Information processing device, information processing system, information processing method, and program
The described system facilitates real-time assessment of ion exchange resin deterioration using machine learning, enabling efficient filler state recognition and timely replacements.
Patent Information
- Application Number
- JP2020183558
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2020-11-02
- Publication Date
- 2024-11-01
- Estimated Expiration
- 2040-11-02
AI Technical Summary
Existing methods for determining the physical deterioration of ion exchange resins require stopping the ion exchange equipment, which is time-consuming and makes it difficult to assess the state of the filler in containers like resin towers.
An information processing device and system that uses a camera to capture images of the filler, processes the images using a learned model through machine learning to identify deterioration, and outputs a notification based on the calculated state of the filler.
Enables easy recognition of the state of the filler in containers, allowing for timely replacement decisions without interrupting operations.
Smart Images

Figure 0007576959000001 
Figure 0007576959000002 
Figure 0007576959000003
Abstract
Description
[Technical field]
[0001] The present invention relates to an information processing device, an information processing system, an information processing method, and a program. [Background technology]
[0002] Ion exchange resins are widely used as desalination means for removing ions contained in industrial water, groundwater, etc., and reducing dissolved salts in the water. Such ion exchange resins undergo physical deterioration with use, and when the resins no longer satisfy the required performance due to the physical deterioration, they must be replaced.
[0003] Generally, to determine the physical deterioration of an ion exchange resin, the operation of an ion exchange device (a container including a resin tower and a tank) filled with the ion exchange resin is stopped once, the ion exchange resin is separated and collected from the stopped ion exchange device, and the separated and collected ion exchange resin is observed using a microscope or the like to determine the degree of physical deterioration. Also, a technology has been disclosed in which a predetermined liquid is brought into contact with the ion exchange resin separated and collected from the ion exchange device, and the components of the liquid after contact are analyzed to evaluate the degree of deterioration (see, for example, Patent Document 1). [Prior art documents] [Patent documents]
[0004] [Patent Document 1] JP 2001-228134 A Summary of the Invention [Problem to be solved by the invention]
[0005] In the above-mentioned technology, in order to judge the physical deterioration of the ion exchange resin, the ion exchange device in operation must be stopped. Also, it takes time to obtain the judgment result of the physical deterioration of the ion exchange resin. Therefore, there is a problem that the state of the packing filled in the container such as the resin tower cannot be easily recognized.
[0006] An object of the present invention is to provide an information processing device, an information processing system, an information processing method, and a program that can easily recognize the state of a filling material filled in a container. [Means for solving the problem]
[0007] The information processing device of the present invention comprises: an image data acquisition unit that acquires image data showing an image of a filling material filled in a container into which water flows and from which treated water flows out; A calculation unit that calculates the state of the filling material contained in the image data by inputting the image data acquired by the image data acquisition unit into a trained model that has undergone machine learning; and an output unit that outputs a result of the calculation performed by the calculation unit.
[0008] In addition, the information processing system of the present invention comprises: A camera and an information processing device are included. The camera includes: An imaging unit that captures an image of the filling material filled in the container; a transmission unit that transmits image data representing an image captured by the imaging unit to the information processing device, The information processing device includes: an image data acquisition unit that acquires the image data transmitted from the transmission unit; A calculation unit that calculates the state of the filling material contained in the image data by inputting the image data acquired by the image data acquisition unit into a trained model that has undergone machine learning; and an output unit that outputs a result of the calculation performed by the calculation unit.
[0009] The information processing method of the present invention further comprises: A process of acquiring image data showing an image of the filling material filled in the container; A process of calculating a state of the filling material contained in the image data by inputting the acquired image data into a trained model that has undergone machine learning; and outputting the result of the calculation.
[0010] In addition, the program of the present invention is On the computer, acquiring image data showing an image of the contents filled in the container; A step of calculating a state of the filling material contained in the image data by inputting the acquired image data into a trained model that has undergone machine learning; and a procedure for outputting the result of the calculation. Effect of the Invention
[0011] In the present invention, the state of the contents filled in the container can be easily recognized. [Brief description of the drawings]
[0012] [Figure 1] FIG. 1 is a diagram illustrating a first embodiment of an information processing system according to the present invention. [Diagram 2] 2 is a diagram showing an example of the internal configuration of the camera shown in FIG. 1. [Diagram 3] 2 is a diagram illustrating an example of an internal configuration of the information processing device illustrated in FIG. 1. [Figure 4] FIG. 4 is a diagram showing an example of a change in sphericity over the number of days of use learned by the trained model shown in FIG. [Diagram 5] 2 is a flowchart illustrating an example of an information processing method in the information processing device shown in FIG. [Figure 6] 10 is a flowchart illustrating another example of the information processing method in the information processing device shown in FIG. [Figure 7] FIG. 11 is a diagram illustrating a second embodiment of an information processing system according to the present invention. [Figure 8] 8 is a diagram illustrating an example of an internal configuration of the information processing device illustrated in FIG. 7. [Figure 9] 8 is a flowchart illustrating an example of an information processing method in the information processing device shown in FIG. 7. [Figure 10] FIG. 13 is a diagram illustrating a third embodiment of an information processing system according to the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0013] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. (First embodiment)
[0014] FIG. 1 is a diagram showing a first embodiment of an information processing system of the present invention. As shown in FIG. 1, the information processing system in this embodiment has a camera 100 and an information processing device 200. Water flows into the container 300, and the container 300 discharges the treated water as treated water. The container 300 includes, for example, an ion exchange resin tower or tank filled with ion exchange resin. The size of the container 300 is not particularly limited, and the container 300 may be small or large. The shape and material of the container 300 are not particularly limited, and the container 300 may be, for example, a resin column whose entire surface is assembled with acrylic plates. The container 300 is provided with window parts 310-1 and 310-2 so that the camera 100 can capture an image of the filling inside from the outside. In the example shown in FIG. 1, two window parts 310-1 and 311-2 are provided, but the number is not limited. The window parts 310-1 and 311-2 have a structure that allows the contents inside to be seen more clearly from the outside than the other parts of the container 300 other than the window parts 310-1 and 311-2. For example, the window parts 310-1 and 311-2 may be holes opened in the container 300, or may be made of a transparent material. The contents filled in the container 300 are not particularly limited, and may be, for example, ion exchange resin, activated carbon, synthetic adsorbent, zeolite, etc. At the inlet and outlet of the container 300, pressure gauges 400-1 and 400-2 are provided, respectively, for measuring the pressure of water in the flow path provided in the container 300. The pressure gauges 400-1 and 400-2 may be any pressure gauge that can calculate the pressure difference between the pressure of water before it flows into the container 300 and the pressure of water after it flows out of the container 300. The measurement method of the pressure gauges 400-1 and 400-2 is not particularly limited, and when a digital method is used, the pressure gauges 400-1 and 400-2 may transmit the measured digital values to a control device such as a PLC (Programmable Logic Controller) or to a computing device that performs machine learning processing. In this embodiment, the pressure gauges 400-1 and 400-2 transmit the measured values to the information processing device 200.
[0015] The camera 100 captures an image of the filling material filled in the container 300 through the windows 310-1 and 310-2. The camera 100 is preferably installed at a position where stable imaging can be performed, such as a position with little vibration or a position where the camera 100 can stably capture the filling material. A light source that irradiates light onto the filling material may be provided so that the camera 100 can capture a clearer image of the filling material. The type of light source is not limited, but a white LED (Light Emitting Diode) may be used. An infrared light source may also be used as the light source, and in that case, it is preferable to use an infrared-compatible camera 100. By using an infrared-compatible camera 100, it is possible to capture physical deterioration that cannot be captured by a visible light camera, and there is an advantage in that more feature quantities can be obtained.
[0016] Fig. 2 is a diagram showing an example of the internal configuration of camera 100 shown in Fig. 1. As shown in Fig. 2, camera 100 shown in Fig. 1 has an imaging unit 110 and a transmission unit 120. Note that Fig. 2 shows only main components related to this embodiment among the components of camera 100 shown in Fig. 1.
[0017] The imaging unit 110 captures an image of the filling material filled in the container 300. The imaging unit 110 may capture a still image or a video. There are no particular limitations on the image captured by the imaging unit 110, and it may be a color image or a grayscale image as long as it can capture physical deterioration of the filling material.
[0018] The transmission unit 120 transmits image data indicating an image captured by the imaging unit 110 to the information processing device 200. The image may be transmitted from the transmission unit 120 via a wired connection or wirelessly. There is no particular limitation on the communication standard used by the transmission unit 120 to transmit the image.
[0019] The information processing device 200 performs a predetermined process using an image transmitted from the camera 100. The information processing device 200 is not particularly limited, and may be a personal computer or a microcomputer as long as it is a device capable of performing machine learning, which will be described later, on the image transmitted from the camera 100. When the information processing device 200 is equipped with a graphics processing unit (GPU), high-speed calculation is possible. FIG. 3 is a diagram showing an example of the internal configuration of the information processing device 200 shown in FIG. 1. As shown in FIG. 2, the information processing device 200 shown in FIG. 1 has a trained model 210, an image data acquisition unit 220, a calculation unit 230, and an output unit 240. Note that FIG. 3 shows only the main components related to this embodiment among the components of the information processing device 200 shown in FIG. 1.
[0020] The image data acquisition unit 220 has a signal receiving function. The image data acquisition unit 220 uses the receiving function to acquire image data transmitted from the camera 100.
[0021] The calculation unit 230 inputs the image data acquired by the image data acquisition unit 220 into the trained model 210, which has undergone machine learning in advance, to calculate the state of the filler included in the image data. The trained model 210 is a model that has undergone learning of physical deterioration such as cracks and chips in the filler using machine learning in the learning phase. The machine learning algorithm used is not particularly limited, and any algorithm capable of detecting objects one by one may be used. Examples of the machine learning algorithm used include an instance segmentation algorithm such as Mask R-CNN and an object detection algorithm such as YOLO. The calculation unit 230 counts the number (A) of fillers included in the image data by inputting the image data acquired by the image data acquisition unit 220 into the trained model 210, which has undergone machine learning in advance. At this time, the calculation unit 230 identifies normal (not physically deteriorated) fillers included in the image data obtained from the trained model 210. The calculation unit 230 also identifies physically deteriorated fillers included in the image data obtained from the trained model 210. Then, the calculation unit 230 may count the number of specified normal (not physically deteriorated) fillings (B) and the number of physically deteriorated fillings (C), and count the sum (B+C) as the number of fillings included in the image data (A). Next, the calculation unit 230 calculates the ratio of the number of normal (not physically deteriorated) fillings to the total number of fillings included in the image data. The following (Equation 1) may be used for this calculation. ((AC) / A)×100[%] … (Formula 1) Here, the calculated ratio is defined as the sphericity ratio. The calculation unit 230 calculates the replacement time of the filler based on the sphericity ratio. Here, the trained model 210 is trained in advance to determine that a filler whose degree of deterioration exceeds a preset standard is a filler that has physically deteriorated. In addition, the trained model 210 is trained in advance to predict a predicted value of a change in deterioration of the filler based on the sphericity ratio and a replacement time according to the predicted value. The trained model 210 may be a model in which the calculation unit 230 predicts the sphericity ratio several months later using statistical processing such as regression analysis or machine learning such as a recurrent network for the sphericity ratio, and predicts the replacement time of the filler based on the predicted value. The trained model 210 may be a model that counts the number of fillers included in the image data and the number of fillers that have physically deteriorated, and learns the ratio of the number of fillers that have physically deteriorated to the number of fillers counted. Also, the trained model 210 may be trained to predict the change in deterioration of the filler based on the difference (differential pressure) between the pressure value transmitted from the pressure gauge 400-1 and the pressure value transmitted from the pressure gauge 400-2, and the sphericity. In this case, the calculation unit 230 inputs the image data acquired by the image data acquisition unit 220, the pressure value transmitted from the pressure gauge 400-1, and the pressure value transmitted from the pressure gauge 400-2 into the trained model 210 that has previously undergone machine learning, thereby calculating the replacement time of the filler included in the image data.
[0022] The output unit 240 outputs the result of the calculation performed by the calculation unit 230. At this time, the output unit 240 performs a predetermined notification when the ratio calculated by the calculation unit 230 falls below a preset threshold value (for example, 90%). When the calculation unit 230 calculates the time to replace the filling material, the output unit 240 outputs information indicating the calculated replacement time. Note that the output format of the information in the output unit 240 may be, for example, audio output, display, printing, or transmission of the information to another device, as long as the information can be recognized from the outside. In addition, the output of information and warning in the output unit 240 may be lighting of a warning light, output of an alarm sound, vibration, etc.
[0023] Fig. 4 is a diagram showing an example of a change in sphericity over the number of days of use learned by the trained model 210 shown in Fig. 3. As shown in Fig. 4, the trained model 210 shown in Fig. 3 has learned how the sphericity changes as the number of days of use of the filler passes. In addition, a threshold value is set in advance for the change in sphericity, and the replacement time is calculated based on the number of days that have passed when the changing sphericity falls below the threshold value.
[0024] An information processing method in the information processing device 200 shown in Fig. 1 will be described below. In the following description, a case will be described in which the filler filled in the container 300, which is an ion exchange resin tower, is ion exchange resin. Fig. 5 is a flowchart for explaining an example of an information processing method in the information processing device 200 shown in Fig. 1. Here, the trained model 210 has already learned in the machine learning learning phase that an ion exchange resin whose degree of deterioration, such as cracks or chips, exceeds a preset standard from image data of an ion exchange resin is an ion exchange resin that has physically deteriorated.
[0025] First, when the image data acquisition unit 220 acquires image data transmitted from the camera 100 (step S1), the calculation unit 230 inputs the image data acquired by the image data acquisition unit 220 to the trained model 210 (step S2). The calculation unit 230 acquires the state of the ion exchange resin output from the trained model 210 to calculate the total number of ion exchange resins included in the image data (step S3). Here, the calculation unit 230 identifies ion exchange resins that are physically deteriorated from the trained model 210, and calculates the total number of ion exchange resins by adding the number of normal ion exchange resins that are not physically deteriorated and the number of ion exchange resins identified as being physically deteriorated. Next, the calculation unit 230 calculates a sphericity ratio, which is the ratio of the number of normal ion exchange resins that are not physically deteriorated to the total number of ion exchange resins included in the image data (step S4). The above-mentioned (Equation 1) may be used to calculate this sphericity ratio. After calculating the sphericity ratio, the calculation unit 230 determines whether the calculated sphericity ratio is below a preset threshold value (step S5). If the sphericity is below the predetermined threshold, the output unit 240 outputs a predetermined notification (step S6).
[0026] In addition, the trained model 210 may have learned the sphericity, which is the ratio of normal ion exchange resins that are not physically deteriorated among the ion exchange resins included in the image data, during the learning phase of machine learning. In this case, the calculation unit 230 acquires the sphericity from the trained model 210, and the output unit 240 uses the sphericity acquired by the calculation unit 230 from the trained model 210 as a condition for outputting a predetermined notification.
[0027] 6 is a flowchart for explaining another example of the information processing method in the information processing device 200 shown in FIG. 1. Here, the trained model 210 has learned in the learning phase of machine learning that an ion exchange resin whose degree of deterioration such as cracking or chipping exceeds a preset standard from image data of an ion exchange resin is an ion exchange resin that has physically deteriorated. Furthermore, the trained model 210 has calculated a sphericity ratio, which is the ratio of normal ion exchange resin that has not physically deteriorated among the ion exchange resins contained in the image data, calculated a predicted value of the change in deterioration of the ion exchange resin based on the calculated sphericity ratio, and learned the replacement time based on the calculated predicted value.
[0028] First, when the image data acquisition unit 220 acquires image data transmitted from the camera 100 (step S11), the calculation unit 230 inputs the image data acquired by the image data acquisition unit 220 to the trained model 210 (step S12). The calculation unit 230 calculates the replacement time of the ion exchange resin (filler) by acquiring the replacement time output from the trained model 210 (step S13). Then, the output unit 240 outputs the replacement time calculated by the calculation unit 230 (step S14).
[0029] In this way, the filler filled in the resin tower is photographed using a camera through a window provided in the resin tower, image data captured by the camera is acquired, and the state of the filler is input to a trained model that has previously undergone machine learning based on the acquired image data, thereby acquiring the state of the filler and issuing a predetermined notification according to the acquired state. This makes it easy to recognize the state of the filler filled in the resin tower. In addition, by outputting the replacement time based on the image data of the captured filler, it is easy to recognize the replacement time of the filler. (Second embodiment)
[0030] FIG. 7 is a diagram showing a second embodiment of the information processing system of the present invention. As shown in FIG. 7, the information processing system in this embodiment has an information processing device 201 and a monitor 501. The container 300, the window parts 310-1 and 310-2 provided in the container 300, and the pressure gauges 400-1 and 400-2 are the same as those in the first embodiment. The position where the information processing device 201 is installed is preferably a position where stable imaging can be performed, such as a position with little vibration or a position where the information processing device 201 can stably capture the filling. Note that a light source that irradiates light onto the filling may be provided so that the information processing device 201 can capture the filling more clearly. The light source may be the same as that described in the first embodiment.
[0031] The information processing device 201 has the functions of the camera 100 in the first embodiment and the functions of the information processing device 200. FIG. 8 is a diagram showing an example of the internal configuration of the information processing device 201 shown in FIG. 7. As shown in FIG. 8, the information processing device 201 shown in FIG. 7 has an imaging unit 251, a trained model 210, an image data acquisition unit 220, a calculation unit 230, and an output unit 240. Note that FIG. 8 shows only the main components related to this embodiment among the components of the information processing device 201 shown in FIG. 7. The trained model 210, the image data acquisition unit 220, the calculation unit 230, and the output unit 240 are the same as those in the first embodiment.
[0032] The imaging unit 251 captures an image of the filling material filled in the container 300 through the windows 310-1 and 310-2. The imaging unit 251 may capture a still image or a video. The image captured by the imaging unit 251 is not particularly limited and may be a color image or a grayscale image as long as it can capture physical deterioration of the filling material. The image data acquisition unit 220 acquires the image data captured by the imaging unit 251 and performs the same process as the process performed by the image data acquisition unit 220 in the first embodiment.
[0033] The information processing device 201 may be a personal computer or a microcomputer, as long as the imaging unit 251 captures an image and the captured image is subjected to machine learning using the trained model 210. If the information processing device 201 is provided with a graphics processing unit (GPU), high-speed calculation is possible.
[0034] The monitor 501 may be, for example, a display of a mobile terminal such as a smartphone, or may be a device capable of displaying general information such as a display of a personal computer. The monitor 501 is connected to the information processing device 201 wirelessly or via a wire, and displays information transmitted from the information processing device 201 (for example, a notification screen, a warning screen, etc.).
[0035] An information processing method in the information processing device 201 shown in Fig. 7 will be described below. In the following description, a case will be described in which the filler filled in the container 300, which is an ion exchange resin tower, is ion exchange resin. Fig. 9 is a flowchart for explaining an example of an information processing method in the information processing device 201 shown in Fig. 7. Here, the trained model 210 has already learned in the machine learning learning phase that an ion exchange resin whose degree of deterioration, such as cracks or chips, exceeds a preset standard from image data of an ion exchange resin is an ion exchange resin that has physically deteriorated.
[0036] First, the imaging unit 251 captures an image of the ion exchange resin filled in the container 300 through the window unit 310-1 or 310-2 (step S21). Then, the image data acquisition unit 220 acquires the image data captured by the imaging unit 251 (step S22). Next, the calculation unit 230 inputs the image data acquired by the image data acquisition unit 220 to the trained model 210 (step S23). The calculation unit 230 calculates the total number of ion exchange resins included in the image data by acquiring the state of the ion exchange resin output from the trained model 210 (step S24). Here, the calculation unit 230 identifies ion exchange resins that are physically deteriorated from the trained model 210, and calculates the total number of ion exchange resins by adding the number of normal ion exchange resins that are not physically deteriorated and the number of ion exchange resins identified as being physically deteriorated. Next, the calculation unit 230 calculates a sphericity ratio, which is the ratio of the number of normal ion exchange resins that are not physically deteriorated to the total number of ion exchange resins included in the image data (step S25). The sphericity may be calculated using the above-mentioned (Equation 1). After calculating the sphericity, the calculation unit 230 determines whether the calculated sphericity is below a preset threshold (step S26). If the sphericity is below the predetermined threshold, the output unit 240 outputs a predetermined notification (step S27).
[0037] Note that the trained model 210 may have learned the sphericity, which is the ratio of ion exchange resins that have physically deteriorated among the ion exchange resins included in the image data, during the learning phase of machine learning. In this case, the calculation unit 230 acquires the sphericity from the trained model 210, and the output unit 240 uses the sphericity acquired by the calculation unit 230 from the trained model 210 as a condition for outputting a predetermined notification.
[0038] In this way, the filler filled in the resin tower is imaged through a window provided in the resin tower, and the image data is input to a trained model that has previously learned the state of the filler based on the image data, thereby acquiring the state of the filler and issuing a predetermined notification according to the acquired state. This makes it easy to recognize the state of the filler filled in the resin tower. (Third embodiment)
[0039] Fig. 10 is a diagram showing a third embodiment of the information processing system of the present invention. As shown in Fig. 10, the information processing system in this embodiment has an electron microscope 101 and an information processing device 200. The electron microscope 101 captures an image of the filler that is separated and collected from the resin tower after the filler filled in the resin tower is separated and collected. The electron microscope 101 has a communication function and transmits the captured image data to the information processing device 200. The information processing device 200 is the same as that in the first embodiment. (Example of creating and validating a trained model)
[0040] A Keyence VHX-6000 microscope was used as the imaging means to capture images of several types of ion exchange resins separated and collected from the ion exchange resin tower. A portion of the image data was extracted from the captured image data to create model validation data, which is intended to verify the accuracy of a trained model created using machine learning. In addition, image data other than the model validation data was padded, and the padded image data was randomly divided into a training dataset and a validation dataset in a ratio of 7:3. Each ion exchange resin with physical deterioration such as cracks and chips was annotated for this divided dataset to create a dataset for creating a trained model. In order to detect physical deterioration such as cracks and chips one by one, a trained model was created using the annotated dataset using MaskRCNN (licensed under The MIT License) as a machine learning algorithm. When the model validation data was inferred from the created trained model, physical deterioration such as cracks and chips could be detected one by one with high accuracy. Other devices that can be used include those that cut out the spheres (circles) that make up the filler, and those that detect cracks.
[0041] In addition, in the case of a system in which the container 300 is configured redundantly and an active system and a standby system are provided, the operating system may be automatically switched from the active system to the standby system after a notification from the output unit 240. In this way, the system can be operated efficiently.
[0042] Although the above description has been given by allocating each function (processing) to each component, this allocation is not limited to the above. Furthermore, the configuration of the components is also not limited to the above-described form, and the above-described form is merely an example. Furthermore, a combination of each embodiment may be used. Furthermore, a plurality of trained models according to the environment such as the installation location of the container 300 and the weather may be prepared in advance, and the trained model to be used may be switched according to the environment at the time of measurement.
[0043] The above-mentioned processing performed by the information processing devices 200 and 201 may be performed by logic circuits that are created according to the purpose. Also, a computer program (hereinafter referred to as a program) in which the processing contents are described as a procedure may be recorded on a recording medium that can be read by the information processing devices 200 and 201, and the program recorded on the recording medium may be read and executed by the information processing devices 200 and 201. The recording medium that can be read by the information processing devices 200 and 201 refers to a removable recording medium such as a floppy (registered trademark) disk, a magneto-optical disk, a DVD (Digital Versatile Disc), a CD (Compact Disc), a Blu-ray (registered trademark) Disc, and a USB (Universal Serial Bus) memory, as well as a memory such as a ROM (Read Only Memory), a RAM (Random Access Memory), and a HDD (Hard Disc Drive) that are built into the information processing devices 200 and 201. The program recorded on the recording medium is read by a CPU provided in the information processing devices 200 and 201, and the same processing as that described above is performed under the control of the CPU. Here, the CPU operates as a computer that executes a program read from a recording medium on which the program is recorded. [Explanation of symbols]
[0044] 100 Cameras 101 Electron Microscope 110,251 Imaging section 120 Transmitter 200, 201 Information processing device 210 trained models 220 Image data acquisition unit 230 Calculation Unit 240 Output section 300 containers 310-1, 310-2 Window section 400-1,400-2 Pressure Gauge 501 Monitor
Claims
1. an image data acquisition unit that acquires image data showing an image of a filling material filled in a container into which water flows and from which treated water flows out; A calculation unit that calculates the state of the filling material contained in the image data by inputting the image data acquired by the image data acquisition unit into a trained model that has undergone machine learning; an output unit that outputs a result of the calculation performed by the calculation unit; The calculation unit uses the trained model to identify fillers whose degree of deterioration exceeds a standard among the fillers included in the image data, predicts a change in deterioration of the fillers based on a ratio of the number of the identified fillers whose degree of deterioration exceeds a standard, and calculates a replacement time for the fillers based on the predicted change. The output unit is an information processing device that outputs the replacement time calculated by the calculation unit.
2. 2. The information processing device according to claim 1, The output unit is an information processing device that performs a predetermined notification when the ratio calculated by the calculation unit exceeds a preset threshold value.
3. 3. The information processing device according to claim 1, The calculation unit is an information processing device that uses the trained model to predict changes in deterioration of the filling material based on the differential pressure between the water pressures at the inlet and outlet of the container and the calculated ratio.
4. 4. The information processing device according to claim 1, The data processing device, wherein the filler is an ion exchange resin.
5. A camera and an information processing device are included. The camera includes: An imaging unit that captures an image of a filling material filled in a container into which water flows and from which treated water flows out; a transmission unit that transmits image data representing an image captured by the imaging unit to the information processing device, The information processing device includes: an image data acquisition unit that acquires the image data transmitted from the transmission unit; A calculation unit that calculates the state of the filling material contained in the image data by inputting the image data acquired by the image data acquisition unit into a trained model that has undergone machine learning; an output unit that outputs a result of the calculation performed by the calculation unit; The calculation unit uses the trained model to identify fillers whose degree of deterioration exceeds a standard among the fillers included in the image data, predicts a change in deterioration of the fillers based on a ratio of the number of the identified fillers whose degree of deterioration exceeds a standard, and calculates a replacement time for the fillers based on the predicted change. The output unit is an information processing system that outputs the replacement time calculated by the calculation unit.
6. 6. The information processing system according to claim 5, The container has a window, The camera is an information processing system that captures an image of the filling material filled in the container through the window portion.
7. A process for acquiring image data showing an image of a filling material filled in a container into which water flows and from which treated water flows out; A process of calculating a state of the filling material contained in the image data by inputting the acquired image data into a trained model that has undergone machine learning; A process of identifying, among the fillings included in the image data, fillings whose degree of deterioration exceeds a standard using the trained model; A process of predicting a change in deterioration of the packing based on a ratio of the number of packings whose degree of deterioration exceeds a standard using the trained model; A process of calculating a replacement time of the packing based on the predicted change; and outputting the calculated replacement time.
8. On the computer, A step of acquiring image data showing an image of a filling material filled in a container into which water flows and from which treated water flows out; A step of calculating a state of the filling material contained in the image data by inputting the acquired image data into a trained model that has undergone machine learning; Using the trained model, a step of identifying a filler whose degree of deterioration exceeds a standard among the fillers included in the image data; A step of predicting a change in deterioration of the packing based on a ratio of the number of packings whose degree of deterioration exceeds a standard using the trained model; calculating a replacement time for the packing based on the predicted change; and a step of outputting the calculated replacement time.
Citation Information
Patent Citations
Method and apparatus for detecting boundary surface of ion exchange resins
JP1992083538A
Evaluation method for performance of cation exchange resin and management method of water treatment system using it
JP2001228134A
Apparatus, program, and method for management of reuse separation membrane
JP2011212608A
Information processing device
WO2019059310A1
Performance evaluation system, performance evaluation method, program, and learned model
WO2022080416A1
