JPH083492B2 - Flow image analyzer - Google Patents

Description

TECHNICAL FIELD The present invention relates to a flow image analysis device for analyzing a flow of a fluid.

[Conventional technology]

Such an image analysis device is, for example, as shown in FIG.
An image pickup device 2 for picking up an image of the flow of the fluid 1 in which particles (tracers) are mixed, and a signal processing section 3 for calculating the movement of particles in the fluid 1 as a velocity vector based on an image signal input from the image pickup device 2. The data output unit 4 outputs the calculated velocity vector in a predetermined display pattern. As shown in FIG. 8, the signal processing unit 3 detects the particles included in the image x (t) input at a certain time point t and 1
The image x (t- input at the time (t-1) before the frame.
Correlate the particles in 1) one by one. This association generally associates the particles at the closest position between the two images. Then, the velocity vector (the distance between the particles is the velocity, and the direction of the straight line connecting the particles is the direction) is determined from the relationship of the center coordinates of the particles.

By accumulating the velocity vector thus obtained in one image, velocity vectors at various points on the screen can be obtained and the motion of the fluid can be analyzed (see, for example, JP-A-61-153774).

[Problems to be solved by the invention]

By the way, in such a conventional flow image analysis apparatus, images x (t) and x (t-
Since the particles are associated between 1) and the velocity vector is obtained based on the center coordinates of the particles in each image,
If there are many particles in one image or if there is a background in the image, there is a case that the correspondence of particles does not work well as shown in Fig. 9 b and c, and the calculation of the velocity vector of the flow is erroneous. There was a problem.

Therefore, an object of the present invention is to enable the velocity vector of particles to be accurately calculated even when a large number of particles are present in one image or when the image has a background.

[Means for solving problems]

In order to achieve the above object and solve the problems of the prior art, the flow image analysis device according to the present invention is a flow image analysis device that images the motion of a fluid mixed with particles to calculate a flow velocity vector, Subtraction processing means for obtaining a particle image by obtaining a frame difference between two images captured in time series based on image information input from an image pickup device, and an image length corresponding to each particle in the particle image. And a directional component calculating means for calculating a directional component of the flow from an inclination angle of an image corresponding to each particle based on the particle image.

[Action]

According to the present invention, a background image that does not move is removed by obtaining a frame difference between two images that are sequentially captured in time based on sequentially input image signals, and only a particle image is obtained. The velocity component of the velocity vector can be calculated by obtaining the length of the image corresponding to each particle in the image, and the direction component of the velocity vector can be calculated by obtaining the tilt of the image corresponding to the particle based on the particle image. it can.

〔Example〕

An embodiment of the present invention will be described below with reference to the accompanying drawings. FIG. 1 shows an example of a flow analysis apparatus according to the present invention.

In the figure, 11 is an image pickup device for picking up the movement of a fluid in which particles (tracers) are mixed, 12 is input image information from the image pickup device 11, and image information of the image picked up at a predetermined time and the image information Is subtracted from the image information of the images one frame before, which are consecutively captured in time, and the absolute value is calculated to obtain 2
A subtraction processing unit for removing the background image (an object that does not move) by binarization, 13 performs thinning processing based on the output of the subtraction processing unit 12, obtains the center line of the particle, and determines the pixel corresponding to the particle. Particle image A (i, j) with 1 "and other pixels as" 0 "
A thinning processing unit for generating a particle image A (i, j)
A velocity vector calculation unit that obtains a velocity vector of a particle based on a velocity component calculation unit 15 that obtains a velocity component by calculating the length of the particle, and a direction component calculation unit 16 that obtains a direction component from the tilt angle of the particle. ing. Reference numeral 17 is a data output unit.

 Next, the operation of this device will be described with reference to FIG.

First, the data output unit 17 uses the image information H (i, j), which will be described later,
(I, j) and (i, j) are cleared (S-1). Next, the subtraction processing unit 12 obtains the frame difference between the two images, binarizes the absolute value (S-2), and then the thinning processing unit 13 performs thinning processing to obtain the particle image A (i, j). Get (S
-3, S-3a). Then, the velocity component calculation unit 15 performs labeling, separates the particles (S-4, S-4a), obtains the length of each particle, and the value of the length is given to the pixel where the particle exists. The placed image L (i, j) is generated (S-5). At this time, all the particles having the same label are imaged as a density with a value of the same length (S-5a). On the other hand, the direction component calculation unit 16
For each pixel of the particle image A (i, j), as shown in FIG. 3, the tilt angle of the central pixel is obtained from the connection relation of the pixels in 8 neighborhoods (3 × 3) (S-6). At this time, the tilt angle θ
(I, j) can be obtained by coding into 8 kinds of angles as shown in FIG. 3 (S-6a), but the tilt angle f (i, j) may be stored in the image memory as it is. Alternatively, instead of the (3 × 3) connection relationship, it is possible to obtain the inclination by using the connection relationship in the (5 × 5) mask to improve the accuracy.

By the above processing, since the velocity component of the velocity vector of the flow is obtained by the image L (i, j) and the direction component of the velocity vector of the particle is obtained by the image θ (i, j),
Hereinafter, the data output unit 17 edits the velocity vector data into a predetermined output data format. First, in order to time-average the velocity vector, it is once decomposed into x-direction and y-direction components, and x (i, j) and y (i, j) are obtained (S-7, S
-8), Accumulate each data. At this time, image information H (i, j) indicating how much data is added at each pixel is also created (S-9). The above process is repeated a predetermined number of times (N times). Then, the cumulative image X (i,
j) and Y (i, j) are divided by the cumulative number of times H (i, j) to obtain averaged component images (i, j) and (i, j) (S-1
0).

From these two images, the averaged velocity vector at each pixel can be obtained.

Then, when the speed is imaged, for example, (i,
The square root of the sum of squares of (j) and (i, j) may be taken and displayed (S-11, S-11a).

Through the above processing, the flow analysis result can be obtained.

FIG. 5 shows another embodiment of the present invention. In the figure, 21 is an imaging device, 22 is a subtraction processing unit, 23 is a thinning processing unit, and 24 is a histogram creation unit that accumulates the output data of this thinning processing unit 23 and creates a histogram.

Next, the operation of this device will be described with reference to FIGS. 5 and 6. First, the histogram creation unit 24 clears the accumulated image V (t) (5-1).

Next, the subtraction processing unit 22 inputs an image from the imaging device 21 (5-2, 6-2b ₁ , 6-2b ₂ ) and the currently input image x
(T) and the previously input image (image one frame before) x
The difference from (t-1) is taken and binarized (5-3, 6-3b).
By doing so, the background (non-moving portion) is all removed, and an image y (x) showing the movement of particles is obtained. That is, the image y (t) contains only the movement of particles. Next, the thinning processing unit 23 thins the image y (x) and obtains the center line Z (t) with a width of 1 (5-4, 6-4b). This process is a process for obtaining an accurate image for details by expressing the streamline as thin as possible.

Next, the histogram creation unit 24 uses the thinned image Z (t)
Then, the sum of the accumulated image V (t) is obtained and replaced with the accumulated image V (t) (5-5, 6-5b).

 The above process is repeated N times.

When N times of images are accumulated, V (t) has a tracer (particle)
The image represents the frequency of the locus of.

That is, V (t) is an image in which the density is high where there are many traces of the tracer, and the density is low where there are few traces of the tracer, that is, where there is not much flow change. Therefore, it can be said that the concentration of V (t) represents the density of the flow.

It is also possible to display the tracer trajectory itself in a follow-up manner by binarizing V (t) and reducing the number of times N.

It should be noted that all of the above processing can be performed by sequentially scanning a straight line and can be realized by hardware for each block, so real-time processing can also be performed.

〔The invention's effect〕

As described above, the flow image analysis apparatus according to the present invention obtains a particle image by taking the frame difference of two images, and calculates the velocity component from the length of the image corresponding to each particle to the image corresponding to each particle. Since the directional component is calculated from the tilt angle to calculate the velocity vector, the velocity vector of the particle can be accurately calculated without the need of associating the particles even if there is a background image or a large number of particles exist in the image. Since it can be obtained, it is possible to accurately analyze the flow even from an image in which the flow is swirling and swirling.

[Brief description of drawings]

FIG. 1 is a block diagram showing an example of a flow analysis apparatus according to the present invention, FIG. 2 is a flow chart showing an operation example of the flow analysis apparatus according to the present invention, and FIG. 3 is an operation of a tilt angle calculation means according to the present invention. FIG. 4 shows an example, FIG. 4 is a block diagram showing a flow image analysis apparatus according to another embodiment of the present invention, and FIGS. 5 and 6 are diagrams showing an operation example of the flow image analysis apparatus shown in FIG. FIG. 7 is a block diagram showing an example of a conventional flow image analyzing device, FIG. 8 is a diagram showing an example of processing contents by a conventional flow image analyzing device, and FIG. 9 is a diagram showing a case where particle association is wrong. It is a figure. 12 ... Subtraction processing unit 13 ... Thinning processing unit 14 ... Velocity vector calculation unit 15 ... Velocity component calculation unit 16 ... Direction component calculation unit 17 ... Data output unit

Claims (1)

[Claims] 1. A flow image analysis apparatus for imaging a motion of a fluid mixed with particles to calculate a velocity vector of a flow, wherein two images are imaged continuously in time based on image information input from the imaging apparatus. Subtraction processing means for obtaining the particle image by calculating the frame difference of the particle image, velocity component calculation means for calculating the velocity component of the flow from the image length corresponding to each particle in the particle image, and each particle based on the particle image And a direction component calculation means for calculating a direction component of the flow from the tilt angle of the image corresponding to the flow image analysis device.