JPH02129765A - Filamentous microorganism measuring instrument - Google Patents

Abstract

PURPOSE:To accurately measure the amount of filamentous funguses by applying a spatial filter for emphasizing various directional lines to image data and extracting filamentous funguses from a difference between the density level of the center part of a spatial filter output and that of both the side parts. CONSTITUTION:The spatial filter 21 for emphasizing various directional lines is applied to image data obtained by photographing liquid Sa containing filamentous microorganisms and filamentous microorganisms are extracted from the difference between the density level of the center part of an output from the filter 21 and that of both the side parts by a linear shape extracting means 22. Thus, the spatial filter for emphasizing various directional lines can be applied to image data obtained by photographing mud and filamentous microorganisms can be extracted from the difference between the density level of the center part of the spatial filter output and that of both the side parts.

Description

[Detailed description of the invention] [Object of the invention] (Industrial application field) The present invention is directed to the treatment of filamentous microorganisms (hereinafter referred to as filamentous fungi) contained in activated sludge in sewage treatment (prior art). In the activated sludge treatment process, the microbial flora of activated sludge is one of the indicators of the sludge treatment status. In order to operate sewage treatment effectively, it is necessary to maintain good settling properties of sludge.

In order for sludge to settle well, it is necessary for filamentous bacteria to propagate appropriately. By the way, filamentous fungi usually gather together to form a floc 1 as shown in FIG. Therefore, when the amount of filamentous fungi decreases, flocs cannot grow large, and this condition is called pinpoint floc.

In this state, the settling properties of the sludge deteriorate.

On the other hand, if the amount of filamentous fungi increases too much, a bulking state occurs in which a large amount of filamentous fungi appears outside the floc.

Even in this state, the settling properties of the sludge deteriorate. However, since sludge with poor settling properties flows out without settling in the settling tank, the quality of treated water deteriorates. Therefore, in order to maintain sewage treatment in good condition, it is necessary to monitor the condition of activated sludge and periodically measure the amount of filamentous bacteria.

Incidentally, activated sludge is monitored by taking a sample of the sludge on a slide, attaching it to a microscope, and visualizing it by an operator. For this reason, accurate monitoring could only be performed by a skilled operator with knowledge of microorganisms and extensive experience. Further, even for such an experienced person, a long time of microscopy is required.

Therefore, a method has been proposed, for example, as shown in Japanese Patent Application Laid-open No. 62-53792, in which sludge is imaged and filamentous bacteria and flocs are separated and extracted by image processing and then measured. In this separation and extraction, the image data obtained by imaging the sludge is applied with each of the 3x3 filters F1 to F12 shown in Figure 17 to emphasize the linear parts, that is, filamentous bacteria, and then binarized. The process is performed to obtain image data containing filamentous fungi and flocs, and then this image data is reduced and enlarged to separate the filamentous fungi and flocs. Note that in each of the filters F1 to F12, rWJ indicates a weight. For example, the filter Fl emphasizes vertical lines, and the filter F2 emphasizes horizontal lines.

However, in the above method, the measurement accuracy is greatly influenced by the image quality of the image data. To explain the reason for this, the density level when filamentous fungi are detected is as shown in Figure 18 (a).
There is a pulse-like change as shown in (b) of the same figure, and the density level when the edge of the floc is detected becomes an edge change as shown in (b) of the same figure. There are many edges with large changes in density level in the peripheral area of the flock. However, when the filters Fl-F12 are applied, the filamentous fungi and the flocs are emphasized without distinction, so that the filamentous fungi and the edges are detected in the same way.

For this reason, it becomes difficult to separate filamentous fungi and flocs, and it becomes impossible to accurately measure filamentous fungi on image data with uneven illumination.

(Problems to be Solved by the Invention) As described above, when each of the filters Fl-F12 was used, filamentous bacteria and floc edges could not be detected without distinction, and filamentous bacteria could not be accurately measured.

Therefore, an object of the present invention is to provide a filamentous microorganism measuring device that can accurately measure filamentous bacteria.

[Structure of the Invention] (Means for Solving the Problem) The present invention obtains second image data in which filamentous microorganisms are extracted from first image data obtained by imaging a liquid containing filamentous microorganisms; In a filamentous microorganism measuring device that measures filamentous microorganisms by removing flocs formed by filamentous microorganisms from this second image data, each spatial filter that emphasizes lines in each direction is applied to the first image data. This is a filamentous microorganism measuring device which attempts to achieve the above object by providing a linear extraction means for extracting filamentous microorganisms from the difference between the density level of the central part and the density level of both sides in these spatial filter outputs.

(Function) By having such a means, a spatial filter is applied to the image data obtained by imaging the sludge, emphasizing lines in each direction, and the gray level in the center and both sides in the output of this spatial filter are Filamentous microorganisms are extracted from the difference in the density level of the part.

(Example) Hereinafter, an example of the present invention will be described with reference to the drawings.

FIG. 1 is a configuration diagram of a filamentous microorganism measuring device.

In the figure, reference numeral 10 denotes an ITV (industrial television) camera, and an image signal output from the ITV camera 10 is converted into a digital image signal by an A/D (analog/digital) converter 11 and stored in an image memory 12. is stored as image data. by the way,
Connected to this image memory 12 are a line shape extraction means 20, a flock center extraction means 30, and a flock periphery extraction means 40. The line shape extracting means 20 applies each spatial filter to emphasize the lines in each direction to the image data stored in the image memory 12, and calculates the gray level of the central part and the gray level of both sides in the output of these spatial filters. It has a function of extracting filamentous microorganisms from the difference, and has the functions of a line emphasis spatial filter 21 and a difference filter 22 in each direction. As the line emphasis spatial filter 21, for example, one having a filter constant as shown in FIG. 2 is used. Note that the spatial filter shown in the figure emphasizes lines in the vertical direction, and explanations of spatial filters in other directions will be omitted. or,
To explain the spatial filter shown in FIG. 2, the differential filter 22 has a filter constant of "1" at the pixel position (x-3).
The difference between the shading level of each point where 1 exists and the shading level of the corresponding point at pixel position (x+3) is determined, and the sum of the values of these differences is determined. Then, linear filamentous fungi are determined from the calculation results of the differential filter 22. In other words, to explain the method for determining filamentous fungi, the edges of filamentous fungi and flocs change in density level as shown in FIG. here,
When paying attention to each position of al and a3 and calculating the difference, if 1al-a31-0, it is determined to be a filamentous fungus, and if at-a31>0, it is determined to be an edge. Moreover, this linear shape extraction means 2
0 is provided with a spatial filter according to the valley width in order to extract filamentous fungi of the valley width.

The floc center extraction means 30 has a function of extracting the center portion of the floc, and has the functions of a slicing section 31 and a line shape removing section 32.

The slicing unit 31 performs binarization processing on image data using a predetermined threshold value to obtain image data that is greater than or equal to this threshold value, and the linear shape removing unit 32 is a linear shape extracting unit from the sliced image data obtained by the slicing unit 31. It has a function of subtracting the image data obtained in step 20 to obtain flock center image data in which only the center portion of the flock remains.

The floc peripheral extraction means 40 has a function of extracting the peripheral part of the floc, and includes a linear smoothing filter 41,
It has the functions of a wide-area spatial filter 42 and a slice section 43. The linear smoothing filter 41 has a function of smoothing the filamentous fungus portion in the image data, and the wide area spatial filter 42 has a 9×9.7×7 and 3×3 filter as shown in FIGS. 4 to 6. It has a function of using each spatial filter and synthesizing each filter output of each of these spatial filters. Further, the slicing unit 43 divides the filter output image data of the wide-area spatial filter 42 into two parts using a predetermined threshold value.
It has a value processing function.

The floc removing unit 50 synthesizes the floc center image data obtained by the floc center extracting means 30 and the floc surrounding image data obtained by the floc surrounding extracting means 40 to obtain floc image data showing only flocs, and performs linear shape extraction. It has a function of subtracting the floc image data from the filamentous fungus image data obtained by the means 20 to obtain final image data of only the filamentous fungi. And filamentous fungus counter 5
1 is for counting filamentous fungi from the final image data obtained by the floc removal section 51 to obtain the amount thereof.

Next, the operation of the apparatus configured as described above will be explained with reference to the thread measurement flowcharts shown in FIGS. 7 to 9.

The ITV camera 10 images the sludge Sa and outputs the image signal. This image signal is converted into a digital image signal by an A/D converter 11 and stored in an image memory 12 as image data. Note that the image of this image data is the 10th image.
As shown in the figure.

When the image data is stored in this manner, the linear shape extraction means 20 extracts filamentous fungi in step s1. Extraction of filamentous fungi is performed using each spatial filter that emphasizes lines in each direction, but for the sake of simplicity, we will explain the case in which the vertical spatial filter shown in FIG. 2 is used and the results shown in FIGS. This will be explained with reference to the flowchart in FIG. That is, the line emphasis spatial filter 21 determines the line shape in step el, but when the spatial filter shown in FIG. 2 is used, the flowchart shown in FIG. 9 is executed. That is, in step f1, the line emphasis spatial filter 21 multiplies the image data stored in the image memory 12 by a spatial filter having a filter constant shown in FIG.
Calculate and find 1-A3.

-2P (x, y+i)+P (x+3.y+i)1≧
C1 -2P (x, y+i) 10P (x+3.y+i)1
≧02-2P (x, y+ i)+P (x+3.y+i)!
≧03 Here, for example, P (x-3, y+i) is the pixel position (
x-3, y+i). And A1 is the sum of the upper half gray levels in the spatial filter,
A2 is the sum of the lower half gradation levels in the spatial filter, and A3 is the sum of A1 and A2. When A1-A3 is obtained in this way, in step r2, a comparison is made to determine whether Al-A3 is larger than each of the constants C1 to c3. By this comparison, each of A1 to A3 becomes each constant C1 to C
If it is larger than 3 and satisfies the condition, the gray level of the pixel position (x, y+i) applied with the spatial filter will be lower than the gray level of the pixel positions (x-3, y+i) and (x+3, y+i) on both sides. It is determined that the Then step r
4, the difference filter 22 calculates and calculates the following B1 to B3, and in the next step r5, the difference filter 22 calculates the following B1 to B3.
and each constant 01 to C3 are compared.

-P (x+3.y+i) l//z≦C4 -P (x+3.y+i) l/A2≦ 05 -P (x+3.y+i) l/A3≦C6 And each 81 to B3 is calculated from each constant 04 to C6 If it is determined that the condition is satisfied by the small value, the pixel positions (x-3, y+i) and (x+3.y
(i) There is no difference in each gray level and the pixel position (x, y
+i, ) and the pixel positions on both sides (x-3, y+i) and (x
+3. It is determined that the difference in each density level from y+i) is large, and it is determined that there is a linear shape. In addition, the judgment was made in half, such as AI + A2, Bl, and B2.
This is because if only A3°B3 is used, the conditions of A 3 + 83 are both satisfied due to point-like noise with a high density level, and the line is incorrectly determined. When it is determined that the shape is linear in this way,
In step e3, it is determined that it is a filamentous fungus.

This completes the extraction of the filamentous fungi existing in the vertical direction. Next, the linear shape extraction means 20 is set at 22.5° with respect to the vertical direction.
Filamentous fungi existing in each direction are extracted using spatial filters that emphasize lines in each direction of 45"...157.5'. As a result, filamentous fungi as shown in Figure 11 were extracted. Filamentous fungi image data is obtained.

However, in the filamentous fungus image data obtained by the above linear shape extraction means 20, some noise appears in the periphery of the floc 1, as shown in FIG. This is because there are many linear portions in areas where the density level is high around the flock.

Therefore, in order to remove this floc 1, the floc is extracted by the floc center extraction means 30 and the floc periphery extraction means 40. First, in step s2, the flock center extraction section 30 reads out the image data stored in the image memory 12 and performs background correction. Therefore, the slope of the background is considered to be a plane, and the image data is divided into four
Divide and find the highest level point of the gray level. The method of finding this highest level point is to divide each of the four divided image data into small regions of 18X1G to eliminate the influence of noise, and calculate the average value of the gray level and its center coordinate for each small region, and then obtain the representative value. shall be. Then, from these representative values, the highest point of the gray level in each of the four divided image data is determined. When the shading levels of the four points are determined in this manner, the slope of the background and its plane are determined using the least squares method using the shading levels of the four points. However, background plane data is subtracted from the image data to perform background correction.

Next, in step S3, the slicing unit 31 performs binarization processing. By the way, the method of determining the threshold value in this binarization process is that when the histogram is divided into two at a certain value,
In this method, a value that maximizes the variance between these two divided globes is set as a threshold value. The method of determining the threshold value will be specifically explained.

When the gray level is 1 to m and the frequency of the gray level is ni, the probability PI for all pixels N and each gray level is Pi-n1/N. Then, the gray level is divided into two levels, Co -fl~kl, C1-(k+1~ml
Then, the probability of occurrence of co is ω. , the probability of occurrence of C1 ω1
Average value of Co μO% CI average value μ, ω] Note 1-C0 μ. −μ (k)/ω (k) μ. −(Rhu μ (k)) ÷ (1−ω (k)). Here, ■ is the average value of all samples. Then, the variance σ2 (k) between each class of gray levels divided into two is C2 (k
)−C0(μ0−μ)2 +ω1 (μm−μ)2 −ω0ω1 (μm−μ0)2 − (μω (k) −μ (k) l 2÷ω
(k) fl−ω (k)). Therefore, the maximum value of this variance σ2 (k) is determined, and K giving this maximum value is set as a threshold value.

However, the processing unit 31 binarizes the image data using this threshold value to obtain flock center image data in which the center of the flock is extracted. Note that this flock center image data contains a small amount of filamentous fungi at the center of the flock. Therefore, in step s4, the line shape removing section 32 receives the floc center image data obtained by the slicing section 31 and the filamentous fungus image data obtained by the line shape extracting means 20, and generates filamentous fungus image data from the floc center image data. pull. As a result, flock center image data of only the center of the flock as shown in FIG. 12 is obtained.

Furthermore, the peripheral part of the floc is obtained at the same time as the central part of the floc is extracted. That is, the linear smoothing filter 41 of the floc periphery extraction means 40 smoothes the filamentous fungus portion in step s5. This is because a spatial filter is used to extract the peripheral part of the floc, but when this spatial filter is applied, there is a risk that the filamentous fungal part will also be detected as the peripheral part of the floc because the density level changes in the filamentous fungal part are large. be. Therefore, since the linear shape extraction means does not detect the fall of the density change, the image portion of the filamentous fungus is enlarged by one pixel and smoothed. The smoothing method is performed using the following equation, and the gray level P (x, y) of the coordinates (x, y) becomes P (x, y). Here, Pt (Ps to P4) is the density level of the part in contact with the filamentous fungus F in four directions as shown in Fig. 13, and di and di are the distances between these four points and the point of interest (x, y). It is. However, the above formula indicates that each gray level P1 is weighted and added by the inverse ratio of the distance. Once the filamentous fungi have been smoothed as described above, the process moves to step s6, where the wide-area spatial filter 42 applies each of the spatial filters shown in FIGS. 4 to 6 to the smoothed image data. Afterwards, these filter outputs are combined. Next, in step s7, the slicing section 43 receives the image data from the wide-area spatial filter 42, and binarizes this image data using the threshold used in the slicing section 31. As a result, flock surrounding image data as shown in FIG. 14 is obtained. However, a small amount of filamentous fungi remains in this image data around the floc. Therefore, in the next step s8, the floc periphery extraction means 40 subtracts the filamentous fungus image data obtained by the linear shape extraction means 20 from the floc periphery image data.

Thereafter, in step s9, the floc removal section 50 receives each image data from the floc center extraction means 30 and the floc periphery means 40, and synthesizes it to create floc image data containing only flocs.

Here, this flock image data is affected by noise and has holes of various sizes inside the flock. Therefore,
In step slo, the flock image data is subjected to minute area removal and closed area removal with a density of "0" to remove noise and holes. Thereafter, the floc removal section 50 subtracts the floc image data from the filamentous fungus image data obtained by the linear shape extraction means 20 to obtain final image data of only the filamentous fungi. Note that at this time, minute areas are removed from the final image data.

As a result, image data of only filamentous fungi as shown in FIG. 15 is obtained. Then, the filamentous fungus counter 51 thins out the filamentous fungus portion in the final image data, and counts pixels having a gray level r255J or higher. As a result, the amount of filamentous fungi can be determined.

In this way, in the above embodiment, a spatial filter is applied to the image data obtained by imaging the sludge to emphasize lines in each direction, and the shading level in the center and the shading level on both sides in the output of this spatial filter are determined. Since the filamentous microorganisms are extracted based on the difference between This makes it possible to measure filamentous fungi with high precision.

Note that the present invention is not limited to the above-mentioned embodiment, and may be modified without departing from the spirit thereof. For example, in the above embodiment, floc center oil is used to remove floc parts from filamentous fungus image data.

Although the pile means and the floc periphery extraction means are used, the flocs may be removed using other means.

[Effects of the Invention] As detailed above, according to the present invention, it is possible to provide a filamentous microorganism measuring device that can accurately measure filamentous bacteria.

[Brief explanation of the drawing]

1 to 15 are diagrams for explaining one embodiment of the filamentous microorganism measuring device according to the present invention, in which FIG. 1 is a configuration diagram, FIG. 2 is a schematic diagram of a line emphasis spatial filter, Fig. 3 is a schematic diagram for explaining the linear detection function, Figs. 4 to 6 are schematic diagrams of wide-area spatial filters, and Figs. 7 to 9
The figure is a filamentous measurement flowchart, Figures 1θ to 15 are diagrams showing the process of measuring filamentous fungi, and Figures 18 to 18 are diagrams for explaining the prior art. 10... ITV camera, 12... Image memory, 20
. . . Line shape extraction means, 21 . . . Line emphasis spatial filter, 22 . . . Difference filter, 30 . . . . Flock periphery extraction means, 41 . . Linear smoothing filter, 42 . . . Wide area spatial filter, 43
... slicing section, 50... floc removal section, 51.
・Filamentous fungus counter. Applicant's representative Patent attorney Takehiko Suzue Figure 4 Figure 5 Figure 5

Claims (1)

[Claims] Obtain second image data in which the filamentous microorganisms are extracted from first image data obtained by imaging a liquid containing filamentous microorganisms, and further remove flocs formed by the filamentous microorganisms from this second image data. In the filamentous microorganism measuring device that measures the filamentous microorganisms, each spatial filter that emphasizes lines in each direction is applied to the first image data, and the density level of the central part and the side parts in the output of these spatial filters are calculated. A filamentous microorganism measuring device comprising a linear extraction means for extracting the filamentous microorganism based on the difference between the density level of the filamentous microorganism.