JP2006194788A - Particle image processing method and device, and its program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To accurately acquire feature information of a particle from a particle image. <P>SOLUTION: This particle image processing method comprises a process of imaging the particle, a process of performing interpolation processing of increasing the number of constituted pixels of a multi-gradation image formed of a plurality of pixels obtained by imaging, and a process of acquiring information showing the feature of the particle based on the interpolation image having been subjected to the interpolation processing. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a particle image processing method and apparatus for obtaining information on particles by imaging particles and analyzing the image of the particles.

In recent years, in the manufacturing process and quality control process of various particulate materials such as fine ceramic particles and toner, pigments, cosmetic powders, food additives, chemicals, various information about each particle of the particulate material has been measured, It is becoming important to analyze particles.
For example, as an apparatus for analyzing particles, a flow cell for converting a flow of a particle suspension into a flow surrounded by a sheath liquid, light irradiation means for irradiating light to the converted suspension flow, and irradiation The image analysis means measures the particle data about the area and the perimeter of each imaged particle image, the image analysis means for analyzing the captured particle image, and the display means. Then, a calculation means for calculating the particle and circularity from the particle data, a histogram based on the particle size frequency data based on the particle size, and a two-dimensional scattergram based on two parameters corresponding to the particle size and the circularity are created. A chart creation means for displaying each on the display means, a storage means for storing each captured particle image, and a particle image call for collectively displaying each particle image stored in the storage means on the display means Particle image analyzer which comprises a means is known (e.g., see Patent Document 1).
JP-A-8-136439

However, in the conventional method or apparatus, when the resolution (resolution) of the particle image cannot be sufficiently obtained for the particle image having a small diameter, the equivalent circle diameter or the circularity obtained from the binarized image of the particle is obtained. In some cases, the calculation accuracy of the value is lowered.
The present invention has been made in view of such circumstances, and binarizes the image by interpolating the pixels of the image data obtained by imaging the particles to increase the resolution of the particle image. It is an object of the present invention to provide a particle image processing technique characterized by improving the calculation accuracy of the equivalent circle diameter and circularity obtained from an image.

  The present invention is based on a step of imaging particles, a step of performing interpolation processing to increase the number of constituent pixels for a multi-gradation image composed of a plurality of pixels obtained by imaging, and an interpolation image subjected to interpolation processing There is provided a particle image processing method including a step of obtaining information representing the characteristics of particles.

  The present invention also provides pixel interpolation means for performing interpolation processing to increase the number of constituent pixels by interpolating pixels with respect to a multi-tone image composed of a plurality of pixels obtained by imaging particles, and interpolation performed by interpolation processing. The present invention provides a particle image processing apparatus comprising analysis means for obtaining feature information that represents the feature of a particle based on an image.

  According to the present invention, pixels are interpolated from the multi-tone image, and the feature information of the particles is obtained based on the interpolated image. Therefore, highly accurate feature information can be obtained.

Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 shows an external view of a particle image analyzer as an embodiment of the present invention. As shown in FIG. 1, the particle image analysis apparatus includes a particle image imaging apparatus 21 and a data analysis apparatus 22. The particle imaging device 21 and the data analysis device 22 are connected by a communication cable, and a captured image captured by the particle imaging device 21 is transmitted to the data analysis device 22 via the communication cable.

  The data analysis device 22 is a particle image processing device 11 that performs image processing on the captured image transmitted from the particle imaging device 21, and a monitor television for displaying the particle image, analysis data, and the like output from the particle image processing device 11. 12, a keyboard 13 for inputting and operating data on the particle image processing apparatus 11, and a program reading unit 15 (see FIG. 4) for reading a program operating on the particle image processing apparatus 11.

  Next, the configuration of the imaging unit of the particle imaging device 21 will be described with reference to FIG. The suction pipette 1 is connected to a pump 18 via a sample filter 2, a valve 16 and a valve 17. A sample charging line 3 for charging the sample is connected between the valve 16 and the valve 17, and the particle suspension charged in the sample charging line 3 is pushed out to the sample charging line 3. A syringe 4 is connected. The sample charging line 3 is connected to the flow cell 5. The flow cell 5 is a cell that can form a flat flow by a hydrodynamic effect by wrapping and flowing the flow of the particle suspension in a sheath liquid. The sample liquid bottle 6 is connected to the waste liquid chamber 14 via the sheath liquid chamber 7 and the flow cell 5.

  The strobe 8 and the video camera 10 are disposed with the flow cell 5 interposed therebetween, and an objective lens 9 is disposed between the flow cell 5 and the video camera 10. In this embodiment, the imaging magnification is 20 times, and the resolution is 512 × 512 pixels.

Next, the operation of the particle imaging device 21 will be described.
In FIG. 2, when the valves 16 and 17 are opened, the particle suspension is sucked from the suction pipette 1 by the pump 18, passes through the sample filter 2, and is drawn into the sample charging line 3 at the top of the flow cell 5. The sample filter 2 removes coarse particle dust in the suspension so that the flow cell 5 having a narrow (narrow) flow path is not clogged. The sample filter 2 also has an effect of loosening coarse agglomerates.

  The particle suspension drawn into the charging line 3 is guided to the flow cell 5 by closing the valves 16 and 17 and operating the syringe 4, and the suspension is discharged from the tip of an internal sample nozzle (not shown). Extruded. At the same time, the sheath liquid is also sent from the sheath liquid bottle 6 to the flow cell 5 through the sheath liquid chamber 7, the particle suspension is surrounded by the sheath liquid, and the suspension is squeezed flatly in terms of fluid dynamics. And is discharged through the waste liquid chamber 14. By periodically irradiating pulse light from the strobe 8 to the suspension flow narrowed in this manner, a still image of particles is captured by the video camera 10 via the objective lens 9.

The optimum solvent for suspending the particles may be selected according to the particle characteristics (particle size and specific gravity).
Further, in order to reliably narrow the flow of the suspension flatly or finely, it is preferable to change the viscosity or specific gravity of the sheath liquid according to the characteristics of the suspension, for example, according to the viscosity or specific gravity of the solvent. Although not shown in FIG. 2, a plurality of types of sheath liquid bottles may be provided, and a mechanism that can easily switch the type of sheath liquid to be used according to the sample to be measured may be added.

  If the flat surface of the suspension flow is imaged by the video camera 10, a particle image can be captured over the entire imaging area of the video camera 10, and a large number of particles can be imaged by one imaging as shown in FIG. . In addition, since the distance between the center of gravity of the imaged particle and the imaging surface of the video camera 10 can be made almost constant, a focused particle image can always be obtained regardless of the size of the particle.

  FIG. 4 is a block diagram showing the configuration of the data analysis device 22. As shown in the figure, the data analysis device 22 performs image processing on the captured image transmitted from the video camera 10, and the particle image and analysis data output from the particle image processing device 11. It is mainly composed of a monitor TV 12 for displaying, a keyboard 13 for inputting and operating data on the particle image processing apparatus 11, and a program reading unit 15 for reading a program operating on the image processing apparatus 11.

  The particle image processing apparatus 11 includes an image interpolating unit that interpolates pixels with respect to a grayscale image composed of a plurality of pixels obtained by imaging by the video camera 10, that is, a high resolution unit 11 f and a high resolution unit. Obtained by edge tracing, a binarization unit 11a that converts the image signal of the interpolation image obtained from the unit 11f into binarized data, an edge trace unit 11b that performs edge tracing of the binarized particle image data An analysis unit 11c for analyzing data, a data storage unit 11d for storing image data and processing data, and a program storage unit 11e for storing a particle image analysis program read from a program recording medium by the program reading unit 15 Prepare.

Here, the program reading unit 15 includes a CD-ROM drive, and the program recording medium includes a CD-ROM.
Then, the image processing unit 11 processes the image signal from the video camera 10 according to the processing condition input from the keyboard 13 and displays the processing result on the monitor television 12.
Here, the particle image processing apparatus 11 is configured by a personal computer including a CPU, a ROM, and a RAM.

The procedure of image processing in the data analyzer 22 is shown in the flowchart of FIG.
First, an image signal from the video camera 10 is captured by the image processing unit 11, A / D converted, and captured as image data that is a grayscale image (step S1). Next, background correction for correcting the intensity unevenness (sudding) of the irradiation light with respect to the suspension flow is performed (step S2).

  Specifically, this background correction is performed by capturing in advance image data obtained by light irradiation when not passing through the particle flow cell 5 before measurement, and the image data and image data of the actual particle imaging screen. Is a process generally known as image processing.

  Next, high resolution processing, that is, pixel interpolation processing is performed (step S3). Specifically, for every pixel of the particle image, an interpolation pixel is inserted between two adjacent pixels. The interpolation pixels are formed so as to have gradations in adjacent pixels, here, average luminance values, and here, so-called linear interpolation is performed.

First, a case where one pixel is interpolated between adjacent pixels will be described. Focusing on the four pixels Pa, Pb, Pc, and Pd that are in contact with each other as shown in FIG. 6A, these are interpolated pixels Pab, Pac, Pcd, Pbd, and Pm as shown in FIG. 6B. Is interpolated by In this case, each brightness of the pixels Pa, Pb, Pc, Pd is
Pa = A
Pb = B
Pc = C
Pd = D
When each of the interpolated pixels Pab, Pac, Pcd, Pbd, Pm is
Pab = (1/2) A + (1/2) B
Pac = (1/2) A + (1/2) C
Pcd = (1/2) C + (1/2) D
Pbd = (1/2) B + (1/2) D
Pm = (1/4) A + (1/4) B + (1/4) C + (1/4) D
It becomes.

Next, a case where two pixels are interpolated between adjacent pixels will be described. Focusing on the four pixels Pa, Pb, Pc, and Pd that are in contact with each other as shown in FIG. 6A, these are interpolated pixels Pab1, Pab2, Pac1, Pac2, and Pcd1 as shown in FIG. 6C. , Pcd2, Pbd1, Pbd2, Pm1, Pm2, Pm3, Pm4. In this case, each brightness of the pixels Pa, Pb, Pc, Pd is
Pa = A
Pb = B
Pc = C
Pd = D
The luminances of the interpolated pixels Pab1, Pab2, Pac1, Pac2, Pcd1, Pcd2, Pbd1, Pbd2, Pm1, Pm2, Pm3, and Pm4 are
Pab1 = (2/3) A + (1/3) B
Pab2 = (1/3) A + (2/3) B
Pac1 = (2/3) A + (1/3) C
Pac2 = (1/3) A + (2/3) C
Pcd1 = (2/3) C + (1/3) D
Pcd2 = (1/3) C + (2/3) D
Pbd1 = (2/3) B + (1/3) D
Pbd2 = (1/3) B + (2/3) D
Pm1 = (2/3) (2/3) A + (2/3) (1/3) B + (1/3) (2/3) C + (1/3) (1/3) D
Pm2 = (2/3) (1/3) A + (2/3) (2/3) B + (1/3) (1/3) C + (1/3) (2/3) D
Pm3 = (1/3) (2/3) A + (1/3) (1/3) B + (2/3) (2/3) C + (2/3) (1/3) D
Pm4 = (1/3) (1/3) A + (1/3) (2/3) B + (2/3) (1/3) C + (2/3) (2/3) D
It becomes.

As described above, in the linear interpolation, the pixel density (luminance in this embodiment) of the interpolation pixel is obtained by a weighted average based on the distance from the surrounding four pixels. The distance between the pixels is the distance between the pixel centers, and the distance between adjacent pixel centers of the surrounding four pixels (pixels before interpolation) is 1. In FIG. 6A, the distance in the Y direction from the pixel center of the pixel Pa to the interpolation pixel center is P, the distance in the X direction from the pixel center of the pixel Pa to the interpolation pixel center is Q, and the pixels Pa, The pixel density of Pb, Pc, Pd is
Pa = K
Pb = L
Pc = M
Pd = N
The pixel density of the interpolation pixel is
(1-P) (1-Q) K + (1-P) QL + P (1-Q) M + PQN
It is expressed as
Therefore, the pixel density (luminance in this embodiment) of the interpolation pixel can be easily obtained by linear interpolation using a weighted average.

  Next, contour enhancement processing is performed as preprocessing for accurately extracting the contour of the particle image (step S3a). Specifically, generally well-known Laplacian enhancement processing is performed.

  Next, when the image data is binarized at a certain appropriate threshold level, each particle image becomes a binarized image as shown in FIG. 7 (step S4). Next, for each binarized particle image, it is determined whether the edge point is a (contour pixel representing the contour), and in which direction is the edge point adjacent to the edge point of interest Information, that is, a chain code is generated (step S5). Next, edge tracing of the particle image is performed with reference to this chain code, and the area (total number of pixels) St, the orthogonal count number Et, the skew count number Es, and the corner count of each particle image are analyzed as the particle image analysis parameters. The number Cn is obtained (step S6).

  Here, the processing in steps S5 and S6 will be described in detail with reference to FIG. 7. The center points a to n of the contour pixels (edge points) forming the contour of the particle image (hatched portion in FIG. 7) are linear. When all the contour pixels are sequentially focused in the counterclockwise direction (or clockwise direction) from the point a, and the straight line connecting the current target contour pixel and the next target contour pixel is directed in the vertical or horizontal direction, the current target contour The pixel is regarded as the first pixel, and when the straight line connecting the current target contour pixel and the next target contour pixel is directed in the oblique direction, the current target contour pixel is regarded as the second pixel, and the number of the first and second pixels is determined. Count.

Here, the total number of first pixels is the orthogonal count number Et, and the total number of second pixels is the skew count number Es.
In FIG. 7, when a straight line connecting one contour pixel of interest and two contour pixels adjacent to both sides thereof intersects with the angle of interest of the contour pixel, the contour pixel is regarded as a bending point, and the entire contour Focusing on each of the pixels, the number of inflection points is counted. The total number of bending points is the corner count number Cn.
In the particle image shown in FIG.
Area St = total number of pixels = 25,
Since the first pixel is a pixel including points c, e, g, h, j, k, l, m, Et = 8,
Since the second pixel is a pixel including points a, b, d, f, i, and n, Es = 6.
Since the bending points are points a, c, d, e, f, g, i, j, and n, Cn = 9.

When the necessary imaging process is completed (step S7), the perimeter length L is first calculated by the following equation based on the analysis parameters obtained for each particle image (step S8).
L = 0.980 * Et + 1.406 * Es-0.091 * Cn ... (1)
(However, the distance between two pixels is 1.)
Equation (1) is known as Vossepoel's equation.
Next, the area S is calculated by the following equation (step S9).
S = St−0.5L (2)
(However, the distance between two pixels is 1.)
Next, the equivalent circle diameter Rs is calculated by the following equation (step S10).
Rs = a × S ^1/2 × k + b (3)
Here, k is the size of one pixel, and a and b are correction coefficients.

Next, the circularity X is calculated by the following equation (step S11).
X = (perimeter of circle having the same area as particle image) / (perimeter of particle image)
= ΠRs / L (4)
It becomes.
As described above, it is possible to obtain the shape information representing the shape of the particle such as the perimeter, area, equivalent circle diameter, and circularity for each of a large number of particle images. Can be analyzed.

FIG. 8 is a histogram of circularity distribution obtained by imaging latex particles having a diameter of 2 μm. FIG. 8A shows a case where pixel interpolation is performed as described above, and FIG. 8B shows a case where pixel interpolation is not performed (comparative example). These histograms were created by extracting 3469 particles from 1000 frames of imaging pixels.
In FIG. 8B, the tail of the circularity extends to approximately 0.9, the number of particles represented by a circularity of 1.0 is extremely large, and the distribution is unnatural. On the other hand, in FIG. 8A, the skirt of the circularity completely exceeds 0.9, while the particles having a circularity of 1.0 are greatly reduced, the peak is approximately 0.98, the distribution width is narrow, It is sharp.
Therefore, FIG. 8A shows that the measurement of the circularity is performed more accurately than in FIG. 8B.

  In this way, the pixel is linearly interpolated into the grayscale image of the particle obtained by imaging the particle to form an enlarged image, so that the pixel is compared with the binary image of the grayscale image before the pixel is linearly interpolated. The binarized image of the enlarged image obtained by linear interpolation increases the number of contour pixels forming the contour. Therefore, the shape information representing the shape of the particle such as the perimeter, area, equivalent circle diameter, and circularity obtained based on the contour pixel more reflects the shape of the particle.

  The image analysis target of the above embodiment includes fine ceramics, pigments, inorganic powders such as cosmetic powders, and organic powders such as food additives, and even particles made of polycrystals. Good.

  In the above embodiment, the grayscale image obtained by imaging particles is used to perform linear interpolation as a high resolution, but using a multi-tone image such as a color image obtained by imaging particles, Linear interpolation may be performed.

  In the above embodiment, the strobe is used as the pulse light source, but a pulse laser light source may be used.

  Further, in the above embodiment, the slob and the video camera are arranged with the flow cell interposed therebetween, but when the particle suspension is converted into a flat flow in the flow cell, the strobe is placed on the flat surface of the particle suspension flow. It is preferable to irradiate light orthogonally and to arrange the imaging unit on the optical axis.

In the above embodiment, a CD-ROM is used as a recording medium. However, a magnetic tape, a cassette tape, a magnetic disk such as a floppy (registered trademark) disk or a hard disk, an optical disk such as a CD-MO / MD / DVD, an IC card (memory). (Including a card) / optical card, or a semiconductor memory such as a mask ROM, EPROM, EEPROM, or flash ROM.
In addition, the particle image analysis apparatus may have a system configuration that can be connected to a communication network including the Internet, and the recording medium may be a medium that fluidly carries the program so as to download the program from the communication network. The download program is stored in advance in the particle image processing apparatus.

1 is an external view of a particle image analyzer according to an embodiment of the present invention. It is principal part structure explanatory drawing of the particle image apparatus of the Example by this invention. It is explanatory drawing of the image by the Example of this invention. It is a block diagram which shows the structure of the data analyzer of the Example by this invention. It is a flowchart which shows operation | movement of the data analyzer of the Example of this invention. It is explanatory drawing of the pixel of the particle image and the pixel for interpolation in the Example of this invention. It is explanatory drawing of the binarized image in the Example of this invention. It is a histogram of circularity distribution of an Example and a comparative example.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Suction pipette 2 Sample filter 3 Sample charging line 4 Sheath syringe 5 Flow cell 6 Sheath liquid bottle 7 Sheath liquid chamber 8 Strobe 9 Objective lens 10 Video camera 11 Particle image processing device 12 Monitor television 13 Keyboard 14 Waste liquid chamber 21 Particle imaging device 22 Data analyzer

Claims (10)

  A step of imaging particles, a step of performing interpolation processing to increase the number of constituent pixels for a multi-tone image composed of a plurality of pixels obtained by imaging, and a feature of particles based on the interpolated interpolation image A particle image processing method.   Pixel interpolation means for performing interpolation processing to increase the number of constituent pixels for a multi-tone image made up of a plurality of pixels obtained by imaging particles, and feature information representing the characteristics of the particles based on the interpolated interpolation image A particle image processing apparatus comprising:   The particle image processing apparatus according to claim 2, wherein the pixel interpolation unit performs interpolation by inserting at least one interpolation pixel between adjacent pixels.   The particle image processing apparatus according to claim 3, wherein the interpolation pixel includes a gradation value obtained based on a gradation in the adjacent pixel and a weighted average that performs weighting according to a distance between the pixels.   The particle image processing apparatus according to claim 4, wherein the gradation value is luminance.   The analysis unit generates a binarization unit that converts the interpolation image into a binarized image, and a chain code that tracks contour pixels that form the contour of the particle image in the binarized image. 6. The particle image processing apparatus according to claim 2, further comprising a calculation unit configured to calculate feature information representing a feature of the particle based on the particle information.   The particle image processing apparatus according to claim 2, wherein the characteristic information is at least one of a circle-equivalent diameter and a circularity.   The particle image processing apparatus according to claim 2, wherein the multi-tone image is a grayscale image.   A flow cell that forms a flow of a suspension containing particles, a light source that irradiates light to the flow of the suspension, an imaging unit that images particles irradiated with light, and a plurality of images obtained by the imaging unit An image processing unit having pixel interpolation means for performing interpolation processing for increasing the number of constituent pixels for a multi-tone image composed of pixels, and analysis means for obtaining feature information representing the characteristics of particles based on the interpolated interpolation image. A particle image analyzer provided.   A step of performing interpolation processing for increasing the number of constituent pixels on a multi-tone image obtained by imaging particles with a particle imaging device, and a step of obtaining feature information representing particle characteristics based on the interpolated interpolation image An image processing program for causing a computer to execute.

Images