High-resolution three-dimensional flow field testing method and system based on dual-viewing angles of a single light field camera

Abstract

The present invention proposes a high-resolution three-dimensional flow field measurement method and system based on a single light field camera with dual perspectives. The method includes: using a light field camera to image the flow field to be measured in the front view direction and the top view direction reflected by a 45-degree mirror, thereby obtaining a time-series particle light field image of the flow field to be measured; reconstructing the light field image based on the light field principle to obtain three-dimensional particle images from two perspectives; superimposing the reconstructed front view image and top view image to obtain an accurate three-dimensional particle image; and performing cross-correlation calculation on the time-series three-dimensional particle image to obtain a three-dimensional flow field velocity field. The system utilizes the superposition of two perspectives to improve the axial resolution of light field imaging. The present invention can obtain an accurate three-dimensional velocity vector field of the flow field using a single light field camera. Compared with existing three-dimensional flow field measurement methods such as dual light field cameras, the present invention reduces the cost of the three-dimensional PIV system, simplifies system operation, and is suitable for high-precision measurement of three-dimensional flow fields in restricted optical access spaces.

Description

Technical Field

[0001] The present invention relates to a high-resolution three-dimensional flow field testing method and system based on a single light field camera with dual viewing angles. Background Art

[0002] The earliest research on particle image velocimetry (PIV) dates back to the late 1970s. After more than four decades of research and development, PIV has become a mature flow measurement technology, offering advantages such as non-contact, flow structure visualization, and full-field flow measurement. It is widely used in a wide range of research fields, including experimental fluid dynamics, aerodynamics, biomimetic fluid dynamics, intake mechanics, and combustion. To meet the demand for three-dimensional, three-component velocity field measurements in research, researchers have dedicated the past decade or so to developing various full-field three-dimensional flow measurement technologies.

[0003] Currently, there are several 3D flow field measurement technologies. Scanning PIV uses a 2D laser plane and a specialized mechanical device to scan the 3D flow field, achieving 3D flow field measurement. However, this method is limited by its mechanical structure and scanning time. Defocused PIV uses a specific aperture near the camera lens to eliminate blurring caused by particles not being in the focal plane. However, the density of tracer particles in defocused PIV is severely limited, and the use of an aperture significantly reduces the amount of light collected, making it incapable of imaging high-speed flow fields. Holographic PIV records the interference pattern (hologram) produced by a coherent light beam passing through the particle volume to restore the volume position of the tracer particles. Typically, a special holographic projection film is required to record the interference pattern of the tracer particles, a cumbersome and time-consuming process. In recent years, digital holographic PIV has made great progress, but it still suffers from low resolution and insufficient reconstruction of velocity vectors. Tomographic PIV uses multiple cameras to record tracer particle information from different angles and reconstruct the 3D velocity vector field through tomographic imaging. It has been widely used in the flow field, but it requires a complex and expensive camera array system and tedious system debugging. Synthetic aperture PIV is another multi-camera 3D PIV technique, but compared to tomographic PIV, it requires the use of more cameras (typically 8 to 15). Light field PIV, as an alternative to synthetic aperture PIV, replaces multiple cameras by placing a microlens array (MLA) in front of a single camera. Compared to other methods, light field PIV greatly simplifies the PIV system. However, the image reconstructed by a single light field camera has a problem of lower axial resolution than lateral resolution, making 3D high-resolution flow field measurement very difficult. Although dual light field PIV can effectively solve the problem of low axial image resolution, it increases the cost and complexity of the system and, more importantly, greatly limits its application in spaces with restricted optical access. Summary of the Invention

[0004] The present invention aims to provide a high-resolution, dual-view 3D flow field measurement method and system based on a single light-field camera. By adding a reflector above the imaging area, this system enables simultaneous 3D imaging of the flow field under test from two perspectives using a single light-field camera. By superimposing the dual-view images, high-spatial-resolution 3D imaging based on isotropic geometry is achieved. This system overcomes the poor axial resolution of single light-field cameras, enabling high-precision 3D flow field measurement with a single light-field camera. It also broadens the application of light-field PIV systems in spaces with limited optical access.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] The first aspect of the present application provides a three-dimensional flow field measurement method based on a single light field camera with dual viewing angles, comprising:

[0007] (1) Amplify the laser into a parallel volume beam to excite the tracer particles, and place a 45-degree reflector above the flow field to be measured;

[0008] (2) Use the light field camera to image the flow field in the normal direction and the downward direction reflected by the 45-degree mirror to obtain the time-series particle light field image of the flow field;

[0009] (3) Calibrate the three-dimensional position of the light field system and correct the perspective error of the overhead image;

[0010] (4) Based on the light field principle, the front view image and the top view image of the light field image are reconstructed respectively to obtain the three-dimensional front view image and the three-dimensional top view image of the reconstructed particle field;

[0011] (5) Superimpose the front view image and the top view image of the reconstructed particle field to obtain an accurate three-dimensional distribution image of the particles;

[0012] (6) Perform cross-correlation calculation on the reconstructed time-series three-dimensional particle images to obtain the three-dimensional velocity vector field of the flow field to be measured;

[0013] (7) Post-process the three-dimensional velocity vector field of the flow field, remove the erroneous velocity vectors in the velocity vector field, and insert the correct velocity vector based on the adjacent velocity vector field.

[0014] The second aspect of the present application provides a three-dimensional flow field testing method based on a single light field camera with dual viewing angles, comprising:

[0015] (1) Light field camera, consisting of a macro lens, a microlens array, a main lens, and an industrial camera;

[0016] (2) Volumetric illumination and multi-view imaging, including an optical system consisting of a reflector, a 532 nm laser source, and a lens combination;

[0017] (3) A data processor is used to reconstruct the three-dimensional flow field velocity vector field by obtaining the time-series particle light field image.

[0018] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:

[0019] This application enables high-precision measurement of 3D flow fields within restricted optical access spaces using a single light-field camera, addressing the issue of low axial resolution of the light field. By replacing multiple (light-field) camera systems or specialized optical components, a single light-field camera reduces the cost of 3D PIV systems, greatly simplifies system configuration and operation, and effectively facilitates high-precision measurement research of 3D flow fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 This is a schematic diagram of the process of the single light field camera dual-view three-dimensional flow field velocity measurement method in the present invention;

[0021] Figure 2 This is a schematic diagram of the principle of the dual-view three-dimensional flow field velocity measurement method using a single light field camera in the present invention;

[0022] Figure 3 This is a schematic diagram of the principle of the light field camera in the present invention and its differences from traditional cameras;

[0023] Figure 4 It is a schematic diagram of the light field three-dimensional position calibration method in the present invention. DETAILED DESCRIPTION

[0024] Please combine Figures 1 to 4 As shown, the present invention provides a three-dimensional flow field testing method based on a single light field camera with dual viewing angles, comprising the following steps:

[0025] Step 1: If Figure 2 As shown in (a), tracer particles are evenly spread in the flow field to be measured. A 532nm laser source is magnified into a parallel volume light beam through a concave lens and a convex lens to excite the tracer particles. A reflector is placed above the flow field to be measured at a 45-degree angle to the upper plane of the flow field.

[0026] Step 2: Use the light field camera to image the front view direction and the top view direction reflected by the 45-degree mirror of the flow field to be measured, obtain the time-series particle light field image of the flow field, and divide the light field image recorded in two perspectives into the front view image and the top view image.

[0027] Step 3: Calibrate the three-dimensional position of the light field system and correct the perspective error of the overhead image. Figure 3This embodies the basic characteristics of a light field camera. A light field camera uses the microlens plane (s, t) and the sensor plane (u, v) to construct a 4D light field (s, t, u, v), which reflects the information of the real physical coordinates (X, Y, Z). The calibration process is to construct the relationship between the 4D light field (s, t, u, v) and the real physical coordinates (X, Y, Z), such as Figure 4 As shown in (a) and 4(b), the imaging area of ​​the light field is calibrated using a calibration plate, using a direct light field calibration algorithm. A calibration matrix is ​​constructed by taking multiple points at 21 different depths:

[0028]

[0029] s=[s0(u0,v0) s1(u0,v0) s2(u0,v0) … s0(u1,v0) … s n (u m ,v m )] T

[0030] t=[t0(u0,v0) t1(u0,v0) t2(u0,v0) … t0(u1,v0) … t n (u m ,v m )] T

[0031] Where (X i ,Y i ,Z i ) is the real physical coordinate of the sample, (s i ,t i ) is the micro-lens coordinate behind the main lens, (u i ,v j ) are the pixel coordinates behind the microlens. i, j are integers starting from 0, n is the number of samples taken, and m is the number of pixels along one direction behind the microlens.

[0032] The least squares method is used to solve the calibration coefficient matrix:

[0033] a s =A\s

[0034] a t =A\t

[0035] That is, the functional relationship between the light field image (s, t, u, v) and the real physical coordinates (X, Y, Z) is obtained, and the axial position system error is as follows: Figure 4 As shown in (c), the calibration method can also correct the perspective error of the top-view image so that the top-view image and the front-view image have a one-to-one correspondence between the three-dimensional positions when they are superimposed.

[0036] Step 4: Utilize the refocusing property of the light field image to reconstruct the three-dimensional distribution of particles. Reconstruct the front view image and the top view image of the light field image respectively. The light field camera is equivalent to a compact camera array. According to the equation:

[0037] E(X,Y,Z)=∫∫L(u,v,P s (X,Y,Z,u,v),P t (X,Y,Z,u,v))dudv

[0038] By shifting and superimposing different viewing angles in different directions, the axial direction of the three-dimensional particle field can be reconstructed. i The function is obtained by the calibration coefficients in the third step.

[0039] During the superposition process, the light field image is separated from the particles by setting the background intensity threshold and the effective projection percentage threshold, which can effectively remove the influence of background noise and blurred particles. The method is as follows:

[0040]

[0041]

[0042] where n 有效 is the number of superimposed viewing angles greater than the background intensity threshold, n 总 is the total number of superimposed viewing angles, and V is the effective projection percentage.

[0043] Step 5: If Figure 2 As shown in (b), the particle field reconstructed from a single perspective has an obvious stretching effect along the axial direction. The top view image of the reconstructed particle field is rotated 90 degrees and superimposed with the front view image of the reconstructed particle field to finally obtain a high-precision three-dimensional particle voxel matrix;

[0044] Step 6: Perform cross-correlation calculations on the time-series particle 3D voxel matrix to obtain the 3D velocity vector field of the flow field. Divide the particle 3D voxel matrix into small, uniformly sized sampling volumes. By performing Fourier transform multiplication on the corresponding sampling volumes of two consecutive frames of voxel matrices, and then performing an inverse Fourier transform on the product in the frequency domain, the cross-correlation function of the sampling region is obtained, and the 3D velocity vector field of the entire flow field is further calculated.

[0045] Step 7: Post-process the three-dimensional velocity vector field of the flow field, remove the erroneous velocity vectors in the velocity vector field, and insert the correct velocity vector based on the adjacent velocity vectors.

[0046] The present application also provides a three-dimensional flow field testing system, comprising:

[0047] (1) Light field cameras, including macro lenses, microlens arrays, primary lenses, and industrial cameras;

[0048] (2) Volumetric illumination and multi-view imaging, including an optical system consisting of a reflector, a 532nm laser source, and a lens combination;

[0049] (3) Data processor, which reconstructs the three-dimensional flow field velocity vector field by obtaining the time-series particle light field image.

[0050] In summary, this application improves the axial resolution of a single light-field camera by superimposing dual-view light-field images, enabling a single light-field camera to perform high-precision measurements of three-dimensional flow fields within restricted optical access spaces. A single light-field camera replaces multiple (light-field) camera systems or specialized optical components. Compared to existing multi-(light-field) camera three-dimensional flow field testing methods, this reduces the cost of the three-dimensional PIV system and greatly simplifies the system's configuration and operation. It is particularly suitable for high-precision measurements of three-dimensional flow fields within restricted optical access spaces, and has promoted research on three-dimensional flow field measurements.

[0051] The above detailed description of the specific embodiments of the present invention is intended only as an example, and the present invention is not limited to the specific embodiments described above. It should be noted that those skilled in the art may make various improvements and modifications without departing from the principles of the present invention, and such improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A three-dimensional flow field testing method based on a single light field camera with dual viewing angles, characterized in that: The following steps are involved: (1) Amplify the laser into a parallel volume light beam to excite the tracer particles, and place a 45-degree reflector above the flow field to be measured; (2) Use the light field camera to simultaneously image the flow field in the normal direction and the downward direction reflected by the 45-degree mirror, and obtain the time-series particle light field image of the tracer particles in the flow field; (3) Calibrate the three-dimensional position of the light field system, calibrate the imaging area of ​​the light field with a calibration plate, use a direct light field calibration algorithm, and correct the perspective error of the overhead image; (4) Based on the light field principle, the front view image and the top view image of the light field image are reconstructed respectively to obtain the three-dimensional front view image and the three-dimensional top view image of the reconstructed particle field; (5) Superimpose the three-dimensional front view image and the three-dimensional top view image of the reconstructed particle field to obtain an accurate three-dimensional distribution image of the particles; (6) Perform cross-correlation calculation on the reconstructed three-dimensional particle distribution image to obtain the three-dimensional velocity vector field of the flow field to be measured; divide the three-dimensional particle voxel matrix into small sampling volumes of uniform size, and perform Fourier transform multiplication on the corresponding sampling volumes of two consecutive frames of voxel matrices, and then perform inverse Fourier transform on the product in the frequency domain to obtain the cross-correlation function of the sampling area; (7) Post-process the three-dimensional velocity vector field of the flow field, remove the erroneous velocity vectors in the velocity vector field, and insert the correct velocity vector based on the adjacent velocity vector field.

2. The three-dimensional flow field measurement method based on a single light field camera with dual viewing angles according to claim 1, characterized in that: In step (1), tracer particles are evenly spread in the flow field to be measured, and a 532 nm laser source is used to excite the tracer particles by magnifying the beam into a parallel volume light through a concave lens and a convex lens in sequence.

3. The three-dimensional flow field measurement method based on a single light field camera with dual viewing angles according to claim 1, characterized in that: In step (2), a reflector is placed above the flow field to be measured at an angle of 45 degrees to the top-view plane, and a single light field camera is used to simultaneously image the flow field in the front view direction and the top view direction to be measured, thereby obtaining a time-series particle light field image of the flow field, including an image in the front view direction and an image in the top view direction.

4. The three-dimensional flow field measurement method based on a single light field camera with dual viewing angles according to claim 1, characterized in that: In step (4), the particle field reconstruction refers to restoring the three-dimensional spatial distribution of the tracer particles by utilizing the refocusing property of the light field image.

5. The three-dimensional flow field measurement method based on a single light field camera with dual viewing angles according to claim 4, characterized in that: The particle field reconstruction includes: a light field filtering refocusing algorithm, a perspective error correction algorithm for overhead images, and a calibration algorithm for the spatial position of particle three-dimensional images.

6. The three-dimensional flow field measurement method based on a single light field camera with dual viewing angles according to claim 1, characterized in that: In step (5), the precise three-dimensional particle distribution image refers to the use of the superposition of dual-perspective particle three-dimensional images to eliminate the axial stretching effect of the single-perspective particle three-dimensional image, thereby achieving high-resolution reconstruction of the three-dimensional particle position and shape, thereby obtaining accurate three-dimensional flow field velocity.

7. The three-dimensional flow field measurement method based on a single light field camera with dual viewing angles according to claim 6, characterized in that: The precise three-dimensional distribution of particles includes: an algorithm for realizing superposition of three-dimensional images of particles from different viewing angles and an algorithm for reconstructing the quality assessment of the three-dimensional images of particles.

8. A three-dimensional flow field testing system for implementing the three-dimensional flow field testing method according to any one of claims 1 to 7, characterized in that: include: (1) Light field cameras, including macro lenses, microlens arrays, primary lenses, and industrial cameras; (2) Volumetric illumination and multi-view imaging, including an optical system consisting of a reflector, a 532nm laser source, and a lens combination; (3) Data processor, which reconstructs the three-dimensional flow field velocity vector field by obtaining the time-series particle light field image.

Images