JP2011017603A - Method for measuring particle image flow velocity - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately determine an erroneous vector using a peak value of a cross-correlation value of luminance patterns in two images in a particle image flow velocity measuring method.
A particle image flow velocity measuring apparatus supplies a tracer particle from a tracer particle supply unit to a flow field around an object, irradiates the tracer particle with a time difference, and irradiates a laser beam twice to capture reflected light. The velocity vector of the flow field is obtained based on the two images of the obtained tracer particles. The two images are each divided into a plurality of inspection regions, and the first peak value fp and the second peak value sp of the cross-correlation values of the luminance patterns of the tracer particles in each inspection region of the two images are compared, and the first image is compared. Since it is determined that the reliability of the velocity vector is high when the peak value fp / the second peak value sp ≧ 1.2 is established, it is possible to reliably determine the erroneous vector and increase the measurement accuracy of the flow state.
[Selection] Figure 5

Description

  The present invention is obtained by supplying tracer particles from a tracer particle supply unit to a flow field around an object, irradiating the tracer particles with a time difference, irradiating laser light twice, and imaging reflected light with an imaging unit. The present invention relates to a particle image velocimetry method for obtaining a flow field velocity vector based on two images of tracer particles.

  The flow field of the fluid containing the tracer particles is irradiated with a two-dimensional laser sheet at two times having a minute time difference, and images of the tracer particles in the laser sheet at each time are taken with a camera. A particle image velocity measurement method (PIV: Particle Image Velocimetry) that measures the state of a flow field based on the image of the above is known.

  In such a particle image velocimetry apparatus, it is desirable to image only the reflected light from the tracer particles contained in the fluid, but in reality, the reflected light from the object placed in the fluid is attached to the observation window of the wind tunnel. Since it is inevitable that noise components such as reflected light from the dust and halation are captured together with the reflected light from the tracer particles, an incorrect velocity vector (false vector) is used when calculating the fluid velocity vector from the camera image. May be calculated.

  In order to minimize the calculation of such an error vector, the time average luminance value of each pixel of the image or the minimum luminance value of a plurality of images taken with a time difference is removed as noise from the original image, so that the tracer particle A technique for obtaining an image in which only reflected light is left is known from Patent Document 1 below.

  Also, by filtering the spatial frequency component from the gray image at two consecutive times of the camera with a high-pass filter and a low-pass filter, the maximum moving distance of the two-time gray image can be obtained with high accuracy, leaving only the tracer particle image. It is known from Patent Document 2 below.

JP 2001-74415 A JP 2008-140103 A

  However, although what is described in Patent Documents 1 and 2 can be expected to have a certain degree of effect in eliminating erroneous vectors, there is still room for further improving the accuracy of eliminating erroneous vectors.

  The present invention has been made in view of the above-described circumstances, and an object of the particle image flow velocity measurement method is to accurately determine an erroneous vector using a peak value of a correlation degree of luminance patterns in two images.

  In order to achieve the above object, according to the first aspect of the present invention, the tracer particles are supplied from the tracer particle supply means to the flow field around the object, and the laser beam is emitted with a time difference between the tracer particles. In the particle image flow velocity measurement method for obtaining the velocity vector of the flow field on the basis of the two images of the tracer particles obtained by irradiating the reflected light with the imaging means, the two images are respectively applied to a plurality of inspection regions. The reliability of the velocity vector is determined by comparing the first peak value and the second peak value of the cross-correlation values of the luminance patterns of the tracer particles in each inspection region of the two images. A particle image flow velocity measuring method is proposed.

  According to the invention described in claim 2, in addition to the configuration of claim 1, when the first peak value / the second peak value ≧ 1.2 is established, the reliability of the velocity vector is high. A particle image flow velocity measuring method characterized by determining is proposed.

  The first and second first CCD cameras 32A and 32B in the embodiment correspond to the image pickup means of the present invention.

  According to the configuration of the first aspect, the tracer particles are supplied from the tracer particle supply means to the flow field around the object, the laser light is irradiated twice with a time difference between the tracer particles, and the reflected light is imaged by the imaging means. Then, the velocity vector of the flow field is obtained based on the two images of the obtained tracer particles. Each of the two images is divided into a plurality of inspection regions, and the velocity vector is compared by comparing the first peak value and the second peak value of the cross-correlation values of the luminance patterns of the tracer particles in each inspection region of the two images. Since the reliability is determined, it is possible to reliably eliminate erroneous vectors and increase the flow state measurement accuracy.

  According to the second aspect of the present invention, when the first peak value / second peak value ≧ 1.2 is established, it is determined that the reliability of the speed vector is high. The measurement accuracy of the flow state can be increased.

The whole top view of the particle image flow velocity measuring apparatus provided in the wind tunnel. FIG. 2 is an enlarged view of part 2 of FIG. 1. FIG. 3 is an enlarged sectional view taken along line 3-3 in FIG. 2. FIG. 4 is a sectional view taken along line 4-4 of FIG. Explanatory drawing of a movement of the Z-axis direction of a laser sheet and a CCD camera. Explanatory drawing of the process which calculates a peak ratio from the image of 1st, 2nd time. The figure which shows the relationship between the ratio of a 1st, 2nd peak, and the number of erroneous vectors.

  Hereinafter, embodiments of the present invention will be described with reference to FIGS.

  As shown in FIG. 1, an object 12 such as a model of an automobile body is placed inside a wind tunnel 11 to which a uniform flow of air having a predetermined flow velocity is supplied, and tracer particles installed upstream thereof. Fine oil droplets (tracer particles) having a diameter of several μm are supplied from the supply means 13 into the uniform flow. The uniform flow changes the direction of the flow along the surface of the object 12 and forms a three-dimensional velocity field. A transparent observation window 14 is provided on at least a part of the wall surface of the wind tunnel 11, and a particle image flow velocity measuring device 15 is disposed at a position facing the object 12 with the observation window 14 interposed therebetween.

  As shown in FIGS. 2 to 4, the particle image flow velocity measuring device 15 includes a main guide rail 21 arranged in parallel with the axis of the wind tunnel 11, a sub guide rail 22 arranged in parallel with the main guide rail 21, A main traverser 23 supported movably on the guide rail 21, a pair of first and second sub traversers 24A and 24B supported movably on the sub guide rail 22, and the main traverser 23 and the first sub traverser 24A are connected. The first connecting rod 25A and the second connecting rod 25B for connecting the main traverser 23 and the second sub-traverser 24B are provided, and the first and second connecting rods 25A, 25B have the same length. .

  An endless timing belt 28 is wound around a driving sprocket 26 and a driven sprocket 27 provided at both ends of the main guide rail 21, and one chord portion thereof is one through hole 23 a of the main traverser 23 (see FIG. 4). The other string portion passes through the other through-hole 23b (see FIG. 4) of the main traverser 23 in a non-engagement manner. When the drive sprocket 26 is driven by a motor 29 (see FIGS. 3 and 4), the timing belt 28 wound around the driven sprocket 27 rotates, and the main traverser 23 moves along the main guide rail 21. When the main traverser 23 moves, the first and second sub-traversers 24A and 24B connected thereto via the first and second connecting rods 25A and 25B move along the sub-guide rail 22.

  A laser head 30 is provided on an extension line of the main guide rail 21, and the laser head 30 irradiates the laser beam Lb along the main guide rail 21. Laser sheet irradiation means 31 comprising a mirror and a cylindrical lens is provided on the upper surface of the main traverser 23, and the laser sheet irradiation means 31 converts the laser beam Lb into a laser sheet Ls that irradiates a plane orthogonal to the laser beam Lb. The axes of the laser beam Lb and the wind tunnel 11 extend in the Z-axis direction, and the laser sheet Ls extends in the X-axis and Y-axis directions orthogonal to the Z-axis direction.

  First and second CCD cameras 32A and 32B are provided on the upper surfaces of the first and second sub-traversers 24A and 24B, respectively, and the first and second CCD cameras 32A and 32B are mirror-symmetrical with respect to the laser sheet Ls. Arranged and directed to one point on the laser sheet Ls. The first and second CCD cameras 32A and 32B image a predetermined area of the laser sheet Ls in the vicinity of the object 12, and the first and second CCD cameras 32A and 32B have a focus on the entire imaging area. A Shine Frog adapter is provided. The main traverser 23 on which the laser sheet irradiation means 31 is mounted and the first and second sub-traversers 24A and 24B on which the first and second CCD cameras 32A and 32B are mounted move in the Z-axis direction while maintaining a fixed positional relationship. Therefore, the laser sheet Ls and the first and second CCD cameras 32A and 32B move in the Z-axis direction while maintaining a certain positional relationship.

  The main traverser 23 and the first and second sub-traversers 24A and 24B move intermittently by a predetermined distance in the Z-axis direction, and the laser sheet Ls is irradiated twice at a short time interval Δt each time it stops. In the embodiment, a distance of 1 m in the Z-axis direction is moved intermittently at intervals of 10 cm (see FIG. 5). The time interval Δt is set to be smaller as the flow velocity of the uniform flow in the wind tunnel 11 is higher, and is adjusted so that the moving distance of the tracer particles in the Z-axis direction therebetween becomes a size suitable for measurement. Further, the movement distance of the tracer particles in the Z-axis direction at the time interval Δt is set so as not to exceed the thickness of the laser sheet Ls in the Z-axis direction.

  The uniform flow flowing in the wind tunnel 11 changes its direction around the object 12 to become a three-dimensional flow, and the tracer particles contained therein also move along the air streamline. The first and second CCD cameras 32A and 32B image the tracer particles irradiated on the laser sheet Ls in synchronization with the first irradiation at the time t1, so that 2 of the tracer particles distributed in the irradiation surface from two directions. Get the images. Similarly, the first and second CCD cameras 32A and 32B take images of the tracer particles irradiated on the laser sheet Ls in synchronization with the second irradiation at time t1 ′ = t1 + Δt, so that the tracer distributed in the irradiation surface is obtained. Two images from two directions of particles are acquired.

  In the present embodiment, among various methods of PIV, an “image correlation method” is employed in which the luminance patterns of two images acquired at two times t1 and t1 ′ are compared to obtain the movement vector of the tracer particle group.

  FIG. 6 shows, for example, two images acquired at two times by the first CCD camera 32A. Since the axis of the first CCD camera 32A is inclined with respect to the irradiation surface of the laser sheet Ls, the xy plane of the image is inclined with respect to the XY coordinates of the irradiation surface of the laser sheet Ls. The square frame is one of inspection areas obtained by dividing the image of the first CCD camera 32A into a grid pattern, and the movement vector in the xy plane of the tracer particle group is calculated in each inspection area. That is, the cross-correlation value C indicates which position in the predetermined inspection area of the second image acquired at time t1 ′ the luminance pattern of the tracer particle group in the predetermined inspection area of the first image acquired at time t1. A detection is made using (Δx, Δy), and the movement vector divided by time Δt is set as a two-component velocity vector in the inspection region. Hereinafter, the process of calculating the two-component velocity vector will be described.

  First, the initial values of the first peak value fp and the second peak value sp are both set to zero.

fp ← 0
sp ← 0
Subsequently, a cross-correlation value C (Δx, Δy) defined by the following equation is calculated.

  Here, f (x, y) is a luminance function obtained from the luminance pattern of the first image, and g (x, y) is a luminance function obtained from the luminance pattern of the second image. Therefore, g (x + Δx, y + Δy) corresponds to the luminance distribution of g (x, y) moved by −Δx in the x-axis direction and by −Δy in the y-axis direction. Therefore, Δx and Δy correspond to the movement amount of the luminance pattern in the time interval Δt.

  Each inspection region is a set of n square pixels each having a length of p in the x-axis direction and n in the y-axis direction, sweeping Δx from p to np, and Δy from p to np The cross-correlation value C (Δx, Δy) is calculated while sweeping up to. Each time the cross-correlation value C (Δx, Δy) exceeds the current first peak value fp, the cross-correlation value C (Δx, Δy) becomes the new first peak value fp, and the current first peak value fp becomes the new second peak value fp. Let it be the peak value sp.

  When the cross-correlation values C (Δx, Δy) are calculated for all the values of Δx and Δy in this way, the values of Δx and Δy from which the final first peak value fp was obtained are obtained from the first image. The amount of movement of the luminance pattern to the image. Then, Δx and Δy divided by the time interval Δt at which the first and second images are acquired become vx and vy which are the x component and y component of the two-component velocity vector in the inspection region.

  At this time, when the degree of correlation between the first image and the second image is high, the first peak value fp is prominently large and the second peak value sp is much smaller than the first peak value fp. When the degree of correlation between the first image and the second image is low, the difference between the first peak value fp and the second peak value sp becomes small. In such a case, 2 obtained based on Δx and Δy. The reliability of the component velocity vector is lowered.

  Therefore, in the present embodiment, the peak ratio fp / sp that is the ratio of the first peak value fp and the second peak value sp is calculated, and when the peak ratio fp / sp is greater than or equal to the threshold value 1.2, that is, the first peak value. If fp is 1.2 times or more of the second peak value sp, it is determined that the reliability of the two-component velocity vector is high, and conversely, the first peak value fp is 1. If it is less than twice, it is determined that the reliability of the two-component velocity vector is low, and the two-component velocity vector is deleted as an erroneous vector.

  The horizontal axis in FIG. 7 is the peak ratio, and the vertical axis is the number of erroneous vectors. From this graph, it can be seen that the number of erroneous vectors is extremely small in the region where the peak ratio is 1.2 or more, and the number of erroneous vectors increases rapidly in the region where the peak ratio is less than 1.2.

  The method for calculating the two-component velocity vector in each inspection area from the two images acquired at the two CCD cameras 32A at the two times t1 and t1 'has been described. Similarly, the two CCD cameras 32B can calculate the two component velocity vectors at the two times t1 and t1. By comparing the two images acquired at t1 ′, a three-component velocity vector in each inspection region can be calculated.

  The two-component velocity vector of the inspection area obtained from the image of the first CCD camera 32A and the two-component velocity vector of the inspection area obtained from the image of the second CCD camera 32B are the same as those of the first and second CCD cameras 32A and 32B. The imaging direction differs, that is, the actual three-component velocity vector differs depending on the parallax as seen from different directions.

  Therefore, from the two two-component velocity vectors and the relative positional relationship of the first and second CCD cameras 32A and 32B with respect to the laser sheet Ls, the in-plane velocity (in the X-axis direction) of the laser sheet Ls is obtained by calibration. A three-component velocity vector consisting of the velocity Vx and the velocity Vy in the Y-axis direction and the out-of-plane velocity (velocity Vz in the Z-axis direction) can be calculated corresponding to each position on the irradiation surface.

  When the three-component velocity vector in the irradiation surface of the laser sheet Ls is calculated as described above, the main traverser 23 and the first and second sub-traversers 24A and 24B are moved in the Z-axis direction while maintaining a fixed positional relationship. In other words, by moving the laser sheet Ls and the first and second CCD cameras 32A and 32B in the Z-axis direction while maintaining a fixed positional relationship, the laser sheet Ls irradiates the irradiation surface shifted in the Z-axis direction. Then, a three-component velocity vector at each position on the new irradiation surface is calculated. In FIG. 1, a state in which the irradiation surface of the particle image flow velocity measuring device 15 is at a position on the front end (upstream end) side of the object 12 is indicated by a solid line, and a state at a position on the rear end (downstream end) side of the object 12. Is indicated by a chain line.

  By stacking the results obtained by performing this operation on a plurality of irradiation surfaces separated by a predetermined distance in the Z-axis direction in the Z-axis direction, all three-component velocity vectors in the three-dimensional space around the object 12 can be measured. Thus, the velocity field formed around the object 12 can be measured in detail.

  In addition, since the images of the tracer particles distributed on the irradiation surface are sequentially acquired while moving the position of the laser sheet Ls in the axis Z direction, the time at which each image is acquired is different, but the flow around the object Is measured as a steady flow, the time lag is not a problem.

  If the positions of the first and second CCD cameras 32A and 32B are fixed and only the laser sheet irradiation means 31 is moved in the Z-axis direction, the first and second CCD cameras are moved along with the movement of the laser sheet irradiation means 31. Since the relative positional relationship with 32A and 32B changes, the calibration for calculating the three-component velocity vector from the two two-component velocity vectors obtained from the images of the first and second CCD cameras 32A and 32B is laser. Each time the position of the sheet Ls is moved, there is a problem that the number of calibration steps increases.

  However, according to the present embodiment, since the laser sheet irradiation means 31 and the first and second CCD cameras 32A and 32B move in the Z-axis direction while maintaining a certain positional relationship, every time the position of the laser sheet Ls moves. Calibration man-hours can be reduced by eliminating the need to perform calibration.

  The embodiments of the present invention have been described above, but various design changes can be made without departing from the scope of the present invention.

  For example, in the embodiment, a direct cross-correlation method is used to determine an erroneous vector based on the first and second peak values fp, sp calculated from the cross-correlation values C (Δx, Δy) of two images. In addition to the direct cross-correlation method, other correlation methods such as FFT cross-correlation method and auto-correlation method can be used.

  In the embodiment, the stereo PIV using the first and second CCD cameras 32A and 32B is employed. However, the present invention can also be applied to a PIV using a single CCD camera.

12 Object 13 Tracer particle supply means 32A First CCD camera (imaging means)
32B Third CCD camera (imaging means)
fp first peak value sp second peak value

Claims (2)

The tracer particles are supplied from the tracer particle supply means (13) to the flow field around the object (12), the reflected light is imaged by irradiating the tracer particles with a time difference twice and irradiating the laser light (32A, 32B). In the particle image velocimetry method for obtaining the velocity vector of the flow field based on two images of the obtained tracer particles,
The two images are each divided into a plurality of inspection regions, and the first peak value (fp) and the second peak value (sp) of the cross-correlation value of the luminance pattern of the tracer particles in each inspection region of the two images. To determine the reliability of the velocity vector by comparing the velocity vectors.   The particle image according to claim 1, wherein when the first peak value (fp) / second peak value (sp) ≧ 1.2 is established, it is determined that the reliability of the velocity vector is high. Flow rate measurement method.

Images