Fluid measurement method, fluid measurement program, and fluid measurement system

Abstract

To provide a fluid measurement method, fluid measurement program and fluid measurement system which improve accuracy of flow rate calculation.SOLUTION: A fluid measurement method generates a color pattern using two or more of three light sources of a light source which emits an R component in RGB, a light source which emits a B component and a light source which emits a G component so as to be able to perform emission simultaneously, generates a color pattern image being a two-dimensional image whose color is changed along one axis according to the color pattern, projects the generated color pattern image toward fluid, acquires a photographed image obtained by imaging the fluid with the one axis as the depth direction, and measures a change in the fluid on the basis of the depth position according to the color of the acquired photographed image.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a fluid measurement method, a fluid measurement program, and a fluid measurement system. [Background technology]

[0002] In order to visualize fluid movement, methods for measuring the movement of fluids in three-dimensional space with high accuracy are being developed. For example, one method for visualizing fluid movement is called Particle Tracking Velocimetry (PTV), which involves photographing tracer particles that follow the fluid suspended in the fluid and calculating the particle velocity from the movement of the particles.

[0003] When using particle measurement methods, it is difficult to fully understand the phenomenon of particle movement with simple two-dimensional two-component (2D2C) measurements. When simply using photographed images, the position in the depth direction becomes unclear, making it impossible to measure particle movement in 3D3C.

[0004] Therefore, to obtain positional information in depth, a technology called rainbow PTV is available, in which a rainbow or digital rainbow image of white light is projected in the depth direction using a projector, and the hue of the scattered light is converted into depth to grasp the movement of particles.

[0005] To expand the dynamic range, a technology has been proposed that repeatedly projects a rainbow pattern two or four times in the depth direction to expand the dynamic range in the depth direction. With this technology, the projected image is periodically switched to determine which cycle of the rainbow it is. There is also a technology that combines the repetition of the rainbow pattern with defocusing to determine which cycle of the rainbow it is from the blur of the image, thereby expanding the dynamic range in the depth direction. Furthermore, a technology has been proposed that introduces color judgment using machine learning to improve the accuracy of the judgment.

[0006] Other 3D3CPTV technologies that do not use rainbow images include technology that uses cameras facing in two or more different directions to capture images and recognize depth positions, technology that identifies depth positions from image distortion caused by blur, and technology that recreates three-dimensional positions from holographic images. [Prior art documents] [Non-patent literature]

[0007] [Non-Patent Document 1] T. Watamura, Y. Tasaka, Y. Murai, LCD-projector-based 3D color PTV, Exp. Therm. Fluid Sci. 47 (2013) 68-80. [Retrieved October 19, 2021], Internet<URL:https: / / doi.org / 10.1016 / j.expthermflusci.2012.12.019> [Non-patent document 2] H.J. Park, D. Saito, Y. Tasaka, Y. Murai, Color-coded visualization of microbubble clouds interacting with eddies in a spatially developing turbulent boundary layer, Exp. Therm. Fluid Sci. 109 (2019) 109919. [Retrieved October 19, 2021], Internet<URL:https: / / doi.org / 10.1016 / j.expthermflusci.2019.109919> [Non-patent document 3] HJ Park, S. Yamagishi, S. Osuka, Y. Tasaka, Y. Murai, Development of multi-cycle rainbow particle tracking velocimetry improved by particle defocusing technique and an example of its application on a twisted Savonius turbine, Exp. Fluids. 62 (2021) 71. [Retrieved October 19, 2021], Internet<URL:https: / / doi.org / 10.1007 / s00348-021-03179-7> [Non-patent document 4] D. Noto, Y. Tasaka, Y. Murai, In situ color-to-depth calibration: toward practical three-dimensional color particle tracking velocimetry, Exp. Fluids. 62 (2021) 1-13. [Retrieved October 19, 2021], Internet<URL:https: / / doi.org / 10.1007 / s00348-021-03220-9> Summary of the Invention [Problem to be solved by the invention]

[0008] However, when projecting a digital rainbow image to determine the depth position, it is difficult to accurately identify the position due to the narrow dynamic range. Furthermore, the technology that repeats rainbow patterns, the technology that combines rainbow patterns with blur, and the technology that uses machine learning to determine color have insufficient dynamic ranges, making it difficult to accurately identify the depth position. Therefore, it is difficult to perform highly accurate flow velocity measurements using any of these technologies.

[0009] The disclosed technology has been made in view of the above, and aims to provide a fluid measurement method, a fluid measurement program, and a fluid measurement system that improve the accuracy of flow velocity measurement. [Means for solving the problem]

[0010] In one embodiment of the fluid measurement method, the fluid measurement program, and the fluid measurement system disclosed in the present application, two or more of three light sources, a light source that irradiates the R component of RGB, a light source that irradiates the B component, and a light source that irradiates the G component, are used in a manner that allows simultaneous irradiation. , the intensity ratio of the R component to the B component is continuously changed, and the G component is discretely changed. A color pattern is generated, a color pattern image is generated which is a two-dimensional image in which the color is changed along one axis according to the color pattern, the generated color pattern image is projected toward a fluid, the fluid is photographed with the one axis as the depth direction to obtain a photographed image, and changes in the fluid are measured based on the depth position according to the color of the obtained photographed image. [Effects of the Invention]

[0011] In one aspect, the present invention can improve the accuracy of flow velocity measurement. [Brief explanation of the drawings]

[0012] [Figure 1] FIG. 1 is a diagram showing an outline of a fluid measurement system. [Figure 2] FIG. 2 is a diagram for explaining a color pattern according to the embodiment. [Figure 3] FIG. 3 is a diagram for explaining a comparison between the color pattern according to the embodiment and a digital rainbow. [Figure 4] FIG. 4 is a diagram for explaining the components of a digital rainbow. [Figure 5] FIG. 5 shows the photographic results when the value of a single color is increased. [Figure 6] FIG. 6 is a diagram for explaining correction regarding the minimum value that can be output. [Figure 7] FIG. 7 is a diagram showing linearization correction of a gamma curve. [Figure 8] FIG. 8 is a flowchart of a fluid measurement process performed by the fluid measurement system according to the embodiment. [Figure 9] FIG. 9 is a diagram for explaining the effect when the algorithm for fluid measurement according to the embodiment is used. DETAILED DESCRIPTION OF THE INVENTION

[0013] Hereinafter, embodiments of the fluid measurement method, the fluid measurement program, and the fluid measurement system disclosed in the present application will be described in detail with reference to the drawings. Note that the fluid measurement method, the fluid measurement program, and the fluid measurement system disclosed in the present application are not limited to the following embodiments. [Example]

[0014] 1 is a diagram showing an overview of a fluid measurement system. The fluid measurement system 1 is a system for measuring the velocity of a fluid 200. The fluid measurement system 1 according to this embodiment includes an image generation device 10, an LCD (Liquid Crystal Display) projector 11, a color camera 12, and an image analysis device 13. Here, each color is represented by RGB (Red Green Blue).

[0015] The image generating device 10 generates a color pattern using two or more of three light sources, namely, a light source that emits the R component of RGB, a light source that emits the G component, and a light source that emits the B component, capable of simultaneous illumination. This color pattern will be described in detail later. Next, the image generating device 10 generates image data of a color pattern image 100 that reflects the generated color pattern projected from the LCD projector 11. The color pattern image 100 is a two-dimensional image represented by two axes, in which color changes along one axis and the same color continues along the other axis. In FIG. 1, the y-axis of the coordinates added to the fluid 200 is the direction in which color changes, and the z-axis is the direction in which the same color continues. Various information processing devices such as computers can be used as the image generating device 10.

[0016] The LCD projector 11 receives input of image data of the color pattern image 100 from the image generation device 10. Then, the LCD projector 11 projects the color pattern image 100 toward the fluid 200. For example, the color pattern image 100 is projected toward the fluid 200 contained in a water tank or the like. This LCD projector 11 is an example of a "projection device."

[0017] The color of light irradiated onto the particles P contained in the fluid 200 changes along the y-axis direction in Fig. 1. The particles P reflect the color of the light irradiated onto them. That is, particles P present at different positions along the y-axis reflect different colors.

[0018] The camera 12 is a high-resolution camera such as a normal color camera or a three-chip color camera. The camera 21 captures a planar image of the fluid 200, with the axis in the direction in which the color of the color pattern image 100 changes as the depth direction. That is, the camera 12 captures an image of the xz plane of the fluid 200 in the y-axis direction in FIG. 1. Hereinafter, the y-axis direction in FIG. 1 may be referred to as the depth direction.

[0019] The camera 12 captures the displacement of the particle P contained in the fluid 200 at predetermined intervals by photographing the fluid 200 at predetermined intervals. For example, suppose that the particle P in FIG. 1 moves along the dashed arrow, and the camera 12 photographs the particle P at one-second intervals. In this case, the particle P is at the first point along the arrow in FIG. 1 at t-1 seconds, at the intermediate point at t seconds, and at the final point at t+1 seconds. In this way, the movement of the particle P can be grasped from the images photographed by the camera 12. The camera 12 is an example of an "imaging device."

[0020] However, the image captured by the camera 12 is a planar image. Therefore, while the position of the particle P in the captured image can be determined in the x and z directions, its position in the depth direction (y direction) cannot be determined. Here, due to differences in the reflected light of particles P at different depth positions, it is possible to determine the depth position based on the color reflected by each particle P in the captured image. However, in order to accurately determine the depth position from the color, it is preferable to widen the dynamic range obtained from the color pattern in the color pattern image 100. Therefore, in the fluid measurement system 1 according to this embodiment, the dynamic range in the depth direction of the captured image is widened using a color pattern as described below. The color patterns used in the fluid measurement system 1 according to this embodiment are described in detail below.

[0021] <Color pattern> Fig. 2 is a diagram for explaining a color pattern according to an embodiment. The vertical axis on the left side of the paper in Fig. 2 represents the value of the B component normalized by the sum of the R and B components, and the vertical axis on the right side of the paper represents the value of the G component normalized by the sum of the R and B components. The horizontal axis in Fig. 2 represents the color pattern.

[0022] In the color pattern according to this embodiment, the ratio of the B component value to the total value of the R and B components is continuously changed. As a result, the ratio of the B component to the total value of the R and B components changes as shown in graph 101 in FIG. 2. Here, the B and G components are represented as normalized values, B / (B+R) and G / (B+R). Graph 101 is a graph in which the increase and decrease of B / (B+R) are alternately repeated eight times.

[0023] Furthermore, in the color pattern according to this embodiment, a constant G / (B+R) is set for each interval where B / (B+R) increases or decreases. That is, G / (B+R) takes eight discrete values ​​in accordance with the increase or decrease of B / (B+R), as shown in graph 102 in FIG. 2.

[0024] The color change due to the changes in the R, G, and B components is represented by a color pattern 103. In this color pattern 103, simultaneous changes in two or more colors of the R, G, and B components are recognized, and the hue, lightness, and saturation change.

[0025] As described above, the color pattern according to this embodiment is generated by continuously changing the intensity ratio between the R component and the B component and discretely changing the intensity ratio between the R component, the B component, and the G component. More specifically, the color pattern according to this embodiment is generated by repeating the same continuous change by alternately increasing and decreasing the intensity of the B component, and changing the intensity of the G component with each increase and decrease in the intensity of the B component.

[0026] FIG. 3 is a diagram illustrating a comparison between a color pattern according to an embodiment and a digital rainbow. FIG. 3 shows a three-dimensional space representing colors, with the horizontal axis representing the R component, the axis toward the back representing the B component, and the vertical axis representing the G component. FIG. 4 is a diagram illustrating the components of a digital rainbow. The vertical axis of graph 131 in FIG. 4 represents the RGB components of each color, and the horizontal axis represents the color. There is no definition of a digitally expressed rainbow, but here we will consider a color pattern arranged along the hue angle. Therefore, these RGB components are just an example.

[0027] 3, the grayscale is represented by a straight line connecting the point where the normalized R, G, and B components are all 0 to the point where they are all 1. For example, in the case of an 8-bit representation, the R, G, and B components are all normalized from 255 to 1.

[0028] The color pattern according to this embodiment is represented by a plane 110 in Fig. 3. More specifically, the color changes as indicated by arrows 111 in the plane 110 in response to an increase or decrease in the B component in the graph 101 in Fig. 2 and a corresponding increase in the G component. That is, the color pattern 103 in Fig. 2 corresponds to a line of eight arrows 111 arranged in the same direction from the bottom up. Here, the color pattern 103 shown in Fig. 2 is an incomplete representation because it is in black and white, but in reality it is a color pattern in which the color changes continuously.

[0029] 4, a color pattern 132 representing a digital rainbow is generated by changing the intensity ratio of any one of the R, G, and B components as shown in graph 131. The color pattern 132 corresponds to a change in hue angle.

[0030] In this way, the color pattern 132 representing the digital rainbow is expressed as a line in the three-dimensional space of Fig. 3, whereas the color pattern 103 used in this embodiment is expressed as a plane in the three-dimensional space of Fig. 3. Therefore, the color pattern 103 according to this embodiment can express a wider range of colors than the color pattern 132 representing the digital rainbow.

[0031] Therefore, by projecting a color pattern image 100 using color pattern 103 onto fluid 200, the range of colors that express the depth direction can be made wider than with color pattern 132 representing a digital rainbow, making it possible to expand the dynamic range in the depth direction.

[0032] Here, by increasing the number of repetitions of increase and decrease of the B component in Figure 2 and correspondingly increasing the number of stages of change of the G component, the number of arrows on plane 110 in Figure 3 will increase, making it possible to further expand the range of colors that can be expressed. However, if the number of stages of change of the G component is increased, adjacent G components will be closer to each other, which may reduce the accuracy of color identification and lead to incorrect judgment. For this reason, it is preferable that the number of increases and decreases of the B component and the number of stages of the G component be around eight, as shown in Figure 2.

[0033] In this embodiment, the R and B components are changed in a correlated manner, while the G component is changed independently. This is because by separating the G component, which has an intermediate wavelength, it is possible to suppress color interference when irradiated by the LCD projector 11 or captured by the color camera 12.

[0034] Furthermore, when photographing with the color camera 12, the light reflected from the particles P is photographed. The intensity of each color in the light reflected from the particles P is a value that is easily affected by the environment. For example, in the case of a digital rainbow, the accuracy of the analysis is improved by using hues that are relatively less affected by the environment, etc., in order to suppress intensity fluctuations. In contrast, in this embodiment, in order to expand the dynamic range, a color pattern was created that allows colors of the same hue but different saturations to be recognized as different colors.

[0035] Returning to FIG. 1, the explanation will be continued. The image analysis device 13 acquires the captured image taken by the color camera 12. Thereafter, the image analysis device 13 performs image analysis on the acquired image, and identifies the position of the particle P in the fluid 200 from the position of the particle P in the captured image and the color of the light reflected by the particle P. The image analysis device 13 then calculates the velocity of the fluid 200 from the movement state of the particle P. Various information processing devices such as computers can be used as the image analysis device 13. It is also possible to realize the image generation device 10 and the image analysis device 13 by the same computer.

[0036] Here, in the fluid measurement system 1 according to this embodiment, correction may be made to the output characteristics of the LCD projector 11. The correction to the output characteristics of the LCD projector 11 will be described below.

[0037] FIG. 5 shows the results of imaging when the value of a single color is increased. The horizontal axis of FIG. 5 represents the input value of the G component, and the vertical axis represents the measured value of the imaging result. Both the input and measured values ​​in FIG. 5 are normalized values. Graph 141 in FIG. 5 shows the relationship between the input value and the measured value of the G component. Graph 142 shows the relationship between the input value and the measured value of the R and B components. As shown in graph 141, the output characteristics of the LCD projector 11 are nonlinear with respect to the measured value of the G component input value. Furthermore, due to light leakage from the liquid crystal of the LCD projector 11, the minimum output value of the G component does not become 0, as shown by point 143. Furthermore, although the G component is increased, the R and B components also increase due to crosstalk, as shown in graph 142. Here, the increases in the R and B components caused by crosstalk are approximately the same, so graph 142 shows the states of both the R and B components.

[0038] FIG. 6 is a diagram illustrating correction related to the minimum possible output value. The horizontal axis of FIG. 6 represents the input value of the G component, and the vertical axis represents the measured value of the shooting result. The minimum value of the G component in graph 141 in FIG. 5 is enlarged as shown in FIG. 6. Here, in graph 141, when the input value of the G component is 0, the measured value is greater than 0.05 and less than 0.1, as indicated by lines 151 and 152. Therefore, a correction is applied to the output characteristics of LCD projector 11 to discard information up to the G component output value of approximately 0.1. Similarly, a correction is applied to the output characteristics of LCD projector 11 to discard the portion from 1 at the same rate as the input value truncated from 0.

[0039] FIG. 7 is a diagram showing linearization correction of a gamma curve. The horizontal axis of FIG. 7 represents the input value of the G component, and the vertical axis represents the measured value of the shooting result. By truncating the lower end of the input value corresponding to the portion where the measured value is approximately 0.1, the output of LCD projector 11 is truncated by 10% from the lower limit of the input value below line 161 in FIG. 7. Similarly, the output of LCD projector 11 is truncated by 10% from the upper limit of the input value above line 162. This corrects the minimum value that can be output for the output characteristics of LCD projector 11.

[0040] Furthermore, by applying a linearization correction to the G component gamma curve before the nonlinear correction shown in graph 141 to the output characteristics of LCD projector 11, a correction is applied to linearize the change in the G component as shown in graph 163. Furthermore, crosstalk between the R and B components is also corrected from the output characteristics of LCD projector 11. Existing technology can be used to correct this crosstalk. Additionally, when photographing light emitted from LCD projector 11, errors occur due to the sensitivity of color camera 12, but this sensitivity is automatically corrected by color camera 12, so no correction to the sensitivity is applied.

[0041] 8 is a flowchart of the fluid measurement process by the fluid measurement system according to the embodiment. Next, the flow of the fluid measurement process by the fluid measurement system 1 according to the embodiment will be described with reference to FIG.

[0042] The image generating device 10 generates a color pattern that allows simultaneous use of two or more of three-channel light sources that irradiate the R, G, and B components of RGB (step S1).

[0043] Various corrections are made to the output characteristics of the LCD projector 11, such as correction for the minimum value that can be output, correction for linearizing the gamma curve, and correction for removing crosstalk (step S2).

[0044] Next, the image generating device 10 generates image data of a color pattern image 100 that displays the generated color pattern, and inputs the image data to the LCD projector 11 (step S3).

[0045] The LCD projector 11 uses the input image data to project the color pattern image 100 onto the fluid 200 (step S4).

[0046] The color camera 12 captures the reflected light from the particles P of the fluid 200 onto which the color pattern image 100 is projected (step S5). Thereafter, the color camera 12 outputs the captured image to the image analysis device 13.

[0047] Thereafter, the image analyzer 13 converts the color of each captured particle P into depth, performs image analysis, calculates the flow velocity of the fluid 200, and performs fluid measurement (step S6).

[0048] FIG. 9 is a diagram for explaining the effect of using the fluid measurement algorithm according to the embodiment. Here, the effect of using the fluid measurement algorithm according to the embodiment will be explained with reference to FIG. 9. Color pattern 201 represents a digital rainbow color pattern. Color pattern 202 represents a color pattern when the G component has four levels. Color pattern 203 represents a color pattern when the G component has eight levels. In other words, color pattern 203 is the same as color pattern 103 in FIG. 2. Color patterns 201 to 203 shown in FIG. 9 are incomplete representations because they are black and white, but in reality they are color patterns in which the colors change continuously. Here, a test was conducted in which color patterns 201 to 203 were each projected onto a whiteboard, and the position was determined by the color to measure the dynamic range.

[0049] Graph 210 shows the determination results when color pattern 201 is projected. Graph 220 shows the determination results when color pattern 202 is projected. Graph 230 shows the determination results when color pattern 203 is projected. In each of graphs 210, 220, and 230, the horizontal axis represents the actual position in the captured image, and the vertical axis represents the position based on the determination result.

[0050] Here, the actual resolution was defined as the 95% confidence interval of the value obtained by subtracting the actual position in the captured image from the position determined by the judgment result, and the dynamic range was defined as the value obtained by dividing the number of pixels in the projected pattern by the actual resolution.

[0051] When color pattern 201 is used, the effective resolution in graph 210 is 60, and the dynamic range is 23. In contrast, when color pattern 202 is used, the effective resolution in graph 220 is 31, and the dynamic range is 44. Furthermore, when color pattern 203 is used, the effective resolution in graph 230 is 17, and the dynamic range is 79. Therefore, compared to when a digital rainbow color pattern is used, the dynamic range is wider even when a color pattern with four levels of the G component is used, and the dynamic range is significantly improved when a color pattern with eight levels of the G component is used.

[0052] As described above, in the fluid measurement algorithm according to this embodiment, a color pattern image, which allows simultaneous use of two or more of three-channel light sources that irradiate R, G, and B components, is projected onto the fluid by a projector with corrected output characteristics. Then, an image is taken with a camera, with the direction in which the color of the color pattern changes being the depth direction, and the color of each particle is converted into depth, followed by image analysis to measure the fluid flow velocity and other fluid measurements.

[0053] This allows for a wider dynamic range in the depth direction than when using a color pattern representing a digital rainbow, which allows for more accurate capture of particle movement in the depth direction and improved accuracy in fluid measurement. [Explanation of symbols]

[0054] 1. Fluid measurement system 10 Image generation device 11 LCD projector 12 color cameras 13 Image analysis equipment 100 color pattern images 200 fluid P particles

Claims

1. two or more of three light sources, a light source for irradiating an R component of RGB, a light source for irradiating a B component, and a light source for irradiating a G component, are used in a manner capable of simultaneous irradiation, and the intensity ratio between the R component and the B component is continuously changed and the G component is discretely changed to generate a color pattern; generating a color pattern image, which is a two-dimensional image in which the color is changed along one axis according to the color pattern; projecting the generated color pattern image toward a fluid; The fluid is photographed with the one axis as a depth direction to obtain a photographed image; A change in the fluid is measured based on a depth position corresponding to the color of the acquired photographed image. A fluid measurement method comprising:

2. The fluid measurement method according to claim 1, characterized in that the color pattern is a color pattern that has a wider dynamic range than a digital rainbow, generated by changing the color by simultaneously changing multiple components.

3. The fluid measurement method according to claim 1, characterized in that the color pattern is generated by repeating the same continuous change by alternately increasing and decreasing the intensity ratio of the B component, and by changing the intensity ratio of the G component with each increase and decrease in the intensity ratio of the B component.

4. 4. The fluid measurement method according to claim 1, wherein the color pattern image is projected after correction of the minimum value that can be output, correction of crosstalk, and linearization correction of a gamma curve.

5. two or more of three light sources, a light source for irradiating an R component of RGB, a light source for irradiating a B component, and a light source for irradiating a G component, are used in a manner capable of simultaneous irradiation, and the intensity ratio between the R component and the B component is continuously changed and the G component is discretely changed to generate a color pattern; generating a color pattern image, which is a two-dimensional image in which the color is changed along one axis according to the color pattern; projecting the generated color pattern image toward a fluid; The fluid is photographed with the one axis as a depth direction to obtain a photographed image; A change in the fluid is measured based on a depth position corresponding to the color of the acquired photographed image. A fluid measurement program that causes a computer to execute processing.

6. an image generating device that uses two or more of three light sources, a light source that irradiates an R component of RGB, a light source that irradiates a B component, and a light source that irradiates a G component, in a manner that allows simultaneous irradiation, continuously changes the intensity ratio between the R component and the B component, and discretely changes the G component to generate a color pattern, and generates a color pattern image, which is a two-dimensional image in which the color is changed along one axis according to the color pattern; a projection device that projects the color pattern image toward the fluid; an imaging device that captures an image of the fluid with the one axis as a depth direction; an image analysis device that measures a change in the fluid based on a depth position corresponding to the color of the acquired photographed image; A fluid measurement system comprising:

Images