JP2020134559A - Observation device, observation method and program - Google Patents

Abstract

To provide an observation device capable of reliably observing a subject in a consecutive manner or at a fixed time interval.SOLUTION: An observation device 20 for observing a subject such as microorganism contained in a collected water comprises a light source 23 for irradiating the microorganism with light, and an imaging element 21 having a light receiving part 21a which receives light with which the microorganism is irradiated and converting the received light into an image signal. The light receiving part 21a of the imaging element 21 is protected by protective glass 22.SELECTED DRAWING: Figure 2

Description

The present invention relates to an observation device, an observation method, and a program for observing a subject such as a microorganism.

Conventionally, a wastewater treatment device for treating wastewater such as sewage with activated sludge containing microorganisms that oxidatively decompose organic substances has been known (see, for example, Patent Document 1). The wastewater treatment apparatus described in Patent Document 1 includes a biological treatment tank and a final sedimentation pond. In the biological treatment tank, biological treatment is carried out by activated sludge, and organic substances and the like contained in the wastewater to be treated are removed. The wastewater after the biological treatment has been carried out (hereinafter referred to as "treated water") is then transferred to the final sedimentation pond. In the final sedimentation pond, activated sludge contained in the treated water is settled and separated from the treated water. Subsequently, the treated water in which the activated sludge is settled and separated (hereinafter referred to as "sludge removal water") is disinfected and discharged into a river or the like, and the activated sludge settled and separated from the treated water is discharged into a biological treatment tank. Will be returned to.

By the way, if the operation method regarding the biological treatment executed in the biological treatment tank is mistaken, the bacterial flora indicating the type and distribution of the microorganisms contained in the activated sludge may change, and the wastewater treatment may not be smoothly executed. Specifically, if the operation method of the biological treatment tank is mistaken, filamentous microorganisms that proliferate in the form of filaments may propagate in the activated sludge, and as a result, the activated sludge may be transformed into bulking sludge having poor sedimentation property.

When the activated sludge is changed to bulking sludge, the bulking sludge is difficult to settle in the final sedimentation pond, so that a bulking phenomenon occurs in which the bulking sludge is not settled and separated from the treated water, and the sludge removal water cannot be properly obtained. Therefore, when the bulking phenomenon occurs, it is necessary to take appropriate measures. However, since the causes of the bulking phenomenon are diverse and the microorganisms that make up activated sludge are also diverse, there are multiple measures for the bulking phenomenon. , It is not easy to select an appropriate measure from multiple measures. In addition, when the bulking phenomenon occurs, it is necessary to return the flora to the normal flora before the bulking phenomenon occurred, but it takes a long time to normalize the flora, so the wastewater treatment function is long. Stop for a period.

Correspondingly, observing each change of microorganisms contained in activated sludge continuously or at regular time intervals helps to select appropriate measures against the bulking phenomenon, and also helps to select appropriate measures for the microbial phenomenon. Since changes in the bacterial flora are recognized at an early stage, it is expected that adjustment of the operating conditions of the biological treatment tank will help to eliminate the bulking phenomenon in a short period of time.

Japanese Unexamined Patent Publication No. 2006-247943

However, at present, microorganisms are observed with a microscope, and observation of the microorganisms requires a high degree of specialization. Therefore, observing microorganisms at all wastewater treatment plants where wastewater treatment equipment is installed with a microscope secures human resources. It is difficult from the viewpoint of Therefore, there is a fact that it is practically not possible to reliably observe a subject such as a microorganism continuously or at regular intervals.

An object of the present invention is to provide an observation device, an observation method, and a program capable of reliably observing a subject continuously or at regular time intervals.

In order to achieve the above object, the observation device of the present invention is an observation device for observing a subject, in which the irradiation means for irradiating the subject with light and the subject in close proximity to the subject and the irradiated light. It is characterized by having a light receiving unit that receives light that has passed through the light receiving unit, a conversion means that converts the light received by the light receiving unit into an image signal, and the conversion means having a protective means that protects the light receiving unit. And.

According to the present invention, the subject can be reliably observed continuously or at regular time intervals.

It is a block diagram which shows schematic the wastewater treatment apparatus for treating wastewater observed by the observation apparatus which concerns on embodiment of this invention. It is a figure used for demonstrating the observation apparatus which observes activated sludge contained in a biological reaction tank in FIG. It is a flowchart which shows the procedure of the observation processing of activated sludge executed by the observation apparatus of FIG. It is a photograph showing an image containing microorganisms observed by the observation device of FIG. It is a block diagram which shows schematic structure of the information processing apparatus connected to the observation apparatus of FIG.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram schematically showing a wastewater treatment device 10 for treating wastewater observed by the observation device 20 according to the embodiment of the present invention.

The wastewater treatment device 10 of FIG. 1 includes a sand basin 11, a first settling basin 12, a biological reaction tank 13, a final settling basin 14, and a sterilization tank 15. The wastewater to be treated first flows into the sand basin 11, and relatively large foreign substances such as stones and wood fragments contained in the drainage are removed from the drainage, and the wastewater from which the relatively large foreign substances are removed is first sent to the sedimentation basin 12. Inflow. In the first settling basin 12, relatively small foreign substances (hereinafter, referred to as “primary sludge”) that were not removed in the sand basin 11 are removed, and the wastewater from which the initial sludge has been removed flows into the biological reaction tank 13. In the biological reaction tank 13, organic matter and the like in the wastewater are removed by microorganisms contained in the activated sludge, and the wastewater from which the organic matter and the like have been removed flows into the final sedimentation pond 14.

The biological reaction tank 13 is provided with a pump P, and wastewater in which activated sludge is dispersed is collected from the biological reaction tank 13 by the pump P. In the final sedimentation pond 14, the activated sludge is removed, and the wastewater from which the activated sludge has been removed flows into the sterilization tank 15, is disinfected, and then discharged into a river or the like. In addition, a part of the activated sludge removed in the final settling pond 14 is returned to the biological reaction tank 13 as return sludge.

FIG. 2 is a diagram used to explain the observation device 20 for observing the activated sludge contained in the biological reaction tank 13 in FIG. 1, and FIG. 2 (A) is used to explain the configuration of the observation device 20. 2 (B) is a view showing the appearance of the observation device 20, and FIG. 2 (C) is a partial cross-sectional view of the observation device 20 of FIG. 2 (B).

The observation device 20 of FIG. 2A includes an image sensor 21 (conversion means) which is an image sensor such as a CCD sensor or a CMOS sensor, a protective glass 22 (protective means) for protecting the image sensor 21, and a light source 23 (irradiation means). ), And the protective glass 22 is placed close to the image sensor 21 and fixed using, for example, an insulating adhesive. As a result, since the image sensor 21 does not come into contact with the wastewater, it is possible to prevent the wastewater from being contaminated. The light source 23 is, for example, an LED lamp, and the light emitted from the light source 23 reaches the image sensor 21 (light receiving unit 21a) through the protective glass 22.

The image sensor 21 converts the received light into a digital image signal. The image pickup device 21 is connected to, for example, the information processing device 50, outputs a digital image signal to the information processing device 50, and the information processing device 50 receives the digital image signal converted by the image pickup device 21 and converts it into a digital image signal. Generate based image data. At this time, the observation device 20 may be provided with a storage medium (storage means) such as an HDD, and the digital image signal converted by the image pickup device 21 is stored in the storage medium. The generated image data is stored in the information processing device 50 and is displayed on, for example, a display device 24 connected to the information processing device 50. The number of pixels of the image sensor 21 is preferably 1 million pixels (1MP) or more.

The image sensor 21 and the light source 23 to which the protective glass 22 is fixed are housed in a housing, for example, and the distance between the image sensor 21 and the light source 23 is kept constant (FIGS. 2 (B) and 2 (C)). .. At this time, the light source 23 is positioned so as to irradiate the center of the image pickup device 21 with light, and the optical axis of the light (the optical axis L in FIG. 2A) is the protective glass 22 and the light receiving portion 21a of the image pickup element 21. Is almost orthogonal to. A lens, a reflector, or a light tube (light tube 25 in FIG. 2C) may be used so that the light emitted from the light source 23 is focused on the image sensor 21. Further, the housing may be configured to be open so that the wastewater containing microorganisms can be dropped onto the protective glass 22.

FIG. 3 is a flowchart showing a procedure of the activated sludge observation process executed by the observation device 20 of FIG.

In FIG. 3, first, the pump P collects wastewater (hereinafter referred to as “observation target water”) in which activated sludge containing microorganisms is dispersed from the biological reaction tank 13 (S301), and the observation target water is protected. It is dropped onto the glass 22 (S302). Next, the observation target water dropped from the light source 23 onto the protective glass 22 is irradiated with light (S303), and the image sensor 21 receives the light emitted from the light source 23 and passed through the observation target water (S304). , Light receiving step), the received light is converted into a digital image signal (S305, conversion step).

In this case, when the microorganism is contained in the water to be observed, the light emitted from the light source 23 transmits the microorganism, but the transmittance is not 100%. That is, the light emitted from the light source 23 reaches the image sensor 21 according to the transmittance of the microorganisms contained in the water to be observed. Therefore, the image sensor 21 receives the light emitted from the light source 23 that has passed through the observation target water and the protective glass 22 in that order, and converts the received light into a digital image signal. After that, the digital image signal converted by the image sensor 21 is output to, for example, the information processing device 50 connected to the image sensor 21, and the information processing device 50 generates image data (S306). The generated image data is displayed on, for example, a display device 24 connected to the information processing device 50.

According to the observation process of FIG. 3, the light emitted from the light source 23 and passed through the observation target water is converted into a digital image signal (S305), but when the microorganism is contained in the observation target water, the light is transmitted from the light source 23. The irradiated light reaches the image pickup device 21 according to the transmission rate of the microorganisms contained in the observation target water. As a result, the image data generated from the digital image signal based on the observation target water containing the microorganism includes a shadow or its outline according to the shape of the microorganism that does not transmit at least a part of the light from the light source 23 ( FIG. 4). Therefore, since the shape of the microorganisms contained in the observation target water can be visually grasped, for example, the observation target water is sampled (S301) and the observation target water is dropped onto the protective glass 22 (S302) continuously or constantly. When controlled to be performed automatically every hour, a subject such as a microorganism can be reliably observed continuously or at regular intervals.

Further, the water to be observed is observed by being dropped onto the protective glass 22 (S302). Since the protective glass 22 is fixed to the light receiving portion 21a of the image sensor 21, the water to be observed comes into contact with the protective glass 22 and is extremely close to the image sensor 21. Therefore, since the light that has passed through the observation target water and the protective glass 22 in sequence is not diffused until it is received by the light receiving portion 21a, it is possible to prevent the outline according to the shape of the microorganisms and the like contained in the observation target water from becoming unclear. can do.

By the way, when the observation target water is dropped on the protective glass 22, water droplets of the observation target water having a predetermined thickness are formed on the glass, so that the observation target dropped on the protective glass 22 from the light source 23. When the water is irradiated with light, the light irradiated to the observation target water having a predetermined thickness may reach the image pickup element 21 through a plurality of microorganisms or fine particles in the observation target water. At this time, the image data generated based on the digital image signal converted by the image sensor 21 indicates a shadow in which the shadows of a plurality of objects are combined, and the shadow or its outline according to the shape of the microorganism is the image data. Not grasped from.

Correspondingly, the image sensor 21 may be tilted with respect to the horizontal direction. Specifically, the image sensor 21 is tilted in a range of 0.5 degrees or more and 10 degrees or less with respect to the horizontal direction. As a result, excess observation target water flows in the inclined direction of the image pickup element 21 based on the gravity of the observation target water dropped on the protective glass 22, and water droplets of observation target water optimal for observing individual microorganisms. The thickness of the can be secured on the protective glass 22. As a result, the light irradiated to the observation target water reaches the image pickup device 21 according to the light transmittance by the microorganisms remaining on the protective glass 22, and is based on the digital image signal formed by the image pickup device 21. The shadow or its outline according to the shape of the microorganism can be grasped from the image data generated in the above.

Further, as a different measure, even when the image pickup element 21 is not tilted with respect to the horizontal direction, the observation target water is diluted with water according to the viscosity and turbidity of the observation target water, and the diluted observation is performed. The target water may be dropped onto the protective glass 22. This makes it possible to grasp the same effect as described above, that is, the shadow or its outline according to the shape of the microorganism.

It should be noted that the observation device 20 is used when observing the microorganisms contained in the water to be observed, but the present invention is not limited to this. For example, there is known a clean water treatment applied to raw water in order to obtain drinking water from raw water such as a river. The clean water treatment has a coagulant addition step of adding a coagulant to the raw water in order to remove turbidity in the raw water. In the flocculant addition step, when the flocculant is first added to the raw water, aggregate nuclei for agglomerating the turbidity of the raw water are formed in the raw water, and then the turbidity is aggregated in the formed aggregate nuclei. Flocks are formed and the flocs are dispersed in the raw water. When the raw water in which the flocs are dispersed or the raw water having aggregated nuclei is used as the observation target water, the observation device 20 is also used when observing the flocs or aggregated nuclei contained in the observation target water. That is, the observation device 20 is used when observing a minute solid contained in the liquid as a subject, and can grasp a shadow or its outline according to the shape of the subject contained in the liquid.

FIG. 5 is a block diagram schematically showing the configuration of the information processing device 50 connected to the observation device 20 of FIG.

The information processing device 50 of FIG. 5 includes a CPU 51 (extraction means, identification means), a RAM 52, a ROM 53, and an HDD 54, which are connected to each other, and the CPU 51 is connected to the observation device 20. The ROM 53 or HDD 54 stores programs, various data, and the like. The various data are, for example, a digital image signal output by the image pickup device 21, image data generated based on the digital image signal, and learning data in which information related to each image data is associated with these image data.

Therefore, when the microorganisms contained in the water to be observed are observed by the observation device 20 and the image pickup device 21 outputs a digital image signal, the ROM 53 or HDD 54 displays the digital image signal and the image data generated based on the digital image signal, and Stores learning data associated with active sludge property data that quantifies the bacterial flora state. In addition, other information, for example, water quality data such as the amount of nitrogen and the amount of phosphorus in the observation target water may be associated with the learning data.

Further, when the aggregated nuclei contained in the water to be observed are observed by the observation device 20 and the image pickup device 21 outputs a digital image signal, the ROM 53 or HDD 54 displays the digital image signal and the image data generated based on the digital image signal, and , The learning data associated with the characteristics of flocs formed in the future when aggregated nuclei grow and the turbidity of the water to be observed (hereinafter referred to as "aggregated nucleus-related data") is stored.

The CPU 51 expands the program stored in the ROM 53 or the HDD 54 into the RAM 52 and executes it, and also generates image data based on the digital image signal stored in the ROM 53 or the HDD 54. Further, the CPU 51 executes deep learning by utilizing the learning data. Further, after executing deep learning, the CPU 51 acquires new image data, identifies activated sludge property data from microorganisms contained in the image data, or aggregate nuclei grow from aggregate nuclei contained in the image data. Identify detailed information such as the shape of the formed flocs.

Subsequently, a method of deep learning executed by the CPU 51 and a method of acquiring new image data after deep learning and specifying information about a subject included in the image data will be described. Here, a method will be described in which the CPU 51 learns the image data containing a microorganism and the activated sludge property data associated with the image data, and then identifies the activated sludge property data related to the microorganism from the image data containing the microorganism. ..

The ROM 53 or HDD 54 stores, for example, image data generated based on the digital image signal output by the image sensor 21, and learning data associated with activated sludge property data in which the bacterial flora state is quantified. Here, the feature amount calculated according to the image data is also added to the learning data.

Specifically, the CPU 51 performs processing such as changing the image size such as enlargement, reduction, and trimming, adjusting the brightness, or adjusting the image quality of the color adjustment on the image data. Next, the CPU 51 performs a convolution calculation process, which is a filter process for extracting local features of image data, a pooling calculation process for compressing data while leaving the features extracted by the convolution calculation process, or a convolution calculation process or pooling. All the image data obtained by the arithmetic processing is combined into one, and the all-combined arithmetic processing that converts it into one-dimensional data by the activation function and outputs it is applied to the processed image data to be discriminated to obtain the feature amount. calculate.

In the present embodiment, (1) convolution operation processing, (2) pooling operation processing, (3) multiple times (for example, 2 to 9 times) convolution operation processing, and (4) fully connected operation processing are performed in this order. As a result, a specific value between 0 and 1 is finally calculated as the feature amount for the image data. The calculated feature amount is added to the image data and stored in the ROM 53 or HDD 54 as learning data associated with the water quality data.

On the other hand, when a new state of microorganisms is observed, the image sensor 21 newly forms a digital image signal and outputs it to the information processing device 50. Subsequently, the CPU 51 generates image data (hereinafter, referred to as “discrimination target image data”) based on the digital image signal output by the image sensor 21, and the discrimination target image data is, for example, enlarged, reduced, trimmed, or the like. Perform processing such as changing the image size, adjusting the brightness, or adjusting the image quality of the color adjustment. Next, the CPU 51 applies a convolution calculation process, a pooling calculation process, or a fully coupled calculation process to the processed image data to be discriminated. In the present embodiment, (1) convolution operation processing, (2) pooling operation processing, (3) multiple times (for example, 2 to 9 times) convolution operation processing, and (4) fully connected operation processing are performed in this order. Apply with. As a result, a specific value between 0 and 1 is finally calculated as the feature amount for the image data to be discriminated.

After that, the CPU 51 identifies the activated sludge property data corresponding to the discrimination target image data from the discrimination target image data, the feature amount calculated based on the discrimination target image data, and the learning data stored in the ROM 53 or the HDD 54. The identified activated sludge property data is recorded in ROM 53 as log data in association with the image data to be discriminated and the feature amount, and is stored in ROM 53 or HDD 54 as learning data.

According to the information processing device 50 of FIG. 5, the CPU 51 that has executed the deep learning discriminates from the feature amount calculated based on the discrimination target image data and the discrimination target image data and the learning data stored in the ROM 53 or the HDD 54. Identify the active sludge property data corresponding to the target image data. When the activated sludge property data corresponding to the discrimination target image data is specified, the bacterial flora state of the microorganism is grasped from the discrimination target image data. As a result, even when it is difficult to observe microorganisms using a microscope that requires a high degree of specialization, the microorganisms are automatically observed continuously or at regular intervals, so that the microorganisms in activated sludge can be observed. Changes in the bacterial flora can be easily grasped.

Further, when the information processing apparatus 50 learns learning data in which image data including aggregate nuclei and data related to aggregate nuclei are associated with each other, and then acquires new discriminant target image data including aggregate nuclei, an image containing microorganisms. In the same way as specifying the active sludge property data related to the microorganism from the data, it is possible to specify the aggregated nucleus-related data corresponding to the image data to be discriminated. That is, when the information processing device 50 acquires new image data after learning the subject included in the image data and the information associated with the subject, the information processing device 50 obtains the information to be associated with the subject included in the new image data. Can be identified.

Although the embodiments of the present invention have been described above, the present invention is not limited to these embodiments.

The present invention supplies a program that realizes one or more functions of the above-described embodiment to a system or device via a network or storage medium, and one or more processors in the computer of the system or device reads and executes the program. It can also be realized by processing, and can also be realized by a circuit (for example, ASIC) that realizes one or more functions.

20 Observation device 21 Image sensor 22 Protective glass 23 Light source 24 Display device 50 Information processing device 51 CPU

Claims (8)

In an observation device for observing a subject
An irradiation means for irradiating the subject with light,
It is provided with a light receiving unit that is close to the subject and receives the light that has passed through the subject among the irradiated light, and includes a conversion means that converts the light received by the light receiving unit into an image signal.
An observation device, wherein the conversion means has a protective means for protecting the light receiving portion. The observation device according to claim 1, wherein the subject is a floc generated when removing microorganisms or turbid substances in the collected water. The observation device according to claim 1 or 2, wherein the light receiving portion is inclined with respect to the horizontal direction. The observation device according to any one of claims 1 to 3, wherein the light receiving portion is inclined in a range of 0.5 degrees or more and 10 degrees or less with respect to the horizontal direction. The observation device according to any one of claims 1 to 4, wherein the conversion means includes a storage means for storing the image signal. An information processing device connected to the conversion means and receiving the image signal from the conversion means is provided.
The information processing device
An extraction means that receives the image signal to generate image data and extracts a feature amount of the image data.
The observation device according to any one of claims 1 to 5, further comprising a specific means for specifying information on water including the subject based on the feature amount. In an observation method for observing the subject using an irradiation means for irradiating the subject with light and an observation device including a light receiving unit close to the subject.
A light receiving step in which the light receiving unit receives the light that has passed through the subject among the light irradiated to the subject.
An observation method characterized by having a conversion step of converting the light received by the light receiving unit into an image signal. A program that causes a computer to execute an observation method for observing the subject by using an irradiation means for irradiating the subject with light and an observation device provided with a light receiving unit close to the subject.
The observation method is
A light receiving step in which the light receiving unit receives the light that has passed through the subject among the light irradiated to the subject.
A program characterized by having a conversion step of converting the light received by the light receiving unit into an image signal.