JPH0322935B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a monitoring device for substances suspended in water that automatically recognizes the floc formation status in a water treatment plant using image recognition technology.

[Conventional technology]

At water treatment plants, since the suspended particles in raw water are small in size, a process is carried out in which they are coagulated to form flocs, and the flocs are allowed to settle. The quality of floc formation directly affects the subsequent treatment processes, namely precipitation and filtration. Poor floc formation worsens the floc settling properties in the sedimentation basin. Furthermore, this deterioration of sedimentation leads to an overload phenomenon in the filter basin. Furthermore, if the recognition of excessive load is delayed, it can lead to serious accidents such as micro flocs flowing out from the filtration basin.

Therefore, it is necessary to monitor the flock. In the past, surveillance of flocks was carried out by lifeguards.
An example is shown in FIG. In the floc formation pond 1,
A paddle 2 for stirring the flocs is provided, and the flocs are stirred. A lifeguard visually monitors the flocs in the pond. However, it is difficult to quantitatively understand flocs through this visual inspection.

FIG. 2 shows an example in which an industrial television camera 4 and a floodlight 7 are installed in a floc formation pond 1. Drive circuit 8, intake circuit 5, CRT in building 8A
There are 6. The drive circuit 8 supplies power supply voltage to the projector 7 . The reason for providing the floodlight 7 is that the inside of the floc formation pond 1 is crowded and dark.
This is because the TV camera 4 cannot image the flock using natural light alone. CRT6 is TV camera 4
The image captured by the camera is displayed, and the monitor monitors the displayed image.

This conventional example shows the inside of the floc formation pond 1.
A system is in place to capture images using TV camera 4, but
The final method was to visually observe the CRT6 screen. For this reason, it is still difficult to grasp the status of floc formation quantitatively.

[Purpose of the invention]

SUMMARY OF THE INVENTION An object of the present invention is to provide a monitoring device that can quantitatively grasp the formation status of flocs, that is, substances suspended in water.

[Summary of the invention]

In order to achieve the above object, an imaging means installed under the water surface takes an image of the state of a substance suspended in the water and captures the image information, and a binarization means binarizes the image information captured by the imaging means. and the obtained 2
and determining means for determining the formation status of the substance based on the value information.

The image binarized by the binarization means shows information corresponding to the numerical value of each flock, so that the judgment means can make a quantitative judgment.

〔Example〕

Hereinafter, one embodiment of the present invention will be described with reference to the drawings.

FIG. 3 is a schematic overall configuration diagram of a monitoring device according to an embodiment of the present invention, in which an image recognition device 9 is provided. This image recognition device 9
The image taken by the camera 4 is taken in as a binarized image, and the area calculation for the size of the floc, the distribution density of the floc, etc. are performed to quantitatively understand the floc formation status. The CRT 6 may directly display the captured image or may display quantified data.

FIG. 4 is a cross-sectional view of a floc formation pond 1 in which a monitoring device according to an embodiment of the present invention is installed. The floc formation pond 1 consists of three floc formation ponds 1A and 1.
Consists of B and 1C. Flotsk formation ponds 1A, 1B,
1C are connected in cascade via a flow path 13.
Further, the floc ponds 1A, 1B, and 1C are provided with paddles 2A, 2B, and 2C for stirring the floc. Each paddle 2A, 2B, 2C
has blades 60, 61, 62 in a cruciform configuration, and these blades have the role of stirring the floc. The flow path 13 has the function of a rectifying wall. Although the flow path 13 is a flow hole extending from the upper surface to the bottom of the flock pond, it may be a part thereof.

FIG. 5 is an overall perspective view of the floc formation pond 1. The paddles 2A, 2B, 2C stir the inside of the floc formation ponds 1A, 1B, 1C along the longitudinal direction. Rapid mixing ponds (not shown) are provided before and after the floc formation pond 1, and the raw water and various micro floc groups therein are mixed into the floc formation pond 1.
flows into. The paddle stirs the micro flocs, causes them to collide with each other, and gradually aggregates the flocs. This agglomeration increases the particle size of the flocs. The particle size of the flocs in the floc formation pond 1B is larger than that in the floc formation pond 1A, and the particle size of the flocs in the floc formation pond 1C is larger than that in the floc formation pond 1B. That is, the particle size increases each time the particles pass through one floc formation pond.

Therefore, a floc monitoring device was installed in the last floc pond 1C. This is because the purpose is also to monitor the final degree of floc formation. This floc pond 1C is also the pond before the sedimentation pond. That is, when the purpose is to monitor the final degree of floc formation, the floc image recognition device with this configuration can be used before and after the sedimentation tank, that is, in the floc formation tank 1.
It will be placed near the exit of the third pond. On the other hand, when monitoring the process of floc formation, it goes without saying that it may be placed in the first or second of the floc formation ponds 1. Figures 15, 16, and 17 described below
The figure discusses this. An example in this case is shown in FIG.

In the figure, 20 is an airtight container, 21 is an observation window (usually a glass plate), 22 is a wiper, 23 is a wiper drive device, 24 is a back screen, 25
A, 25B, and 25C are respectively control boards, 4 is an industrial television camera (ITV), 31 is a close-up lens, 7 is a floodlight, 5 is an ITV image capture circuit, 8 is a floodlight drive circuit, 42 is a light-shielding cover, and 9 is a The image recognition device 60 is a control device.

When viewed from the front, the wiper 22 has the function of moving left and right on the surface of the glass surface (observation window) 21 to wipe away dirt.

The operation of this embodiment is as follows. The ITV 30 fixed in the airtight container 20 is a close-up lens 3
1, the floc formation pond 1 is observed through the observation window 21.
The image of the flock 12 within 0 is enlarged and recognized.
The wiper 22 driven by the wiper drive device 23 wipes dirt off the surface of the observation window 21.
Operates regularly.

Further, a back screen 24 is installed in front of the observation window 21 in order to recognize flock groups with high contrast and accuracy. back screen 24
Considering that the color of the flock 3 is white, it is desirable to use a black color in order to recognize the flock group with high contrast and accuracy.

Incidentally, the floc formation pond 1 is opened to the atmosphere for the purpose of constantly monitoring flocs. For this reason,
The amount of light incident on the floc formation pond 1C changes over time and is strongly influenced by the weather. A floodlight 7 is usually installed as a means for constantly monitoring the flock. If the purpose is simply monitoring that relies on the eyesight of a maintenance manager, installation of the floodlight 7 is sufficient.

However, changes in ambient illumination strongly affect the image recognition accuracy of the flock group. For example, if the illuminance is low, the flock will be perceived as small, and conversely, if the illuminance is high, the flock will be perceived as large.
In order to eliminate this influence, it is necessary to avoid being influenced by changes in illuminance as a natural phenomenon. In this embodiment, a light-shielding cover 42 is provided to darken the surrounding area, and the illuminance provided only by the projector 7 is set under a certain condition. Note that if the light-shielding cover 42 is not provided, the illumination conditions may be controlled at appropriate times using the floodlight drive circuit 8. Also, if circumstances permit, the floodlight 7
It is better to install multiple units and irradiate the flock from multiple directions.

Incidentally, the flock group is in a fluid state, and its moving speed is about 5 to 50 cm/sec. Therefore, it is necessary to perform image recognition of the flock group at high speed. However, in the case of 512 x 480 pixels, even if a high-speed image recognition device is applied, it takes 11 milliseconds to recognize the pixels. Therefore, in order to accurately recognize images at the speed of current image recognition devices, it is important to reduce the moving speed of the flock group as much as possible. In this case, it is taken into consideration that water is moving in a fixed direction in the floc formation pond 1C due to stirring by the stirring paddle 2C.

That is, in order to suppress the movement of water as much as possible within the floc formation pond 1C, a barrier board is arranged so as to block the movement direction. FIG. 7 is a three-dimensional view of FIG. 6, showing the case where the water is moving from the top to the bottom as shown by the white arrows, and shows the jamb plates 25A, 25B, 25C.
It shows the arrangement of. At this time, Jiyama board 25B
and 25C are constructed so as not to interfere with the light rays from the projector 7 and to allow the light rays to enter in a plane.

FIG. 8 shows the jamming plates 25A, 25B, 25C,
The arrangement of 25D and 25E is shown. Figure 9 shows
Border plates 25A and 25B when ITV4 is placed across wall surface 10A of floc formation pond 1C
It shows the arrangement of. Further, FIG. 10 is a three-dimensional view thereof.

FIG. 17, which will be described later, shows a situation in which flock formation is actually recognized by the ITV4, taking into account the points to be noted in image recognition of flock groups as explained above. Here, the background of the flock 3 is black because the back screen 24 is arranged. Further, although the floodlight 7 is not shown, illumination is emitted from two locations on both sides in the horizontal direction.

In this way, the flock image information captured by the ITV 4 is transmitted to the image recognition device 9 via the ITV capture circuit 5.

In the image recognition device 9, from the obtained image information,
In order to extract information useful for water quality management at water treatment plants, we perform various calculations such as particle size and distribution of flocs. The specific method will be described later, but for example, 2
The grain size of each floc in the floc group is calculated by value processing, and the representative grain size of the floc group is determined.

The image recognition control device 9 controls this time sequence. That is, the image recognition control device 9
The details will be described later, but the image recognition device 9 and ITV4
The recognition time and number of times of recognition of flock image information are adjusted. Generally, the condition of floc formation does not change rapidly in a short period of time, so it is sufficient to perform the series of floc monitoring operations described above once every 10 minutes to once an hour.

In the device configured in this way, after the image information of the flock group is captured by the ITV4,
A specific process of signal processing within the image recognition device 9 is shown in FIG. 11 and will be described in detail below.

Here, 100 is a recognition timing control circuit, 1
01 is an A/D conversion circuit, 102 is a threshold input circuit,
103 is a binarization circuit, 104 is a labeling circuit,
105 is a particle size calculation mode designation circuit, 106 is a particle size measurement circuit, 107 is a particle size comparison circuit, 108A, 1
08B, 108C, . . . 108Z are number storage circuits, 109 is a recognition number control circuit, and 110 is a particle size distribution calculation/display circuit.

Further, FIG. 12 shows the image plane of the flock group recognized by the ITV4. Since the floc group is a grayscale image, the boundary between floc 3 and water is not actually clear, but for the sake of simplicity, only the outline of the floc group is illustrated. The brightness level of the flock is high and white, while the brightness level of the water is low and black due to the back screen 24 in the background.

In FIG. 11, recognition timing control circuit 1
00 recognizes the flock image obtained through the ITV 4 and ITV controller 5 shown in FIG.
Controls the time interval (frequency) for capturing image information. Next, the A/D conversion circuit 101 receives the obtained analog signal of the brightness information, for example, the screen signal shown in FIG. 12, and converts the signal into a digital signal one by one. The converted digital signal is sent to the threshold input circuit 1.
Based on the threshold specified by 02, the binarization circuit 1
In step 03, binarization processing is performed.

For example, FIG. 13 shows a case where the screen shown in FIG. 12 is scanned by line AA' to display the luminance level distribution. Here, the brightness level is displayed in 8 bits (256 levels), and the brightness is low toward the top of the vertical axis, while the brightness is high toward the bottom. Since the flock 12 is white, the brightness is high. In other words, the portion that becomes a valley in the downward direction represents a flock.

In this brightness distribution, each pixel is binarized by a binarization circuit 103 based on a threshold specified by a threshold input circuit 102, for example, the brightness specified by the BB' line. The threshold specified by the threshold input circuit 102 is maintained constant under constant illuminance, but can also be operated by an operator.

In the binarization circuit 103, a pixel at a luminance level higher than the threshold value is set to "1", and a pixel at a luminance level below the threshold value is set to "0". Then,
As shown in FIG. 14, the part corresponding to floc becomes "1", and the part corresponding to water becomes "0".

Examples of the results of flock recognition in this manner are shown in FIGS. 15, 16, and 17. Figure 15 is a diagram showing the recognition and binarization of floc groups in the first pond 1A of the floc formation pond 1.
Figure 6 is a diagram where the floc group in the second pond 1B of floc formation pond 1 is recognized and binarized, and Figure 17 is a diagram where the floc group in the third pond 1C of floc formation pond 1 is recognized and binarized. It is. From these figures, it can be seen that the process of increasing the floc particle size can be clearly determined.

Once the flocs have been recognized, the representative grain size of the flocs is then calculated, but before that a number is assigned to each floc. That is, the labeling circuit 104 independently recognizes each flock and assigns a number to each flock. Then, the particle size measuring circuit 106 calculates the representative particle size of each floc in the order of its number.

As shown in Fig. 18, the representative particle diameters are the directional diameter D _c , the maximum diameter D _nax , the minimum diameter D _nio , and the area circle equivalent diameter.
There are D _cir and equivalent circle peripheral major axis D _c1 . here,
The constant diameter D _c indicates a certain diameter in the horizontal direction. The area circle equivalent diameter D _cir and the equivalent circle peripheral major axis D _c1 are defined by the following equations.

D _cir = (4S/π) ^0.5 (1) D _c1 = L/2/π (2) Here, S is the area of the flock, and L is the peripheral length of the flock.

The particle size calculation mode designating circuit 105 specifies a representative particle size to be adopted from among these representative particle sizes.
In this way, the grain size of each floc is calculated in accordance with the designated standard of representative grain size.

The grain size comparison circuit 107 compares the grain sizes of each floc, and stores the number of flocs of each grain size in the corresponding storage location, that is, the number storage circuit 108.
The numbers are stored in A, 108B, 108C, . . . 108Z. Since the image of the floc is binarized, the minimum unit for measuring the particle size is one pixel. Therefore, if the number of flocs corresponding to each particle size is N _i , then, for example, the number storage circuit 108
A is the number of flocs whose grain size corresponds to one pixel
N ₁ is stored, the number storage circuit 108B stores the number N ₂ of flocs whose grain size corresponds to 2 pixels, and the number storage circuit 108C stores the number N ₃ of flocs whose grain size corresponds to 3 pixels. Ru. The results are shown in FIG.

By the way, in order to accurately determine the particle size distribution of flocs, it is necessary to sample a wide range within the floc formation pond. The particle size of the flocs in one of the floc formation ponds 1 is approximately 0.01 to 0.1 mm. On the other hand, the third pond naturally contains many small flocs, but the flocs grow to around 1 mm. The growth process of this flock is the 15th
As shown in FIGS. 16 and 17.
To recognize minute flocs of about 0.01 mm, it is necessary to use a microscope, and in fact, it is difficult to recognize them with a close-up lens.

In one of the floc formation ponds 1, the particle size of the flocs is small, so the number of flocs is sufficiently large. However, in the third pond of the floc formation pond 1, the particle size of the flocs is large, so the number of flocs is small. In order to accurately determine the particle size distribution of flocs when the number of flocs is small, it is necessary to recognize as many flocs as possible. The number of flocks is preferably several hundred or more. Therefore, it is possible to enlarge the screen for recognition, but doing so makes it difficult to recognize small flocs. Therefore, there is naturally a limit to the size of the screen that allows both small and large flocks to be recognized in a well-balanced manner. In this way, it can be seen that in the case of water flocs, when determining the particle size distribution with high accuracy, the floc image information obtained by single-screen image recognition is insufficient due to large variations in particle size distribution. .

For these reasons, it is necessary to temporarily store the information on the flock image obtained through image recognition, capture the images multiple times, and determine the particle size distribution based on this stored information. Furthermore, in the case of the floc formation pond 1, since the water is stirred by the stirring paddle 11, it is considered that water containing new floc is constantly flowing in front of the ITV 4. In other words, a floc image is captured at regular time intervals, this is repeated multiple times, and the particle size distribution of the floc is determined based on this information.

The number of times the flock image is recognized is specified in the recognition number control circuit 109 in FIG.
Return to 00. The recognition timing control circuit 100 captures an image at a specified timing, repeats the operations described above, calculates the grain size of the flocs, and stores the number of each floc in the number storage circuits 108A, 108B, 108C, . . . 108Z. to be memorized.

On the other hand, when the number of recognition times control circuit 109 reaches the specified number of times, the particle size distribution calculation display circuit 110 starts the number storage circuits 108A, 108B, 108
Based on the values of C, . . . 108Z, the number concentration distribution for each particle size is calculated. That is, particle size i
The number concentration M _{i of} is calculated by the following formula.

M _i =N _i /V/N _r (3) Here, V is the observation space volume by the ITV 4, and N _r is the number of recognitions specified by the recognition number control circuit 109.

Note that in the number concentration distribution, since there are many flocs with minute particle diameters, it goes without saying that a method of illustrating the flocs as a volume concentration distribution may be adopted, assuming that the flocs are spheres. In addition, the recognition number control circuit 109
The number of times of recognition in , and the recognition timing in the recognition timing control circuit 100 are both appropriately operated from the image recognition control device.

Next, an operation is performed to determine whether or not the grain size of the floc is normal. As shown in FIG. 20, the grain size of the floc that serves as a reference is determined and is defined as this grain size _Ds .
The number concentration ratio of flocs with a particle size larger than this D _s is calculated. FIG. 21 shows a diagram for this calculation process.

Among the signals output by the particle size distribution calculation and display circuit 110, the signal portion corresponding to the particle size is compared with the reference particle size D _s set by the reference particle size setting circuit 111. A comparison is made in circuit 112.
Then, the number of flocs having a grain size larger than the reference grain size _Ds is stored in the growing floc number concentration storage circuit 1.
13. Let this number concentration be Mg. On the other hand, the number of flocs having a particle size smaller than the reference particle size _Ds is stored in the micro floc number concentration storage circuit 114.
to be memorized. Let this number concentration be M _n .

At the same time, the sum M _t of M _g and M _n is calculated and stored in the total flock number concentration storage circuit 115 . The particle size comparison circuit 116 compares whether the ratio of the growing floc number concentration M _g to the total floc number concentration M _t is greater than a predetermined value r.

Mg/Mt≧r (4) That is, when Mg/Mt is larger than the predetermined value r, it indicates that there are many growing flocs, and therefore the flocs are considered to be well formed. on the other hand,
Conversely, when Mg/Mt is smaller than the predetermined value r, it means that there are a large number of minute flocs, and the condition of floc formation is considered to be poor.

Based on these determination results, an alarm output signal can be issued if the floc formation is defective. Similarly, based on the determination result of equation (4), the flocculant injection amount, the alkali agent injection amount, and the stirring paddle 2
It is possible to control the number of rotations.

Figures 22a and 22b show an example of a time chart for flock recognition processing. For example, if the moving speed of the floc 3 flowing in front of the observation window 21 attached to the airtight container 20 as shown in FIG. 6 is v ₁ cm/sec,
Assuming that the size of the area where the image of the flock is recognized is l ₁ square, the time T ₁ required for the flock to pass from one end of the image recognition area to the other is T ₁ = l ₁ /v ₁ ……( 5) becomes. In other words, by setting the flock image capture time interval of the recognition timing control circuit 100 shown in FIG. 11 to T1 _or more, it becomes possible to always capture new flock information. Of course, it is obvious that flock images may be captured at intervals of _T1 time or less.

Further, the recognition number control circuit 109 shown in FIG.
The recognition operation is repeated the number of recognition times N _r set in , and the number concentration distribution calculation (T ₂
The series of processing (which takes time) (hereinafter referred to as the flock recognition step) is completed. These processes are performed at high speed by the image recognition device 9,
The time required for this (T ₁ +T ₂ ) is very short, on the order of several seconds.

Therefore, after the floc recognition step is completed at a certain time, it is possible to repeat the floc recognition step one after another, but in general, the responsiveness of the process control of a water treatment plant (for example, the It is characterized by the fact that the time it takes for the reaction to spread throughout the treatment process is long, on the order of several tens of minutes to several hours. Therefore, there is little need to constantly repeat the flock recognition step, and the flock recognition step is repeated N _s times at every arbitrary time interval T ₃ (requires (T ₁ + T ₂ ) × N _s seconds). The image recognition device 9 may be used to perform other processing (for example, fish tracking processing for monitoring the inflow of poisonous substances) using the time _T4 .
It is highly effective in terms of improving utilization efficiency and multi-purpose effective use.

〔Effect of the invention〕

According to the present invention, it is possible to objectively, quantitatively, and accurately measure the floc particle size distribution, from minute flocs to growing flocs, on-line automatically during the water floc formation process. This makes it possible to save labor and improve reliability in water treatment plant maintenance and management, and even to apply it to control systems.

[Brief explanation of the drawing]

1 and 2 are conventional examples, FIG. 3 is a schematic overall configuration diagram of a monitoring device according to an embodiment of the present invention, and FIG.
The figure is a cross-sectional view of a floc formation pond, FIG. 5 is a perspective view of a floc formation pond of a water treatment plant, FIG. 6 is a diagram showing an embodiment of the present invention, and FIGS. 1
FIG. 0 is a diagram showing details of an embodiment of the present invention, FIG. 11 is a diagram showing a signal writing process in the embodiment, and FIGS. 12, 13, and 14 are diagrams specifically showing a signal processing process. Figures 15, 16, and 17 are views showing binarized images obtained by implementing the present invention, and Figures 18, 19, and 20 are views showing details of the embodiment of the present invention. 21 is a diagram specifically showing the signal processing process, and FIG. 22 is a diagram showing a processing time chart. 1, 1A, 1B, 1C... floc formation pond, 2
... Paddle, 3 ... Flock, 4 ... TV camera, 7 ... Floodlight, 9 ... Image recognition device, 60 ...
...Control circuit, 101...AD converter, 103...
Binarization circuit, 104...Labeling circuit, 105
... Particle size measurement circuit, 107 ... Particle size comparison circuit, 1
08A~108Z...Memory.

Claims (1)

[Claims] 1. Imaging means installed under the water surface for imaging the state of substances suspended in the water and capturing the image information;
Binarizing the image information captured by the imaging means 2
1. A monitoring device for substances suspended in water, comprising a digitizing means and a determining means for determining the state of formation of the substance based on the obtained binarized information. 2. An imaging means installed under the water surface for imaging the state of substances suspended in the water and capturing the image information;
Binarizing means for binarizing the image information of multiple screens captured by the imaging means, and determining the particle size distribution of the substance from the obtained binarized information and determining the formation status of the substance based on the particle size distribution. 1. A monitoring device for substances suspended in water, comprising: determination means for determining. 3. Imaging means installed under the water surface to image the state of substances suspended in the water and capture the image information;
Binarizing the image information captured by the imaging means 2
The method comprises a digitizing means and a determining means for determining the formation status of the substance based on the obtained binarized information, and operates the imaging means, the binarizing means, and the determining means at arbitrary times. A monitoring device for substances suspended in water, characterized in that the formation status of the substances is determined at arbitrary intervals.