Water level detection device, water level detection method, and water level detection program

Abstract

To provide a water level detection device, water level detection method, and water level detection program for accurately detecting the water level.SOLUTION: A water level detection device is provided, comprising an acquisition unit for acquiring a captured image of a staff gauge, multiple detection units for detecting water levels from the captured image on the basis of different color characteristics, and a selection unit for selecting a specific water level form among multiple water levels detected by the multiple detection units on the basis of color characteristics of the captured image.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present disclosure relates to a water level detection device, a water level detection method, and a water level detection program. [Background technology]

[0002] 2. Description of the Related Art In recent years, water level detection devices have been proposed that take an image of a water gauge installed in a river, sea, pond, or the like with a camera and detect the water level from the image.

[0003] For example, Patent Document 1 proposes a water level measurement method in which a tracking process is performed from the calculation start point along the boundary line of the side of a water measuring plate cut out from an input image, and the position where the boundary line disappears is determined to be the water surface position. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 9-161076 Summary of the Invention [Problem to be solved by the invention]

[0005] However, the water level measurement method of Patent Document 1 has difficulty in accurately detecting the water level when images with different color characteristics are acquired.

[0006] The present disclosure aims to provide a water level detection device, a water level detection method, and a water level detection program that accurately detect the water level. [Means for solving the problem]

[0007] The water level detection device of the present disclosure comprises an acquisition unit that acquires a photographed image of a water gauge, a plurality of detection units that detect the water level from the photographed image based on different color characteristics, and a selection unit that selects a specific water level from the plurality of water levels detected by the plurality of detection units based on the color characteristics of the photographed image.

[0008] The water level detection method of the present disclosure includes acquiring a photographed image of a water gauge, detecting a water level from the photographed image based on different color characteristics, and selecting a specific water level from the detected water levels based on the color characteristics of the photographed image.

[0009] The water level detection program of the present disclosure includes acquiring a photographed image of a water gauge, detecting the water level from the photographed image based on different color characteristics, and selecting a specific water level from the detected multiple water levels based on the color characteristics of the photographed image.

[0010] The water level detection device of the present disclosure comprises an acquisition unit that acquires a photographed image of a water gauge, and a selection unit that selects a specific detection unit from among a plurality of detection units that detect water levels based on different color characteristics according to the color characteristics of the photographed image, and causes the specific detection unit to detect the water level of the photographed image. [Effects of the Invention]

[0011] According to the present disclosure, the water level can be detected accurately. [Brief explanation of the drawings]

[0012] [Figure 1] FIG. 1 is a diagram showing an overview of the first embodiment. [Figure 2] FIG. 2 is a diagram illustrating the functional configuration of the water level detection device according to the first embodiment. [Figure 3] FIG. 3 is a diagram illustrating the hardware configuration of the water level detection device according to the first embodiment. [Figure 4] FIG. 4 is a flowchart showing the water level detection process of the first embodiment. [Figure 5] 5A and 5B are diagrams showing how the water level is stored in the storage unit in the first embodiment. [Figure 6] FIG. 6 is a diagram showing a time series graph of water level. [Figure 7]FIG. 7 is a diagram showing an outline of the second embodiment. [Figure 8] FIG. 8 is a diagram showing the functional configuration of the water level detecting device according to the second embodiment. [Figure 9] FIG. 9 is a flowchart showing the water level detection process of the second embodiment. [Figure 10] FIG. 10 is a diagram showing how the water level is stored in the storage unit in the second embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0013] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.

[0014] First Embodiment [overview] Conventionally, if an image of a water gauge captured by a camera has different color characteristics, the water level may not be accurately detected. For example, a water gauge may be captured using visible light during the day and infrared light at night when there is insufficient light. In this case, visible light capture produces a color image, while infrared capture produces a monochrome image. Generally, water level detection devices detect water levels based on specific color characteristics, such as color images. Therefore, if an image captured with different color characteristics is input, the water level may not be accurately detected. Furthermore, because the spectral absorptance of water for infrared light is higher than that of visible light, an image captured with infrared light exhibits different shading of the water surface compared to an image captured with visible light. Therefore, if a water level detection device is configured to detect water levels based on shading of a color image, the water level may not be accurately detected from an image captured with infrared light.

[0015] Therefore, as shown in FIG. 1, when a water gauge M installed to indicate the water level S of a river, for example, is photographed by a photographing device C, the water level detection device 1 acquires the photographed image I. The water level detection device 1 detects multiple water levels L1 and L2 from the photographed image I based on different color characteristics. For example, the water level detection device 1 may detect the water level L1 from the photographed image I based on color characteristics consisting of three or more colors, and detect the water level L2 from the photographed image I based on color characteristics consisting of two monochrome colors. Then, the water level detection device 1 selects a specific water level from the water levels L1 and L2 based on the color characteristics of the photographed image I. For example, if the photographed image I is a monochrome image, the water level detection device 1 may select the water level L2 from the water levels L1 and L2. When the water level detection device 1 selects the specific water level L2, it may transmit the specific water level L2 to the user's terminal T.

[0016] [Water level detection device] Next, the functional configuration of the water level detection device 1 according to the present disclosure will be described. As shown in Fig. 2, the water level detection device 1 has an acquisition unit 2, a first detection unit 3, a second detection unit 4, a storage unit 5, a communication unit 6, and a selection unit 7. The acquisition unit 2 is connected to each of the first detection unit 3, the second detection unit 4, and the selection unit 7, and the first detection unit 3 and the second detection unit 4 are connected to the communication unit 6 via the storage unit 5. In addition, the selection unit 7 is connected to each of the storage unit 5 and the communication unit 6.

[0017] The acquisition unit 2 acquires a photographed image I of a water gauge M. For example, the acquisition unit 2 may be connected to a photographing device C that photographs the water gauge M, and acquire the photographed image I from the photographing device C. The acquisition unit 2 may acquire the photographed image I, for example, at predetermined time intervals (for example, every 10 minutes). Here, the photographing device C may be composed of a camera or the like. Furthermore, the photographing device C may be configured to acquire photographed images I having different color characteristics.

[0018] For example, the image capture device C may have both visible light and infrared light capture functions, and may be configured to capture visible light images during the day when the ambient light is bright, and capture infrared images by irradiating infrared light at night when the ambient light is dark. The image capture device C captures a color image I when capturing visible light, and a monochrome image I when capturing infrared light. Note that the monochrome image I may be composed of two colors and is not limited to black and white. For example, the monochrome image I may be composed of blue and red, or may be composed of grayscale.

[0019] The photographing device C may be composed of a plurality of photographing devices, including one dedicated to visible light photographing and one dedicated to infrared photographing. The photographed image I may be either a still image or a video.

[0020] The first detector 3 and the second detector 4 detect the water levels L1 and L2 from the captured image I based on different color characteristics. That is, the first detector 3 and the second detector 4 constitute the multiple detectors of the present disclosure.

[0021] For example, the first detection unit 3 may detect the water level L1 based on color characteristics of three or more colors. Specifically, the first detection unit 3 may have a first detection model trained with training data including a data set of an image composed of three or more colors and a water level indicated by a water gauge M. The first detection unit 3 may then input the photographed image I acquired by the acquisition unit 2 into the first detection model, thereby detecting the water level L1 of the water gauge M in the photographed image I. Here, the first detection model may be realized by, for example, a convolutional neural network.

[0022] The second detection unit 4 may detect the water level L2 based on a monochrome color characteristic composed of two colors. Specifically, the second detection unit 4 may have a second detection model trained with training data including a data set of an image composed of two colors and the water level indicated by the water gauge M. The second detection unit 4 may then input the photographed image I acquired by the acquisition unit 2 into the second detection model, thereby detecting the water level L2 of the water gauge M in the photographed image I. Here, the second detection model may be realized by, for example, a convolutional neural network. Note that the monochrome color characteristic may be a color characteristic composed of two colors and is not limited to a color characteristic composed of black and white. For example, the monochrome color characteristic may be a color characteristic composed of blue and red, or a color characteristic composed of grayscale.

[0023] The storage unit 5 stores the water levels L1 and L2 detected by the first detection unit 3 and the second detection unit 4. The storage unit 5 may be configured from, for example, a database, a cloud, or the like.

[0024] The communication unit 6 communicates with a user's terminal T. The communication unit 6 may communicate with the terminal T via a communication network in a wired or wireless manner, for example.

[0025] The selection unit 7 selects a specific water level from the multiple transitions L1 and L2 detected by the first detection unit 3 and the second detection unit 4 based on the color characteristics of the captured image I. For example, the selection unit 7 may determine the color characteristics of the captured image I based on the difference in intensity of the colors included in the captured image I. When the selection unit 7 selects the specific water level, it may update the multiple transitions L1 and L2 stored in the memory unit 5 to the specific water level. In addition, the selection unit 7 may control the communication unit 6 to notify the user's terminal T of the water level with the selected characteristics.

[0026] Next, we will explain in detail the hardware configuration of the water level detection device 1. For example, as shown in Fig. 3, the water level detection device 1 has a storage device 8, a processor 9, a user interface (UI) device 10, and a communication device 11, which are interconnected via a bus B.

[0027] The programs or instructions for realizing the various functions and processes of the water level detection device 1 may be downloaded from an external device via a network, etc. Also, the programs or instructions for realizing the various functions and processes of the water level detection device 1 may be provided from a removable storage medium such as a CD-ROM (Compact Disk-Read Only Memory) or flash memory.

[0028] The storage device 8 may be realized by one or more non-transitory storage media, such as random access memory, flash memory, or a hard disk drive, and may store installed programs or instructions as well as files or data used in the execution of the programs or instructions.

[0029] The processor 9 may be realized by one or more central processing units (CPUs), graphics processing units (GPUs), processing circuitry, etc., each of which may be configured with one or more processor cores. The processor 9 executes various functions and processes of the water level detection device 1 in accordance with data such as programs, instructions, programs, or parameters (e.g., parameters required to execute instructions) stored in the storage device 8.

[0030] The user interface device 10 may be composed of input devices such as a keyboard, mouse, camera, or microphone, output devices such as a display, speaker, headset, or printer, and input / output devices such as a smartphone, tablet, or touch panel.

[0031] The communication device 11 is realized by various communication circuits that execute wired and / or wireless communication processing with an external device, the Internet, a LAN (Local Area Network), a cellular network, or other communication network.

[0032] It should be noted that the above-described hardware configuration is merely an example, and the water level detection device 1 according to the present disclosure may be realized by any other appropriate hardware configuration.

[0033] [Water level detection process] Next, the water level detection process performed by the water level detection device 1 will be described with reference to the flowchart shown in FIG.

[0034] First, in step S1, the acquisition unit 2 acquires a photographed image I of a water gauge M. For example, when the photographing device C photographs a water gauge M installed in a river or the like, it transmits the photographed image I to the acquisition unit 2 at predetermined time intervals. This allows the acquisition unit 2 to acquire the photographed image I at predetermined time intervals. The acquisition unit 2 outputs the acquired photographed image I to the first detection unit 3 and the second detection unit 4, respectively.

[0035] When the first detection unit 3 and the second detection unit 4 each receive the captured image I, they detect the water levels L1 and L2 from the captured image I based on the different color characteristics in step S2.

[0036] For example, the first detection unit 3 may detect the water level L1 of the water gauge M in the photographed image I by inputting the photographed image I into a first detection model that detects the water level L1 based on the color characteristics of an image consisting of three or more colors. On the other hand, the second detection unit 4 may detect the water level L2 of the water gauge M in the photographed image I by inputting the photographed image I into a second detection model that detects the water level L2 based on the color characteristics of an image consisting of two colors.

[0037] In this way, the first detection unit 3 detects the water level L1 based on color characteristics consisting of three or more colors, and the second detection unit 4 detects the water level L2 based on color characteristics consisting of two colors. Therefore, the first detection unit 3 and the second detection unit 4 can detect the water level in the captured image I from different color characteristics.

[0038] Furthermore, the first detector 3 has a first detection model, and the second detector 4 has a second detection model, so the first detector 3 and the second detector 4 can easily detect the water levels L1 and L2 in the captured image I.

[0039] The first detection unit 3 and the second detection unit 4 store the detected water levels L1 and L2 in the memory unit 5. For example, as shown in FIG. 5A, when the first detection unit 3 and the second detection unit 4 detect new water levels L1a and L2a, they may input the water levels L1a and L2a into the water level list 12 stored in the memory unit 5. In this way, the memory unit 5 stores the two water levels L1 and L2 detected by the first detection unit 3 and the second detection unit 4, so the selection unit 7 can easily select a specific water level.

[0040] Next, in step S3, the selection unit 7 selects a specific water level from the two water levels L1 and L2 detected by the first detection unit 3 and the second detection unit 4 based on the color characteristics of the captured image I acquired by the acquisition unit 2.

[0041] For example, the selection unit 7 determines the color characteristics of the captured image I based on the difference in intensity of colors contained in the captured image I. Specifically, the selection unit 7 may sample pixels of the captured image I, and determine that the captured image I is a color image composed of three or more colors if the difference between the maximum and minimum values of RGB values (for example, in the range of 0 to 255) indicating the intensity of red, green, and blue is equal to or greater than a predetermined threshold. On the other hand, the selection unit 7 may determine that the captured image I is a monochrome image composed of two colors if the difference between the maximum and minimum values of the RGB values is less than a predetermined threshold. This allows the selection unit 7 to easily determine the color characteristics of the captured image I.

[0042] Next, if the selection unit 7 determines that the captured image I is a color image, it selects from the two water levels L1 and L2 the specific water level L1 detected by the first detection unit 3 that corresponds to its color characteristics. On the other hand, if the selection unit 7 determines that the captured image I is a monochrome image, it selects from the two water levels L1 and L2 the specific water level L2 detected by the second detection unit 4 that corresponds to its color characteristics.

[0043] In this way, the first detection unit 3 and the second detection unit 4 detect the water levels L1 and L2 from the captured image I based on their mutually different color characteristics. Then, the selection unit 7 selects a specific water level from the multiple water levels L1 and L2 detected by the first detection unit 3 and the second detection unit 4 based on the color characteristics of the captured image I. This allows the water level detection device 1 to accurately detect the water level indicated by the water gauge M. Furthermore, whether the captured image I is color or monochrome (whether it is taken during the day or at night), the water level detection device 1 can select the specific water level detected by the appropriate detection unit, and can determine the water level with high accuracy. Furthermore, there is no need to prepare any special equipment, and the water level can be detected using a general water gauge M.

[0044] Here, it is assumed that the selection unit 7 selects the specific water level L2 detected by the second detection unit 4. When the selection unit 7 selects the specific water level L2, it may update the two water levels L1 and L2 stored in the storage unit 5 to the specific water level L2. For example, as shown in FIG. 5B, the selection unit 7 may overwrite the water level L1a in the water level list 12 with the water level L2a. As a result, the selection unit 7 determines the water level at 18:00 to be the water level L2a. In this way, when the selection unit 7 selects the specific water level L2a, it updates the two water levels L1a and L2a stored in the storage unit 5 to the specific water level L2a, thereby storing the accurate water level L2a and preventing the water level L1a from being output. Note that the selection unit 7 is not limited to overwriting with the water level L2 as long as it can update to the water level L2a. For example, the selection unit 7 may provide a new determination field for inputting the determined water level and input the final water level L2a into the determination field.

[0045] The selection unit 7 may also output the selected specific water level L2a to the user. For example, as shown in FIG. 6, the selection unit 7 may create a time series graph 13 of water levels including the specific water level L2a based on the water level list 12, and output the time series graph 13 to the user in response to an output request from the user. The selection unit 7 may also control the communication unit 6 to transmit the time series graph 13 to the terminal T when an output request is sent from the user's terminal T. This allows the user to check the detected specific water level L2a without being aware of whether the captured image I is in color or monochrome.

[0046] The selection unit 7 may also automatically output the selected specific water level L2a to the user. For example, the selection unit 7 may control the communication unit 6 to notify the user of information based on the specific water level L2a (e.g., the water level list 12, whether or not there is an abnormality in the water level, etc.) a predetermined number of times within a predetermined period. The selection unit 7 may also control the communication unit 6 to notify the user's terminal T if the specific water level L2a is outside a predetermined range. For example, if the predetermined range is set based on an abnormality in the water level, the selection unit 7 may notify the user's terminal T of the abnormality in the water level. This allows the user to easily notice changes in the water level.

[0047] According to this embodiment, the first detection unit 3 and the second detection unit 4 detect the water levels L1 and L2 from the captured image I based on their different color characteristics. Then, the selection unit 7 selects a specific water level from the multiple water levels L1 and L2 detected by the first detection unit 3 and the second detection unit 4 based on the color characteristics of the captured image I. This allows the water level detection device 1 to accurately detect the water level indicated by the water gauge M.

[0048] Second Embodiment [overview] As shown in FIG. 7, when the water level detection device 1 acquires a photographed image I, it selects a specific detection unit from among a plurality of detection units according to the color characteristics of the photographed image I. Here, the multiple detection units detect the water level based on different color characteristics. When the water level detection device 1 selects a specific detection unit, it causes the specific detection unit to detect the water level L of the photographed image I. The water level detection device 1 may transmit the detected water level L to the user's terminal T.

[0049] [Water level detection device] Next, the functional configuration of the water level detection device 1 according to the present disclosure will be described. As shown in Fig. 8, the water level detection device 1 may have a selection unit 7 disposed between the acquisition unit 2 and the first and second detection units 3 and 4. The selection unit 7 selects a specific detection unit corresponding to the captured image I from the first and second detection units 3 and 4, which detect water levels based on different color characteristics. The selection unit 7 then causes the selected specific detection unit to detect the water level L in the captured image I.

[0050] [Water level detection process] Next, the water level detection process performed by the water level detection device 1 will be described with reference to the flowchart shown in FIG.

[0051] First, in step S11, the acquisition unit 2 acquires a captured image I of a water gauge M, as in the first embodiment. Here, the acquisition unit 2 outputs the acquired captured image I to the selection unit 7.

[0052] When the selection unit 7 inputs the captured image I, in step S12, it selects a specific detection unit from the first detection unit 3 and the second detection unit 4, which detect the water level L based on different color characteristics, that corresponds to the color characteristics of the captured image I.

[0053] For example, the selection unit 7 may determine the color characteristics of the captured image I based on the difference in intensity of colors included in the captured image I. Specifically, the selection unit 7 may sample pixels of the captured image I, and if the difference between the maximum and minimum values of RGB values (for example, in the range of 0 to 255) indicating the intensity of red, green, and blue is equal to or greater than a predetermined threshold, determine that the captured image I is a color image composed of three or more colors. On the other hand, if the difference between the maximum and minimum values of the RGB values is less than a predetermined threshold, the selection unit 7 may determine that the captured image I is a monochrome image composed of two colors. This allows the selection unit 7 to easily determine the color characteristics of the captured image I.

[0054] If the selection unit 7 determines that the captured image I is a color image, it selects the first detection unit 3 as a specific detection unit according to the color characteristics of the captured image I. On the other hand, the selection unit 7 selects the second detection unit 4 as a specific detection unit according to the color characteristics of the captured image I. Note that the configurations of the first detection unit 3 and the second detection unit 4 are the same as those in the first embodiment, and therefore description thereof will be omitted.

[0055] When the selection unit 7 selects a specific detection unit, in step S13, the selection unit 7 causes the specific detection unit to detect the water level L in the captured image I. For example, when the selection unit 7 selects the first detection unit 3 as the specific detection unit, the selection unit 7 transmits the captured image I to the first detection unit 3. When the first detection unit 3 receives the captured image I, it detects the water level L from the captured image I based on the color characteristics of the color image composed of three or more colors. The first detection unit 3 stores the detected water level L in the memory unit 5. For example, as shown in FIG. 10 , when the first detection unit 3 detects the water level L, it may input the water level L into a water level list 12 stored in the memory unit 5.

[0056] On the other hand, when the selection unit 7 selects the second detection unit 4 as the specific detection unit, it transmits the captured image I to the second detection unit 4. When the second detection unit 4 receives the captured image I, it detects the water level L from the captured image I based on the color characteristics of the monochrome image composed of two colors. The second detection unit 4 stores the detected water level L in the memory unit 5.

[0057] In this way, the selection unit 7 selects a specific detection unit from the first detection unit 3 and the second detection unit 4, which detect the water level L based on different color characteristics, according to the color characteristics of the captured image I. The selection unit 7 then causes the specific detection unit to detect the water level L in the captured image I. This allows the water level detection device 1 to accurately detect the water level L indicated by the water gauge M. Furthermore, because the water level L is detected by either the first detection unit 3 or the second detection unit 4, the number of times the first detection unit 3 and the second detection unit 4 are used can be reduced, and detection costs can be reduced.

[0058] When a specific detection unit detects the water level L, the selection unit 7 may output the water level L to the user, as in the first embodiment. For example, as shown in FIG. 6 , the selection unit 7 may create a time series graph 13 of the water level L based on the water level list 12, and output the time series graph 13 to the user in response to an output request from the user. Furthermore, when an output request is transmitted from the user's terminal T, the selection unit 7 may control the communication unit 6 to transmit the time series graph 13 to the terminal T. This allows the user to check the detected water level L without being aware of whether the captured image I is in color or monochrome.

[0059] Furthermore, the selection unit 7 may automatically output the detected water level L to the user. For example, if the detected water level L is outside a predetermined range, the selection unit 7 may control the communication unit 6 to notify the user's terminal T. For example, if the predetermined range is set based on an abnormality in the water level, the selection unit 7 may notify the user's terminal T of the abnormality in the water level L. This allows the user to easily notice a change in the water level L.

[0060] According to this embodiment, the selection unit 7 selects a specific detection unit from the first detection unit 3 and the second detection unit 4, which detect the water level L based on different color characteristics, according to the color characteristics of the captured image I. Then, the selection unit 7 causes the specific detection unit to detect the water level L in the captured image I. Therefore, the water level detection device 1 can accurately detect the water level L indicated by the water gauge M.

[0061] Although specific examples of the present disclosure have been described in detail above, these are merely examples and do not limit the scope of the claims. The technology described in the claims includes various modifications and alterations of the specific examples exemplified above. [Industrial Applicability]

[0062] The water level detection device according to the present disclosure can be used as a device for detecting the water level indicated by a water gauge M. [Explanation of symbols]

[0063] 1 Water level detection device 2 Acquisition part 3 First detection unit 4 Second detection unit 5 Storage section 6. Communications Department 7 Selection section 8 Storage device 9 processors 10 User Interface Device 11. Communications equipment 12 Water Level List 13 Time Series Graph C. Imaging device I Photographed image L,L1,L1a,L2,L2a Water level M water mark S water surface T-Terminal

Claims

1. an acquisition unit that acquires a photographed image of a water gauge; a plurality of detectors for detecting water levels from the captured image based on mutually different color characteristics; A water level detection device comprising: a selection unit that selects a specific water level from the multiple water levels detected by the multiple detection units based on the color characteristics of the captured image.

2. The water level detection device according to claim 1, wherein the plurality of detection units include a first detection unit that detects the water level based on the color characteristics consisting of three or more colors, and a second detection unit that detects the water level based on the color characteristics consisting of two colors.

3. the first detection unit has a first detection model trained with training data including a data set of an image composed of three or more colors and a water level indicated by the water gauge; The water level detection device according to claim 2 , wherein the second detection unit has a second detection model trained with training data including a data set of an image composed of two colors and a transition indicated by the water level gauge.

4. The water level detection device according to claim 1 , further comprising a storage unit that stores a plurality of water levels detected by the plurality of detection units.

5. The water level detection device according to claim 4 , wherein when the selector selects the specific water level, the selector updates the plurality of water levels stored in the storage unit to the specific water level.

6. The water level detection device according to claim 1 , wherein the selection unit outputs the specific water level in response to an output request from a user.

7. Further comprising a communication unit for communicating with a user terminal, The water level detection device according to claim 1 , wherein the selection unit controls the communication unit to notify the terminal when the specific water level is outside a predetermined range.

8. The water level detection device according to claim 1 , wherein the selection unit determines the color characteristics based on a difference in intensity of colors included in the captured image.

9. Acquiring a photographed image of a water gauge; Detecting a water level from the captured image based on the different color characteristics; and selecting a particular water level from the plurality of detected water levels based on color characteristics of the captured image.

10. Acquiring a photographed image of a water gauge; Detecting a water level from the captured image based on the different color characteristics; and selecting a specific water level from the plurality of detected water levels based on color characteristics of the captured image.

11. an acquisition unit that acquires a photographed image of a water gauge; A water level detection device comprising: a selection unit that selects a specific detection unit according to the color characteristics of the captured image from among a plurality of detection units that detect water levels based on mutually different color characteristics, and causes the specific detection unit to detect the water level of the captured image.

12. The water level detection device according to claim 11, wherein the plurality of detection units include a first detection unit that detects the water level based on the color characteristics consisting of three or more colors, and a second detection unit that detects the water level based on the color characteristics consisting of two colors.

13. the first detection unit has a first detection model trained with training data including a data set of an image composed of three or more colors and a water level indicated by the water gauge; The water level detection device according to claim 12, wherein the second detection unit has a second detection model trained with training data including a data set of an image composed of two colors and a transition indicated by the water level gauge.

Images