JP2000338029A - Particle size measuring apparatus and method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a particle size-measuring apparatus adaptable even to a granular object to be inspected wherein particles overlap each other closely and capable of reducing calculation cost. SOLUTION: The average value of the brightness of the image signal subjected to A/D conversion of a granular object to be inspected is calculated and a plurality of sets comprising first set values equal to or more than the average value and second set values equal to or less than the average value at least one of which are different from each other are calculated on the basis of the average value. When the image signal subjected to A/D conversion changes across the set value selected according to a predetermined rule among the first and second set values of one set in a horizontal or vertical direction, the pixel thereof is detected as a change point and the number of pixels detected as change points is counted and the particle size of the granular object 10 to be inspected is calculated on the basis of the number of pixels counted as change points with respect to each of a plurality of sets.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a particle size measuring apparatus and a particle size measuring method for measuring, by image processing, the average particle size of a granular specimen observed in a state where many particles overlap each other.

[0002]

2. Description of the Related Art Conventionally, when measuring the particle size of a granular specimen composed of a large number of particles, it has been a problem that particles are observed overlapping each other. As a method for separating images of overlapping particles, for example, “separation of overlapping particle images by iterative operation” (Sakagami, Takagi, IPSJ Transactions, 24
Vol. 5, No. 5, 1983).

According to this, for example, granules, cell clumps, T
In digital image processing of images of a large number of granular subjects such as iO ₂ fine particles, muscle fiber cross-sectional images, red blood cells, carbon fine particles, etc., a mask of a focused particle mass is obtained from the background of the original image, Find the contour points, assign a coordinate value to each contour point, a parameter of the center of the circumference where each contour point is assumed and the probability of this assumption, and sequentially update the value of this parameter by an iterative operation, At the time of convergence, clustering is performed by the central coordinate value to separate and identify the particles, and the particle diameter of each particle is obtained.

However, such a conventional particle size measurement can be applied when individual particles or several overlapping particles can be easily detected from the background, but can be applied when particles do not overlap and there is no background. However, since it is difficult to extract a particle mass, there is a problem that it is difficult to separate and identify particles. In addition, the calculation cost is large because it is obtained by iterative operation,
There is also a problem that it takes time.

In order to solve such a problem, the present applicant has already disclosed in Japanese Patent Application Laid-Open No. H11-63938 that the present invention can be applied to a granular object in which particles are overlapped without gaps, or can be calculated or calculated. We have proposed a particle size measurement device that can reduce costs. The particle size measuring device disclosed in the above publication includes an imaging unit that images a granular object and outputs an image signal, an A / D converter that A / D converts an image signal from the imaging unit, and an A / D converter. Average value calculation means for calculating an average value of brightness of the D-converted image signal; and a plurality of brightness values set based on the average value or the average value in a predetermined direction from the A / D converted image signal. A change point detecting means for detecting a pixel as a change point when changing over a specific one of the set values, and a counting means for counting the number of pixels detected as a change point by the change point detecting means Means, and a particle size calculating means for calculating the particle size of the granular object based on the number of pixels counted by the counting means.

In the fourth to sixth embodiments of the above publication, the change point detecting means sets the first set value fave + β larger than the average value fave based on the average value fave of the brightness.
And a second set value fave-β smaller than the average value fave
For example, when the binary image signal of an adjacent pixel takes “0” indicating that the brightness is small, and the brightness of the pixel of interest is larger than the first set value fave + β, The binary image signal of the corresponding pixel is set to “1”, indicating that the binary image signal of the adjacent pixel has high brightness.
When “1” is taken, if the brightness of the pixel of interest is smaller than the second set value fave−β, binarization is performed to set the binary image signal of the corresponding pixel to “0”. "0" to "1" between pixels where the binary image signal is adjacent.
Alternatively, a pixel that changes from “1” to “0” is detected as a change point. If a change point is detected on the condition that the average value fave is straddled, a slight change in brightness near the average value irrespective of the particle size will be affected.
Thus, the first set value fave + which is larger than the average value fave +
By setting β and a second set value fave−β smaller than the average value, minute brightness changes near the average value fave are removed, and only brightness changes that should contribute to particle size measurement are detected. ing.

[0007]

The larger the value of .beta., Which is the difference between the average value fave and the first and second set values, the less the effect of lightness changes near the average value fave, the less the above effect. growing. On the other hand, if the value of β is too large, there is a problem in that a change in brightness that should contribute to particle size measurement cannot be detected.

The present invention has been made in view of such a problem, and the invention according to claims 1 to 8 can be applied to a granular object in which particles are overlapped without gaps. It is an object of the present invention to further improve a particle size measuring device and a particle size measuring method capable of reducing calculation cost.

[0009]

In order to achieve the above object, a particle size measuring apparatus according to the present invention comprises an image pickup means for picking up an image of a granular object and outputting an image signal, and an image signal from the image pickup means. A / D converter for A / D conversion, average value calculating means for calculating an average value of brightness of the A / D converted image signal, and an average value calculated by the average value calculating means. A set of a first set value equal to or greater than the average value and a second set value equal to or less than the average value, wherein at least one of the first set value and the second set value is different from each other A set value calculating means for calculating a set of the first set value and a second set value of each set in the predetermined direction from the A / D converted image signal in a predetermined direction The pixel when changing across the set values selected according to the rules And change point detection means for detecting a reduction point, counting means for counting the number of detected pixels as change point by the change point detection means, for each of said plurality of sets,
A particle size calculation unit configured to calculate a particle size of the granular object based on the number of pixels counted by the counting unit.

Further, according to the particle size measuring method of the present invention, an image of a granular object is imaged, an image signal is A / D-converted, and an average value of brightness of the A / D-converted image signal is calculated. On the basis of the average value obtained, a first set value equal to or more than the average value and a second set value equal to or less than the average value are used as a set of the first set value and the second set value. At least one of a plurality of sets different from each other is calculated, and from the A / D-converted image signal, the brightness is determined in advance in a predetermined direction among the first set values and the second set values of each set. Detecting the pixel as a change point when changing over the set value selected according to the determined rule, counting the number of pixels detected as the change point, for each of the plurality of sets, Calculate the particle size of a granular object based on the number of pixels counted as a transition point That.

[0011] It is considered that the number of pixels counted by the apparatus and method of the present invention increases as the particle size of each particle of the granular specimen decreases. This is because the brightness pattern of the granular object image has finer and higher frequency components as the particle size of each particle is smaller, and changes vertically over the first set value or the second set value set based on the average value. This is because the number of times to be performed increases. Therefore, by counting the number of pixels that are the change points,
The average particle size can be determined based on the number of pixels.

For example, the average particle size r is determined by the number n of pixels to be counted.

[0013]

[Formula 1] r = k / n (1) Here, k is a constant, and k is, for example, an image of a granular object having a known particle size rs, A / D conversion of an image signal, a change point is detected, and a pixel detected as a change point is detected. Can be determined by the following equation: k = rs · ns

As described above, the larger the difference β between the average value and the first set value or the difference β between the average value and the second set value, the more contributed to the particle size measurement near the average value fave. While small lightness changes that should not be possible can be eliminated, the difference β
Is too large, it becomes impossible to detect a change in lightness that should contribute to the particle size measurement. Therefore, the difference β needs to be an appropriate value according to the measurement target. By the way, measuring the particle size from an image is based on the premise that the particle size is a certain size or more. For example, if the particle size is less than the image resolution, no particles can be observed,
Naturally, the particle size cannot be measured. According to the particle size measurement method of the prior application, the particle size needs to be large with respect to the image resolution so that a change in brightness that should contribute to the particle size measurement can be detected. This measurable minimum particle size increases as the value of β increases for the above-described reason. Conversely, when the value of β is small and the grains are small, a slight change in brightness rarely causes a problem as noise.

That is, by setting a value of β effective for removing a minute change in brightness, which is noise, a smaller value of β (extremely β = 0) can be measured for a small particle that can be measured. The particle size cannot be measured. Here, “the particle size of a small particle cannot be measured” means that a change in lightness that should contribute to the particle size measurement cannot be reliably detected. Therefore, the smaller the particle size, the smaller the number of pixels detected as a change point. This means that the relationship that the number of particles becomes larger does not hold. For example, when trying to find the particle size by the above-described formula (1), the particle size r becomes larger than the actual value because n is smaller than the original value. That is.

In the present invention, a plurality of sets each including a first set value and a second set value are calculated, and for each of the plurality of sets,
The problem is solved by counting the number of pixels detected as a change point and calculating the particle diameter based on the counted number of pixels.

For example, when the average value is fave, two sets (fave + β1, fave−β1), (fave + β2,
fave-β2), the number of pixels n1 and n2 at the change point detected respectively are obtained. Here, it is assumed that β1 is a value effective for removing a minute change in brightness, which is noise, and β1
Let 2 be a value that includes 0 and is smaller than β1. The particle size is determined by r = k1 / n1 for the range of the particle size that is not too small, and r = k2 / n2 for the small particle size. However, the size of the particle size is generally known as a measurement result, and it is not known in advance which method should be used to determine the particle size. Therefore, judgment is made based on the result of obtaining n1 and n2. FIG. 5 schematically shows the relationship between the number n of pixels at the change point and the particle size r due to the difference in the value of β.
When the set value is fave ± β1, if the particle size is smaller than r0, it is not possible to reliably detect a change in brightness that should contribute to the particle size measurement. It has become larger than the value of. on the other hand,
When the set value is fave ± β2 (<β1), when the particle size is large (r≫r0), the influence of a minute change in brightness, which is noise, becomes large and deviates from the actual value. Therefore, for example, a particle size r0 near the lower limit at which an accurate particle size is obtained using a set value fave ± β1 effective for removing a minute change in brightness, which is noise, is set as a threshold, and r2 = k2 / n
When the value r2 obtained by the equation (2) is equal to or smaller than the threshold value r0, r2 is adopted as the particle diameter. When the value r2 is equal to or larger than the threshold value r0 or larger than r0, the value r1 obtained by the equation of r1 = k1 / n1 is used as the particle diameter. The above can be realized by employing

The average brightness value calculated by the average value calculation means may be the average brightness value of the entire image or the average brightness value of a part of the image. In order to detect the pixel as a change point when changing over a set value selected based on an average value of brightness and selected according to a predetermined rule, the effect of brightness fluctuation due to illumination change Can be excluded.

Optionally, said change point detecting means comprises:
Based on the brightness of the pixel of interest and the binary image signal of the pixel adjacent to the pixel in the horizontal direction, the “first value” or “second value”
In which the binary image signal of an adjacent pixel has low brightness.
When the brightness of the pixel of interest is greater than the first set value, the binary image signal of the pixel of interest is set to “first value”, and the binary image signal of the adjacent pixel is When the brightness of the target pixel is smaller than the second set value, the binary image signal of the target pixel is set to the "second value" when the "first value" indicating that the brightness is large is taken. In cases other than the above, the horizontal binarizing means for making the binary image signal of the pixel of interest the same as the adjacent pixels, and the binary image signal obtained by the horizontal binarizing means is used in the horizontal direction.
From "second value" to "first value" or "from first value"
A horizontal change point detecting means for detecting a pixel which changes to a second value as a change point, and two pixels of a pixel adjacent to the pixel of interest in the brightness and vertical direction from the A / D converted image signal. A binary image signal of a "first value" or a "second value" is generated for a pixel of interest based on the value image signal. When the brightness of the pixel of interest is larger than the first set value, the binary image signal of the pixel of interest is set to the “first value” when the “second value” representing that The binary image signal of the matching pixel indicates that the brightness is high.
When the brightness of the pixel of interest is smaller than the second set value, the binary image signal of the pixel of interest is set to the “second value”. In other cases, the pixel of interest is Vertical binarizing means for making the binary image signal the same as an adjacent pixel, and the binary image signal obtained by the vertical binarizing means is converted from a "second value" to a "first value" in the vertical direction. Vertical change point detecting means for detecting, as a change point, a pixel which changes from the "value" or the "first value" to the "second value", wherein the counting means includes the horizontal change point detecting means and At least one of the vertical change point detection means may count the number of pixels detected as a change point.

[0020] Optionally, the change point detecting means may determine, from the A / D-converted image signal, the brightness of a pixel of interest and a binary image signal of a pixel adjacent to the pixel in the horizontal direction. The “first value” or “second value” for the pixel of interest
In which the binary image signal of an adjacent pixel has low brightness.
When the brightness of the pixel of interest is greater than the first set value, the binary image signal of the pixel of interest is set to “first value”, and the binary image signal of the adjacent pixel is When the brightness of the target pixel is smaller than the second set value, the binary image signal of the target pixel is set to the "second value" when the "first value" indicating that the brightness is large is taken. In cases other than the above, the horizontal binarizing means for making the binary image signal of the pixel of interest the same as the adjacent pixels, and the binary image signal obtained by the horizontal binarizing means is used in the horizontal direction.
From "second value" to "first value" or "from first value"
Horizontal change point detection means for detecting a pixel changing to a second value as a change point, wherein the counting means counts the number of pixels detected as a change point by the horizontal change point detection means. You can do it.

Optionally, the average value calculating means includes
The average value of brightness is calculated for each horizontal line in the A / D-converted image signal, and the change point detecting means calculates the brightness of a pixel of interest from the A / D-converted image signal. And generating a “first value” or “second value” binary image signal for a pixel of interest based on a binary image signal of a pixel adjacent to the pixel in the horizontal direction. When the binary image signal of the adjacent pixel takes a “second value” indicating that the brightness is low, the brightness of the pixel of interest is changed with respect to the horizontal line in the vicinity of the processing or in the vicinity thereof. If the value is larger than the first set value calculated in step S2, the binary image signal of the pixel of interest is set to the “first value”, and the binary image signal of the adjacent pixel is “1” indicating that the brightness is large. Value, the brightness of the pixel of interest is being processed Alternatively, when the value is smaller than the second set value calculated for the horizontal line in the vicinity thereof, the binary image signal of the pixel of interest is set to “second value”. Horizontal binarization means for making a binary image signal of a pixel the same as an adjacent pixel; and a binary image signal obtained by the horizontal binarization means, wherein the binary image signal is converted from a "second value" to a "first" And a horizontal change point detecting means for detecting, as a change point, a pixel which changes from the "value" or the "first value" to the "second value", and the counting means includes the horizontal change point detecting means. And the number of pixels detected as a change point can be counted.

[0022]

Embodiments of the present invention will be described below with reference to the drawings. (First Embodiment) FIG. 1 is a block diagram showing a first embodiment of a particle size measuring apparatus 12 according to the present invention. In the figure, reference numeral 10 denotes a granular specimen, which is contained in a case in which many particles are overlapped.

The particle size measuring device 12 for measuring the particle size of the granular object 10 mainly includes a CCD camera 14 (imaging means), an A / D converter 16, an image average value calculating section 40, and a set value calculating section. 42, a change point detection unit 17, a counting unit 30, and a particle size calculation unit 32. The change point detection unit 17 further includes a frame memory 24, a horizontal binarization unit 33, a horizontal change point detection unit 34, A vertical binarizing unit 35 and a vertical change point detecting unit 36 are provided. Among them, the image average value calculation unit 40,
The set value calculation unit 42, the change point detection unit 17, the counting unit 30, and the particle size calculation unit 32 include, for example, a logic circuit, a memory, and a CP.
U, a dedicated arithmetic LSI, or the like. For example, when a CPU and a memory are used, it can be configured by a function realized by a program stored in the CPU or the memory in advance.

The CCD camera 14 captures an image of the granular object 10 and outputs an image signal, for example, an NTSC signal. FIG. 2 shows an example of an image of the granular object 10 captured by the CCD camera 14.

The NTSC image signal is supplied to the A / D converter 1
6 and converted into a digital signal representing, for example, 256 levels of brightness, and then subjected to processing in the image average value calculation unit 40 and the change point detection unit 17, respectively.

The image average value calculating section 40 includes an A / D converter 1
The average value of the brightness of the entire image is calculated from the image of the granular object 10 converted into the digital signal in step 6. Specifically, assuming that the horizontal coordinate value is i, the vertical coordinate value is j, and the brightness of the pixel (i, j) is f (i, j), the average brightness fave is

[0027]

Fave = ｆij f (i, j) / N Here, Σij represents the sum of all pixels, and N represents the total number of pixels.

The set value calculating section 42 includes an image average value calculating section 4
Based on the average value fave calculated at 0, a set of a first set value equal to or larger than the average value and a second set value equal to or smaller than the average value is calculated. As the plural sets, for example,
The first set (fave + β1, fave−β1) and the second set (fave + β1, fave−β1)
ave + β2, fave−β2) are calculated. Then, the difference β1 between the average value fave of the first set and the set value is assumed to be a value effective for removing a minute change in brightness, which is noise, and the difference β1 between the average value fave of the second set and the set value is set. The difference β2 is β
It is good to set it to a value smaller than 1. Β2 can also be set to 0.

The frame memory 24 of the change point detector 17 to which the image signal is sent from the A / D converter 16 stores
The D-converter 16 stores the image of the granular object 10 converted into a digital signal. The frame memory 24 stores image signals sequentially output from the A / D converter 16, and an image average value calculation unit 40 calculates an average value of brightness from the brightness of all pixels, and a set value calculation unit 42. Sets of a plurality of set values are calculated, and based on the set values of each set, the horizontal binarization unit 33 and the vertical binarization unit 34 perform a binarization process to perform a binarization process. The image signal is stored so that reading from the vertical binarization unit 34 can be performed a plurality of times.

The horizontal binarizing section 33 includes a frame memory 24
Generates a binary image in which the value of the pixel of interest is “1” or “0” based on the brightness of the pixel of interest and the brightness of the pixel adjacent to the pixel in the horizontal direction from the digital image signal stored in Specifically, when the binary image signal of the immediately preceding pixel adjacent in the horizontal direction is “0” indicating that the brightness is small, the brightness of the pixel of interest is changed to the first brightness.
When the binary image signal of the pixel of interest is set to “1” and the binary image signal of the immediately preceding pixel takes a value of “1” indicating high brightness, When the brightness of the pixel of interest is smaller than the second set value, the binary image signal of the pixel of interest is set to "0", otherwise, the binary image signal of the adjacent pixel is set to the same binary image signal. is there. This binarization process is performed by setting each set value (fave + β1,
fave-β1) and (fave + β2, fave-β2).

FIG. 3 shows a specific example of this, and a black circle (1)
To (10) correspond to each pixel. Note that the binary image signal of the first pixel in the horizontal direction is “1” when the brightness of the pixel is larger than the average value fave, and “1” when the brightness is smaller than the average value fave.
0 ". Therefore, pixel (1) is set to"
Assuming a value of 1 ", the pixel (2) is smaller than the average value fave, but not smaller than the second set value fave-β1, so that the same binary image signal" 1 "as the pixel (1). Similarly, the pixels (3) to (9) have the same binary image signal “1” as that of the immediately preceding pixel.In the pixel (10), the brightness is the second set value fave−. Because it is smaller than β1,
The binary image signal is "0".

The horizontal change point detecting section 34 is provided with a horizontal binarizing section 3
3 in the horizontal direction.
A pixel which changes from "0" to "1" or "1" to "0" is detected as a change point.In the example of Fig. 3, the pixel (10) is a change point.

Similarly, the vertical binarizing section 35 calculates a pixel of interest from the digital image signal stored in the frame memory 24 based on the brightness of the pixel of interest and the brightness of a pixel adjacent to the pixel in the vertical direction. To generate a binary image with the value of “1” or “0”. Specifically, it indicates that the binary image signal of the immediately preceding pixel adjacent in the vertical direction has low brightness. When the brightness of the target pixel is larger than the first set value when the value is “0”, the binary image signal of the target pixel is set to “1”, and the binary image signal of the immediately preceding pixel is set. Takes a value of “1” indicating that the brightness is high, the brightness of the pixel of interest becomes the second set value (fave−
If the value is smaller than β), the binary image signal of the pixel of interest is set to “0”; otherwise, the binary image signal is set to the same binary image signal as that of the adjacent pixel. This binarization processing is performed by setting each set value (fave + β1, fave−β1), (fave + β
2, fave-β2).

The vertical change point detecting unit 36 determines whether the binary image signal obtained by the vertical binarizing unit 35 changes from "0" to "1" or from "1" to "0" in the vertical direction. Is detected as a change point.

Next, the counting section 30 includes the horizontal change point detecting section 3.
4 and the number of pixels detected as a change point in at least one of the vertical change point detection unit 36 are counted. Granular subject 10
Since the brightness pattern of the image of (1) is finer as the particle size of each particle is smaller and has a higher frequency component, the number of times that the brightness goes above and below the first set value and the second set value is also increased. Therefore, the counting unit 30 counts the number of times of this change as a change point of the binary image.

The particle size calculation unit 32 calculates the average particle size of the particles of the granular test object 10 based on the number of pixels counted by the counting unit 30 for each of the plurality of set value sets. is there.

In general, the larger the particle size of each particle of the granular object 10, the coarser the brightness pattern of the image becomes.
The number of pixels counted at 0 decreases, and conversely, the granular object 1
The smaller the particle diameter of each particle of 0, the larger the number of pixels counted by the counting unit 30, and the particle diameter becomes a decreasing function of the number of pixels. Here, if the particles are arranged without gaps and without overlapping, it is clear that the frequency of the lightness pattern is proportional to the number of particles and inversely proportional to the particle size of the particles. In the case of an actual granular specimen, although the particles overlap, it can be expected that the inverse relationship between the particle size and the particle size is the same, and that the inverse relationship is actually statistically established. Has been obtained.

Accordingly, for the numbers of pixels n1 and n2 obtained using the first and second sets of the first and second set values, respectively, the above-mentioned equation (1) is used. Are used to determine the average particle diameters r1 and r2 as follows.

[0039]

R1 = k1 / n1 r2 = k2 / n2 k1 and k2 are considered to be constant constants regardless of the size of the particle size if optical conditions such as the magnification of the CCD camera 14 are predetermined. Therefore, each set (fave + β1, fave−β1), (fav +) of the first set value and the second set value
e + β2, fave−β2), the known particle sizes rs1,
The granular object having rs2 is imaged by the CCD camera 14,
By performing the same series of processing as described above to determine the number of pixels ns1 and ns2 counted by the counting unit 30, k1,
k2 can be predetermined. That is, k1, k2
Is

[0040]

## EQU4 ## Since k1 = rs1 · ns1 and k2 = rs2 · ns2, k1 and k2 can be stored in the particle size calculator 32 in advance and used for calculating the particle size.

Then, the particle diameter calculation unit 32 calculates the particle diameters r1 and r2 using the set of each of the first set value and the second set value, and then calculates the particle diameter threshold value r0 in FIG. On the other hand, when r2 <r0, r2 is adopted and output as a particle size, and when r2 ≧ r0, r1 is adopted and output as a particle size.

In this way, even if the particles overlap, the average particle size can be obtained by a simple process, and the calculation cost can be reduced. Then, by calculating the particle size based on the number of pixels counted using the set value of the optimal set value set from a plurality of set value set values, the influence of minute brightness change unrelated to the particle size measurement is obtained. By counting only the pixels that reflect the frequency of the lightness pattern that is considered to contribute to the particle size by removing the particle size, more accurate particle size measurement can be performed, and the influence of the lightness change when the particle size is small is reduced. It is possible to prevent a situation in which the measurement is inaccurate due to the removal.

The first set value and the second set value are the average value fave of the brightness of the entire image obtained by the image average value calculation unit 40.
Since the binarization process is performed by the horizontal binarization unit 33 and the vertical binarization unit 35 using these values as thresholds, even if the illumination changes and the brightness fluctuates, Since the set value also changes in proportion to, accurate measurement can be performed without being affected by variations in brightness.

(Second Embodiment) The particle size measuring apparatus of the present invention may be configured such that the vertical binarization section 35 and the vertical change point detection section 36 are removed from the configuration of FIG. That is, the change point detection unit 17 includes the frame memory 24, the horizontal binarization unit 33, and the horizontal change point detection unit 34. Counting unit 30
Counts the number of pixels detected as a change point by the horizontal change point detection unit 34. Since the image of the granular object 10 has no directionality, only the pixels corresponding to the change points in the horizontal direction of each particle are counted, and the particle diameter can be obtained based on the number of pixels. Although the number of pixels counted by the counting unit 30 is smaller than the number of pixels counted in the first embodiment, the same holds true for the number of pixels counted being inversely proportional to the particle diameter of particles. Therefore, taking this into account in the particle size calculation unit 32, if r is the average particle size and n ′ is the number of pixels,

[0045]

R1 = k1 ′ / n1 ′ r2 = k2 ′ / n2 ′ The average particle diameters r1 and r2 are obtained from the relationship. Here, k1 ′ and k2 ′ are constants, and k1 ′ <k1, k
2 ′ <k2. For k1 'and k2', CCD
If the optical conditions such as the magnification of the camera 14 are determined in advance, the constants are considered to be constant irrespective of the size of the particle size. Therefore, each set (f) including the first set value and the second set value
ave + β1, fave−β1), (fave + β2, fave−β)
By using 2), the granular object having the known particle diameters rs1 and rs2 in the range in which the accurate particle diameter is required is imaged by the CCD camera 14, and the same series of processing as described above is performed. Counted number of pixels ns1 ', n
By obtaining s2 ', k1' and k2 'can be determined in advance.

Then, the particle size calculation unit 32 uses the set of the first set value and the second set value to set the particle size r.
After obtaining r 1 and r 2, when r 2 <r 0, r 2 is adopted and output as the particle size, and when r 2 ≧ r 0, r 1 is adopted and the particle size is determined with respect to the particle size threshold r 0 in FIG. Is the same as in the first embodiment.

As described above, by limiting the direction in which the change point is detected to only one direction, the processing can be simplified and the calculation cost can be further reduced. In addition, the CCD was limited to the horizontal direction instead of the vertical direction.
This is because the image signal sent from the camera 14 is generally a signal that is continuous in the horizontal direction, and the horizontal change point detection unit 34 that has received this signal can perform sequential processing, and further processing can be performed. Simplification can be achieved.

(Third Embodiment) A third embodiment will be described with reference to FIG. The particle size measuring device 12 mainly includes a CCD camera 14 (imaging means), an A / D converter 1
6, a horizontal average value calculation unit 44, a change point detection unit 17, a counting unit 30, and a particle size calculation unit 32.
Reference numeral 7 includes a line memory 26, a horizontal binarizing unit 33, and a horizontal change point detecting unit 34. Horizontal average value calculation unit 44
Calculates the average value of the brightness of each line in the horizontal direction of the image signal of the granular object 10 converted to a digital signal by the A / D converter 16. That is, assuming that the horizontal coordinate value is i, the vertical coordinate value is j, and the brightness of the pixel (i, j) is f (i, j), the average value fave (j) of each horizontal line is given. )

[0049]

Calculated by fave (j) = Σif (i, j) / Nx. Here, Σi represents the sum of all the pixels in one horizontal line, and Nx represents the number of pixels in one horizontal line.

The set value calculation unit 42 includes an image average value calculation unit 4
Based on the average value fave (j) for each line calculated at 0, a set of a first set value equal to or larger than the average value and a second set value equal to or smaller than the average value is calculated.

The line memory 26 stores one horizontal line of the image signal of the granular object 10 converted into a digital signal by the A / D converter 16.
The line memory 26 stores the image signals sequentially output from the A / D converter 16 and stores the horizontal average value calculation unit 4.
In step 4, an average value of the brightness is calculated from the brightness of one line in the horizontal direction, a set of a plurality of set values is calculated in the set value calculating unit 42, and the next horizontal binarizing unit is calculated based on the set values of each set. The image signal is stored so that reading from the horizontal binarization unit 33 can be performed a plurality of times in order to perform the binarization processing at 33.

The horizontal binarizing section 33 performs the same operation as the first embodiment from the digital image signal stored in the line memory 26.

The processing in the horizontal change point detecting section 34, the counting section 30, and the particle size calculating section 32 is the same as that in the second embodiment.

As described above, the processing limited to the horizontal direction including the calculation of the average value of the brightness can further simplify the processing.

In each of the above embodiments, the horizontal binarization unit 33 and the vertical binarization unit 35 of the change point detection unit 17 perform the frame memory multiple times in accordance with the number of set value sets. 24
Alternatively, reading is performed on the line memory 26, but the present invention is not limited to this.
7, a plurality of systems including a horizontal binarization unit 33, a horizontal change point detection unit 34, a vertical binarization unit 35, a vertical change point detection unit 36, and a counting unit 30 corresponding to the number of set value sets, Parallel processing can also be performed. In this case, in the third embodiment, except for the line memory 26, the average value fave of the brightness of each line is calculated as the average value in the horizontal direction, for example, one line before the line for which the change point is being detected. fave
(J-1), and the first set value and the second set value can be determined based on the average value fave (j-1) in the horizontal direction one line before. Since the average value of lightness is not expected to change significantly from line to line,
There is no problem in practical use, and the calculation cost can be further reduced. Alternatively, when the subject can be imaged a plurality of times, the CCD camera 14 can perform the imaging a plurality of times with one processing system regardless of the presence or absence of the line memory 26.

[0056]

As described above, according to the first aspect of the present invention,
According to the invention as set forth in any one of (8) to (8), in the predetermined direction, the brightness of the image signal of the granular object crosses the set value selected from the first set value and the second set value according to a predetermined rule. The pixel at the time of change is detected as a change point, the number of pixels at the change point is counted, and the average particle size is determined based on the number of pixels. The present invention can also be applied to granular subjects that overlap without being overlapped. Further, since no repetitive operation or the like is required, the calculation cost can be reduced.

By calculating the particle size based on the number of pixels counted from a plurality of sets of set values using the set value of the optimum set of set values, a minute change in lightness unrelated to the particle size measurement is obtained. By counting only the pixels that reflect the frequency of the brightness pattern that is considered to contribute to the particle size by eliminating the effect, more accurate particle size measurement can be performed, and the effect of brightness change when the particle size is small Can be prevented from being inaccurately measured due to the removal of.

Further, in order to detect, as a change point, a pixel that changes over a specific one of a plurality of set values set based on the average value of lightness, a parameter setting of the change point detection reference is performed. Can be eliminated, the adjustment can be made easily, and it is possible to avoid the influence of the fluctuation of the brightness due to the change of the illumination.

Further, according to the third, fourth, seventh and eighth aspects of the present invention, generally, a picked-up image signal is continuous in the horizontal direction. The processing can be further simplified and speeded up. In particular, according to the fourth and eighth aspects of the present invention, since the average value of lightness is obtained for each horizontal line,
The calculation cost can be further reduced as compared with the case where the average brightness of the entire image is obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a first embodiment of a particle size measuring apparatus according to the present invention.

FIG. 2 is an explanatory diagram showing an example of an image of a granular object to be imaged.

FIG. 3 is a graph showing a brightness pattern on a certain horizontal line L;

FIG. 4 is a block diagram showing a third embodiment of the particle size measuring device according to the present invention.

FIG. 5 shows the number of pixels n and the particle size r at a change point due to a difference in set values.
6 is a graph schematically showing the relationship with.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 Granular object 12 Particle size measuring device 14 CCD camera (imaging means) 16 A / D converter 17 Change point detection part (change point detection means) 20 Horizontal change point detection part (horizontal change point detection means) 22 Vertical change check Output part (vertical change point detecting means) 30 Counting part (counting means) 32 Particle size calculating part (particle size calculating means) 33 Horizontal binarizing part (horizontal binarizing means) 34 Horizontal changing point detecting part (horizontal change check) Output means) 35 vertical binarization section (vertical binarization means) 36 vertical change point detection section (vertical change point detection means) 40 image average value calculation section (average value calculation means) 44 horizontal average value calculation section (average value) Calculation means)

Continued on front page F term (reference) 2F065 AA26 BB05 CC00 DD00 DD06 FF04 JJ03 JJ26 NN19 QQ03 QQ06 QQ08 QQ24 QQ25 QQ26 QQ27 QQ28 QQ42 QQ51 UU05 5B057 AA07 AA10 BA02 CA08 CB06 DA13 DB08 DC02 DC16

Claims (8)

[Claims] 1. A particle size measuring device for measuring a particle size of a granular object, comprising: an image pickup means for picking up an image of the granular object and outputting an image signal; and A / D converting the image signal from the image pickup means. An A / D converter, an average value calculating means for calculating an average value of brightness of the A / D-converted image signal, and an average value based on the average value obtained by the average value calculating means. A plurality of sets in which at least one of the first set value and the second set value is different from each other is calculated in a set including the first set value described above and the second set value equal to or less than the average value. Setting value calculating means for selecting, from the A / D-converted image signal, a brightness in a predetermined direction from the first set value and the second set value of each set in accordance with a predetermined rule. When the pixel changes over the set value, the pixel is detected as a change point. Change point detecting means, counting means for counting the number of pixels detected as change points by the change point detecting means, and, for each of the plurality of sets, based on the number of pixels counted by the counting means. And a particle size calculating means for calculating the particle size of the granular object. 2. The method according to claim 1, wherein the change point detecting unit is configured to determine a pixel of interest based on the brightness of the pixel of interest and a binary image signal of a pixel adjacent to the pixel in the horizontal direction from the A / D converted image signal. To generate a binary image signal of “first value” or “second value”, and the binary image signal of an adjacent pixel indicates “second value” indicating that brightness is small. When the brightness of the pixel of interest is greater than the first set value, the binary image signal of the pixel of interest is set to “first value”, and the binary image signal of the adjacent pixel is When the brightness of the target pixel is smaller than the second set value, the binary image signal of the target pixel is set to the “second value” when the “first value” indicating that the brightness is large is set, In cases other than the above, a horizontal binarization method in which the binary image signal of the pixel of interest is the same as that of the adjacent pixel And the binary image signal obtained by the horizontal binarizing means is converted from a “second value” to a “first value” or “first value” in the horizontal direction.
Horizontal change point detection means for detecting a pixel that changes from a first value to a second value as a change point; and detecting, from the A / D-converted image signal, a brightness and a vertical direction of a pixel of interest. Based on the binary image signal of the pixel adjacent to the pixel, the “first value” or “
It is to generate a binary image signal of the second value "indicating that the binary image signal of the adjacent pixel has low brightness."
When the brightness of the target pixel is larger than the first set value when the second value is taken, the binary image signal of the target pixel is set to the “first value”, and the binary image of the adjacent pixel is set. The signal
When the brightness of the target pixel is smaller than the second set value, the binary image signal of the target pixel is set to the “second value” when the “first value” indicating that the brightness is large is set, In cases other than the above, a vertical binarizing unit that makes the binary image signal of the pixel of interest the same as an adjacent pixel, and the binary image signal obtained by the vertical binarizing unit 2 value "to" first value "or"
Vertical change point detecting means for detecting a pixel changing from a first value to a second value as a change point, wherein the counting means includes the horizontal change point detecting means and the vertical change check. The particle size measuring device according to claim 1, wherein the number of pixels detected as a change point is counted in at least one of the output means. 3. The change point detecting means, based on the brightness of a pixel of interest and a binary image signal of a pixel adjacent to the pixel in the horizontal direction, from the A / D-converted image signal. To generate a binary image signal of a “first value” or a “second value”, wherein the binary image signal of an adjacent pixel is a “second value” indicating that the brightness is low. When the brightness of the pixel of interest is greater than the first set value, the binary image signal of the pixel of interest is set to “first value”, and the binary image signal of the adjacent pixel is When the brightness of the pixel of interest is smaller than the second set value, the binary image signal of the pixel of interest is set to the "second value". In other cases, the horizontal binarization unit sets the binary image signal of the pixel of interest to be the same as the adjacent pixel. The binary image signal obtained by the horizontal binarization means is changed from a "second value" to a "first value" or "first value" in the horizontal direction.
Horizontal change point detection means for detecting a pixel that changes from the first value to the second value as a change point, wherein the counting means detects the change point by the horizontal change point detection means. 2. The particle size measuring device according to claim 1, wherein the number of pixels is counted. 4. The average value calculating means calculates an average value of lightness for each horizontal line in the A / D-converted image signal, and the change point detecting means comprises: From the converted image signal, a “first value” or “1” is assigned to the pixel of interest based on the brightness of the pixel of interest and the binary image signal of a pixel adjacent to the pixel in the horizontal direction.
It is to generate a binary image signal of the second value "indicating that the binary image signal of the adjacent pixel has low brightness."
When the brightness of the pixel of interest is greater than the first set value calculated for the horizontal line in the vicinity of processing or when the value of the pixel of interest is taken as the "second value", When the binary image signal is set to the “first value” and the binary image signals of the adjacent pixels take the “first value” indicating that the brightness is large, the brightness of the pixel of interest is processed. If the value is smaller than the second set value calculated for the horizontal line near or in the vicinity thereof, the binary image signal of the pixel of interest is set to the “second value”. Horizontal binarization means for making a binary image signal of a pixel to be the same as an adjacent pixel; and a binary image signal obtained by the horizontal binarization means, from a "second value" to a "second value" in the horizontal direction. Value of 1 "or"
Horizontal change point detection means for detecting a pixel that changes from the first value to the second value as a change point, wherein the counting means detects the change point by the horizontal change point detection means. 2. The particle size measuring device according to claim 1, wherein the number of pixels is counted. 5. A particle size measuring method for measuring a particle size of a granular object, comprising: imaging a granular object; A / D converting an image signal; and measuring a brightness of the A / D converted image signal. Calculating an average value, based on the calculated average value, a first set value that is equal to or greater than the average value, and a second set value that is equal to or less than the average value; A plurality of sets in which at least one of the value and the second set value are different from each other are calculated. From the A / D-converted image signal, the first set value and the second set value of each set are set in a predetermined direction in the brightness direction. Among the set values of 2, the pixel is detected as a change point when changing over a set value selected according to a predetermined rule, the number of pixels detected as the change point is counted, and For each set of pixels, based on the number of pixels counted as And calculating the particle size of the granular specimen. 6. The detection of the change point is performed based on the brightness of a pixel of interest and a binary image signal of a pixel adjacent to the pixel in the horizontal direction from the A / D-converted image signal. Represents that the brightness is low.
When the brightness of the pixel of interest is greater than the first set value, the binary image signal of the pixel of interest is set to “first value”, and the binary image signal of the adjacent pixel is When the brightness of the target pixel is smaller than the second set value, the binary image signal of the target pixel is set to the "second value" when the "first value" indicating that the brightness is large is taken. In other cases, the binary image signal of the pixel of interest is subjected to horizontal binarization to be the same as an adjacent pixel, and the binary image signal obtained by the horizontal binarization is horizontal. A pixel that changes from the “second value” to the “first value” or the “first value” to the “second value” in the direction is detected as a horizontal change point, and is subjected to the A / D conversion. From the image signal, the brightness of the pixel of interest and the binary image signal of the pixel adjacent to the pixel in the vertical direction are used. When the binary image signal of an adjacent pixel takes a “second value” indicating that the brightness is low, if the brightness of the pixel of interest is larger than the first set value, the binary value of the pixel of interest is reduced. When the value image signal is “first value” and the binary image signal of the adjacent pixel takes “first value” indicating that the brightness is large, the brightness of the pixel of interest is changed to the second set value. If smaller, the binary image signal of the pixel of interest is set to the “second value”; otherwise, the binary image signal of the pixel of interest is the same as the adjacent pixel. And the binary image signal obtained by the vertical binarization is converted from the “second value” to the “first value” or the “first value” to the “second value” in the vertical direction. Is detected as a vertical change point, and the number of pixels detected as a change point is counted. Serial horizontal and counts the number of detected pixels as change point at least one of the vertical direction, the particle diameter measuring method according to claim 5, wherein a. 7. The method according to claim 7, wherein the detecting of the change point comprises detecting, based on the brightness of the pixel of interest and a binary image signal of a pixel adjacent to the pixel in the horizontal direction, from the A / D-converted image signal. Takes a “second value” indicating that the brightness is small, and if the brightness of the pixel of interest is greater than the first set value, the binary image signal of the pixel of interest is changed to “ When the binary image signal of an adjacent pixel takes a “first value” indicating that the brightness is large, when the brightness of the pixel of interest is smaller than the second set value, The binary image signal of the pixel of interest is set to a “second value”, and in other cases, the binary image signal of the pixel of interest is subjected to horizontal binarization to be the same as an adjacent pixel. The binary image signal obtained by the horizontal binarization is "second value" in the horizontal direction. Et "first value", or "first value" to detect a pixel which changes to "second value" as the change point, the particle diameter measuring method according to claim 5, wherein a. 8. The calculation of the average value of the brightness of the A / D-converted image signal is to calculate the average value of the brightness of each horizontal line in the A / D-converted image signal. The change point is detected based on the brightness of a pixel of interest and a binary image signal of a pixel adjacent to the pixel in the horizontal direction from the A / D-converted image signal. Takes a “second value” indicating that the brightness is small, the brightness of the pixel of interest is the first set value calculated for the horizontal line that is being processed or in the vicinity thereof When the binary image signal of the pixel of interest is larger than “1”, the binary image signal of the pixel of interest takes the “first value” indicating that the brightness is large. The brightness of the pixel to be processed is in the horizontal direction If the value is smaller than the second set value calculated for the line, the binary image signal of the pixel of interest is set to the “second value”; otherwise, the binary image signal of the pixel of interest is set to the next value. The binarization in the horizontal direction is performed so as to be the same as the matching pixel. The binary image signal obtained by the binarization in the horizontal direction is converted from the “second value” to the “first value” or “1” in the horizontal direction. 6. The particle size measuring method according to claim 5, wherein a pixel that changes from the first value to the second value is detected as a change point.