent frames of images captured by the RGB camera are interpolated to obtain the high-speed camera image acquisition time point t within the time period of the two frames. h,1 ,t h,2 ,…,t h,N The corresponding high-resolution image b r,1 , b r,2 , …, b r,N The grayscale values ​​corresponding to the interpolated image and the sampling light are used to construct a new reconstruction equation, which is added to the reconstruction equation group corresponding to the original high-speed camera image, thereby improving the spatial resolution of the reconstruction.

[0107] Specific implementation eight: This implementation is different from specific implementation six or seven in that the radiation transfer equation is specifically:

[0108] Since the exposure time of the image detector is extremely short, the change of the measured distribution over time can be ignored. Therefore, the radiation transfer process inside the flame can be described by the steady-state radiation transfer equation.

[0109] Considering the absorption, emission and scattering of the radiation transmission path medium:

[0110]

[0111] Among them, I bλ (s,s i ′) is the blackbody radiation intensity of the flame at position s and in the direction s′, I bλ (s) is the blackbody radiation intensity of the flame at position s, β λ (s) is the attenuation coefficient, unit is m -1 , κ λ (s) is the absorption coefficient, in m -1 , σ sλ (s) is the scattering coefficient, in m -1 ,Φ(s′i ,s′) is s′ i The scattering phase function of the flame radiation incident from the direction to the s′ direction, Ω i is the integration interval;

[0112]

[0113] Where C1 is the first radiation constant, C2 is the second radiation constant, λ is the wavelength of the flame radiation, T(s) is the temperature of the flame at position s, and e is the base of the natural logarithm;

[0114] Since the scattering of flame soot is much smaller than the absorption, the flame scattering is ignored. According to formula (8), the steady-state radiation transfer equation of the microelement is obtained as follows:

[0115]

[0116] Among them, I λ is the spectral radiation intensity of the flame radiation, I bλ is the blackbody spectral radiation intensity of the flame microelement, τ is the optical thickness of the flame radiation light passing through the flame microelement;

[0117] τ=κ·δ (15)

[0118] Among them, κ is the absorption coefficient of the flame microelement through which the flame radiation passes, and δ is the geometric length of the flame microelement through which the flame radiation passes.

[0119] The other steps and parameters are the same as those in the sixth or seventh embodiment.

[0120] The cone angle of the flame radiation imaging beam received by a single pixel in the image captured by the image detector is very small. The flame radiation light can be sampled through the center of the microlens and the center of the pixel to approximately represent the direction of the flame radiation imaging beam received by the pixel. The flame radiation intensity received by a single pixel is the intensity of the flame radiation sampling light, so the relationship between the flame radiation intensity value and the grayscale value of the image captured by the image detector is established through the flame radiation light. From the image detector to the inside of the flame, the flame radiation light is traced starting from the pixel to determine the trajectory of the flame radiation light in the flame. The specific ray tracing process is as follows: Figure 3 shown.

[0121] Specific implementation method 9: This implementation method is different from any one of specific implementation methods 6 to 8 in that the reconstruction equation group is established based on the radiation transfer equation and the flame radiation ray tracing results, specifically:

[0122] For any image in the RGB camera image and the interpolated image, the total number of flame micro-element bodies that the radiation light corresponding to any pixel point m on the image passes through is recorded as n. According to the order in which the radiation light passes through each flame micro-element body, each flame micro-element body is numbered 1, 2, ..., n in sequence. Then the radiation transmission equation of the flame radiation light corresponding to the pixel point m is:

[0123]

[0124] Among them, I λ (m) is the spectral radiation intensity of the flame radiation corresponding to pixel m, I bλ (i) is the blackbody spectral radiation intensity of the i-th flame element, i = 1, 2, ..., n, A m,i is the calculation coefficient;

[0125] Based on the flame radiation ray tracing results of step 3, the radiation transfer equations of all flame radiation rays are combined into a reconstruction equation group;

[0126]

[0127] Among them, b h,i is a vector composed of the spectral radiation intensity of the flame radiation light corresponding to each pixel in the image taken by the high-speed camera, b r,i is the vector of the spectral radiation intensity of the flame radiation light corresponding to each pixel in the interpolation image, x is the vector of the blackbody spectral radiation intensity of each flame microelement, A r is the coefficient matrix of RGB camera detection, A h is the coefficient matrix detected by the high-speed camera.

[0128] The other steps and parameters are the same as those in any one of the sixth to eighth embodiments.

[0129] After all the measured values ​​are obtained according to the above method, the equations can be solved. The tomography problem is usually ill-posed, that is, there may not be a unique solution that satisfies the equations. The underdetermined problem has infinite candidate solutions, each of which can satisfy the equations without residuals. On the other hand, the solution of this type of inverse problem is highly sensitive to small disturbances such as noise in the initial data. In order to improve the stability of solving the inverse problem of tomography, a regularization method can be used to provide additional information for the solution of the tomography problem. In actual measurement applications, tomography is affected by errors and noise. The initial reconstruction equations are shown in equation (14), where e represents the measurement noise and error, and the rank of the matrix A depends on the setting of the measurement optical path and the discretization method of the measurement area. One method to solve this linear equation system is the least squares method, that is, minimizing the squared L2 norm. The goal of this method is to minimize the difference between the estimated value Ax and the measured data b. However, the least squares inversion will amplify the model error in the matrix A and the noise in b into non-physical disturbances in x. Although the least squares method has greater fidelity to the measured data, the amplification of noise usually leads to unstable reconstruction results. Therefore, additional information must be added to the reconstruction process to obtain a true estimate of the solved field.

[0130] Ax+e=b (18)

[0131] One way to solve this problem is to introduce a minimization objective function consisting of two components, including the least squares term and the regularization penalty term. As shown in Equation (15), the first term represents the least squares criterion, and the second term is composed of the regularization parameter α and the prior information about the solution composition.

[0132]

[0133] The Tikhonov regularization method uses the L2 norm of the squared spatial gradient of the reconstructed distribution as a priori. In discrete form, the Tikhonov regularization method essentially uses a finite difference form to obtain the sum of the squared differences of the differences between adjacent values. This method usually uses a discrete Laplace operator Δ. The minimization objective function of Tikhonov regularization is shown in Equation (16):

[0134]

[0135] In the formula, 0 represents a column vector consisting of 0 elements equal to the number of micro-elements, 1 represents a column vector consisting of 1 elements equal to the number of micro-elements, and Δ is the Laplace matrix, which represents the relationship between adjacent voxels, i.e., the smoothness condition. The Tikhonov regularization coefficient λ can be determined by the L-curve method, by calculating the value of λ for different values. and The curve is obtained. A point that maximizes the curvature of the curve is selected as parameter λ, and the selected parameter λ meets two of the criteria.

[0136] Specific implementation method 10: This implementation method is different from any one of specific implementation methods 6 to 9 in that the calculation coefficient A m,i for:

[0137]

[0138] Among them, τ j It is the optical thickness of the jth flame element through which the flame radiation light corresponding to pixel point m passes.

[0139] The other steps and parameters are the same as those in Specific Embodiments 6 to 9.

[0140] The above calculation examples of the present invention are only used to explain the calculation model and calculation process of the present invention in detail, and are not intended to limit the implementation methods of the present invention. For ordinary technicians in the relevant field, other different forms of changes or modifications can be made based on the above description. It is impossible to list all the implementation methods here. All obvious changes or modifications derived from the technical solution of the present invention are still within the scope of protection of the present invention.

Claims

1. A light field tomography system based on dual-camera fusion, characterized in that: The light field tomography system comprises a first light field tomography component and a second light field tomography component, wherein: The first optical field tomography component comprises a main lens (3), a microlens array (2), a relay lens (6) and an image detector (1); the main lens (3), the microlens array (2), the relay lens (6) and the image detector (1) share a common optical axis, that is, the center of the main lens (3), the center of the microlens array (2), the center of the relay lens (6) and the center of the image detector (1) are located on the same straight line; The structure of the second light field tomography component is the same as that of the first light field tomography component, and both the first light field tomography component and the second light field tomography component are placed obliquely above the flame to be measured in the horizontal direction.

2. The light field tomography system based on dual-camera fusion according to claim 1, characterized in that: The image detector of the first light field tomography component is an RGB camera.

3. The light field tomography system based on dual-camera fusion according to claim 1, characterized in that: The image detector of the second light field tomography component is a high-speed camera.

4. The light field tomography system based on dual-camera fusion according to claim 1, characterized in that: The relay lens (6) is composed of two identical lenses.

5. The light field tomography system based on dual-camera fusion according to claim 1, characterized in that: The optical field tomography system comprises a plurality of cage plates (5) and cage rods (7); The cage plate (5) is provided with through holes at the center and around the periphery, and the main lens (3) and the relay lens (6) are mounted on the cage plate (5) through the through holes at the center of the cage plate (5); The cage rods (7) pass through the through holes around each cage plate (5).

6. A flame three-dimensional temperature field reconstruction method based on a dual-camera fusion optical field tomography system according to claim 1, characterized in that: The method specifically comprises the following steps: Step 1: adjusting the main lens so that the first light field tomography component is focused on the flame to be measured, and using the RGB camera of the first light field tomography component to shoot the light field image of the flame to be measured; adjusting the main lens so that the second light field tomography component is focused on the flame to be measured, and using the high-speed camera of the second light field tomography component to shoot the light field image of the flame to be measured; Step 2: interpolate the measured flame light field image taken by the RGB camera to obtain the measured flame light field interpolation image corresponding to each image taken by the high-speed camera; Step 3: The space area where the flame to be measured is located is taken as the measurement area, and the measurement area is divided into various micro-elements; According to the relative position between the light field tomography system and the measurement area, flame radiation ray tracing is performed starting from each pixel point on the light field image taken by the high-speed camera and the interpolation image; Step 4: Establish a reconstruction equation group based on the radiation transfer equation and the flame radiation ray tracing results, and obtain the blackbody spectral radiation intensity of each flame microelement at each moment by solving the reconstruction equation group; Step 5: Convert the blackbody spectral radiation intensity of each flame microelement into the temperature of each microelement in the measurement area to complete the reconstruction of the three-dimensional temperature field of the flame at each moment.

7. The flame three-dimensional temperature field reconstruction method based on the dual-camera fusion optical field tomography system according to claim 6 is characterized in that: The interpolation of the measured flame light field image taken by the RGB camera adopts a cubic spline interpolation method.

8. The flame three-dimensional temperature field reconstruction method based on the dual-camera fusion optical field tomography system according to claim 7 is characterized in that: The radiation transfer equation is specifically: Considering the absorption, emission and scattering of the radiation transmission path medium: Among them, I bλ (s,s i ′) is the blackbody radiation intensity of the flame at position s and in the direction s′, I bλ (s) is the blackbody radiation intensity of the flame at position s, β λ (s) is the attenuation coefficient, unit is m -1 , κ λ (s) is the absorption coefficient, in m -1 , σ sλ (s) is the scattering coefficient, in m -1 ,Φ(s i ′,s′) is s i The scattering phase function of the flame radiation incident from the ′ direction to the s′ direction, Ω i is the integration interval; Where C1 is the first radiation constant, C2 is the second radiation constant, λ is the wavelength of the flame radiation, T(s) is the temperature of the flame at position s, and e is the base of the natural logarithm; According to formula (8), the steady-state radiation transfer equation of the microelement is: Among them, I λ is the spectral radiation intensity of the flame radiation, I bλ is the blackbody spectral radiation intensity of the flame microelement, τ is the optical thickness of the flame radiation light passing through the flame microelement; τ=κ·δ (15) Among them, κ is the absorption coefficient of the flame microelement through which the flame radiation passes, and δ is the geometric length of the flame microelement through which the flame radiation passes.

9. The flame three-dimensional temperature field reconstruction method based on the dual-camera fusion optical field tomography system according to claim 8 is characterized in that: The reconstruction equations are established based on the radiation transfer equation and the flame radiation ray tracing results, specifically: For any image in the RGB camera image and the interpolated image, the total number of flame micro-elements that the radiation light corresponding to any pixel point m on the image passes through is recorded as n, and the radiation transmission equation of the flame radiation light corresponding to the pixel point m is: Among them, I λ (m) is the spectral radiation intensity of the flame radiation corresponding to pixel m, I bλ (i) is the blackbody spectral radiation intensity of the i-th flame element, i = 1, 2, ..., n, A m,i is the calculation coefficient; Based on the flame radiation ray tracing results of step 3, the radiation transfer equations of all flame radiation rays are combined into a reconstruction equation group; Among them, b h,i is a vector composed of the spectral radiation intensity of the flame radiation light corresponding to each pixel in the image taken by the high-speed camera, b r,i is the vector of the spectral radiation intensity of the flame radiation light corresponding to each pixel in the interpolation image, x is the vector of the blackbody spectral radiation intensity of each flame microelement, A r is the coefficient matrix of RGB camera detection, A h is the coefficient matrix detected by the high-speed camera.

10. The flame three-dimensional temperature field reconstruction method based on the dual-camera fusion optical field tomography system according to claim 9 is characterized in that: The calculation coefficient A m,i for: Among them, τ j It is the optical thickness of the jth flame element through which the flame radiation light corresponding to pixel point m passes.

Citation Information

Cited By

  • Improved cage type light field imaging system and debugging method

    CN121037672A

  • High-speed flame emission chromatography device based on reflector array and calibration and three-dimensional reconstruction method

    CN121962475A

  • A high-speed flame emission tomography device based on a mirror array and its calibration and three-dimensional reconstruction method

    CN121962475B