JP2024040830A - Turbidity estimation apparatus, turbidity estimation system, drug injection control system, turbidity estimation method, and program - Google Patents

Abstract

To efficiently identify turbidity of target water early.SOLUTION: A turbidity estimation apparatus includes: a first information generation unit which generates, by turbidity, first information indicating a relationship between depth of water and predetermined features on the basis of a plurality of captured images different in depth for each turbidity, obtained by imaging multiple kinds of water different in turbidity from above the water surface for each of different depths of water; a second information generation unit which generates second information indicating a relationship between the depth of water and the features on the basis of a plurality of captured images different in depth obtained by imaging water to be subjected to turbidity estimation from above the water surface for each of the different depths of water; and a turbidity estimation unit which estimates turbidity of the water to be subjected to the turbidity estimation, using the first information for each turbidity and the second information.SELECTED DRAWING: Figure 8

Description

Embodiments of the present invention relate to a turbidity estimation device, a turbidity estimation system, a drug injection control system, a turbidity estimation method, and a program.

Generally, at water treatment plants in cities, inspectors and others regularly go to the upstream dams, rivers, and other water sources to sample water and conduct water quality tests (turbidity measurements, etc.).

Kazuhide Igarashi, 6 others, "Basic study for turbidity measurement by image analysis", 2017, Summary collection of the 2017 Sabo Society research presentation, p. 726-727

The above-mentioned water sampling and water quality testing (turbidity measurement, etc.) are not always carried out on site, so if, for example, a certain level of turbidity occurs at a water source, the situation can be quickly detected at a water treatment plant, etc. It is difficult to grasp the situation, which delays the preparation of chemicals necessary for water purification and water purification operations.

In addition, in order to understand the turbidity status of water sources, people must go to the site regularly, and turbidity meters must be used to understand the turbidity, which requires manual labor and It is costly and time consuming, and cannot be said to be efficient.

The problems to be solved by the present invention are a turbidity estimation device, a turbidity estimation system, a drug injection control system, and a turbidity estimation method that make it possible to efficiently and quickly grasp the turbidity of target water. , and to provide programs.

The turbidity estimation device of the embodiment estimates water depth and a predetermined value based on a plurality of photographed images of different water depths for each turbidity obtained by photographing a plurality of types of water with different turbidities from above the water surface at different water depths. a first information creation unit that creates first information for each turbidity indicating the relationship with the feature quantity; a second information creation section that creates second information indicating a relationship between water depth and the feature quantity based on a plurality of photographed images; and a turbidity estimating section that estimates the turbidity of the water of which turbidity is to be estimated using the information.

FIG. 1 is a diagram showing an example of the configuration of an entire system including a turbidity estimating device according to a first embodiment. FIG. 2A is a diagram showing a first example of an imaging method for creating a database. FIG. 2B is a diagram showing a second example of an imaging method for creating a database. FIG. 3 is a diagram showing an example of the hardware configuration of the turbidity estimation device 2. As shown in FIG. FIG. 4 is a diagram showing an example of the configuration of a turbidity estimation function (a function related to turbidity estimation) realized by a program executed by the processing unit 22 in the first embodiment. FIG. 5 is a diagram showing the relationship between turbidity and hue. FIG. 6 is a diagram showing a method for extracting "hue for each water depth" from a photographed image. FIG. 7A is a diagram showing an example of a multimodal histogram. FIG. 7B is a diagram showing an example of a histogram with a single peak. FIG. 8 is a diagram showing an example in which information in table G13 created for each turbidity is plotted on a graph. FIG. 9 is a flowchart illustrating an example of a processing procedure for creating the database 220 in the first embodiment. FIG. 10 is a diagram illustrating an example of a processing procedure for estimating the turbidity of water as a turbidity estimation target in the first embodiment. FIG. 11 is a diagram showing a method for extracting "average brightness of the surface of the plate for each water depth" and "edge strength of the contour portion of the plate for each water depth" from the photographed image. FIG. 12 is a diagram illustrating an example of the configuration of a turbidity estimation function (a function related to turbidity estimation) realized by a program executed by the processing unit 22 in the second embodiment. FIG. 13A is a diagram showing an example in which information (relationship between water depth and average brightness) of table G24 created for each turbidity is plotted on a graph. FIG. 13B is a diagram showing an example in which information (relationship between water depth and edge strength) of table G24 created for each turbidity is plotted on a graph. FIG. 14 is a flowchart illustrating an example of a processing procedure for creating the database 220 in the second embodiment. FIG. 15 is a flowchart illustrating an example of a processing procedure for estimating the turbidity of water as a turbidity estimation target in the second embodiment.

Embodiments will be described below with reference to the drawings.

In each of the embodiments described below, for example, the turbidity of water at the site is measured by remote sensing from images taken by an imaging device installed at the site, without the need for an investigator or the like to go to the site to sample water or measure turbidity. make it possible to understand the situation. In order to achieve this, in each embodiment, a specific feature amount is obtained from a captured image transmitted from the site to the turbidity estimating device, and the turbidity of water at the site is estimated from the feature amount.

In the first embodiment, an example is shown in which the value of "hue" is adopted as the feature amount. In the second embodiment, an example will be shown in which the value of "average brightness" or the value of "edge strength" is adopted as the feature amount.

<First embodiment>
First, a first embodiment will be described.

(Entire system configuration)
FIG. 1 shows an example of the configuration of the entire system including the turbidity estimating device according to the first embodiment.

As shown in FIG. 1, a management device 1, a container 101, a member 102, a surveillance camera (imaging device) 11, and the like are installed in a management area of a site 100 such as a dam or river that is a water source. The site 100 is unmanned, and water is constantly being supplied to the water container 101 at the site and photographed by the surveillance camera 11, and photographed images are constantly transmitted from the management device 1 to the turbidity estimating device 2. It has become.

The container 101 contains water obtained from a dam, a river, etc. and whose turbidity is to be estimated. It is assumed that water from a dam, a river, etc. always automatically flows into and out of the container 101, and the height of the water level in the container 10 is kept constant.

The container 101 includes a predetermined member 102 inside. The member 102 is a member having a plurality of step-like steps, and has a plurality of horizontal surfaces having different heights. This member 102 is configured so that the depth from the water surface to the member 102 (hereinafter referred to as "water depth") is different (that is, a plurality of types of water depths are formed). Details of the member 102 will be described later.

In addition, in the example of FIG. 1, multiple types of water depths are realized in one container by using the member 102 having a plurality of steps, but the present invention is not limited to this example. For example, without using the member 102, multiple types of water depths may be realized using multiple containers. Furthermore, without using the member 102, a single container 101 having a plurality of bottoms with different heights may be used to realize a plurality of types of water depths.

The surveillance camera 11 is installed above the surface of the water contained in the container 101. The surveillance camera 11 photographs the water contained in the container 101 from above the water surface at different water depths at regular time intervals, and generates "a plurality of photographed images at different water depths." A general-purpose network camera may be applied to the surveillance camera 11.

The management device 1 transmits “a plurality of captured images at different water depths” obtained from the surveillance camera 11 to the turbidity estimation device 2. At that time, information indicating the respective water depth values may also be transmitted. The turbidity estimating device 2 corresponds to, for example, a computer installed at a business office in a predetermined area. The container 101, surveillance camera 11, management device 1, and turbidity estimation device 2 described above constitute a turbidity estimation system.

The turbidity estimation device 2 according to the first embodiment acquires “hue for each water depth” from “a plurality of captured images at different water depths” transmitted from the management device 1.

Specifically, the turbidity estimation device 2 calculates the ``relationship between water depth and hue'' from ``a plurality of captured images with different water depths'' transmitted from the management device 1, and calculates the ``relationship between water depth and hue''. The turbidity of the water in the container 101 is estimated by comparing the information indicating ``with each information in a database for turbidity estimation, which will be described later. The database for turbidity estimation (hereinafter referred to as the "database") is a turbidity estimation database that calculates turbidity from "multiple images taken at different water depths" for each turbidity that have been taken in advance for each type of water with different turbidity. This includes information indicating another "relationship between water depth and hue." A more specific turbidity estimation method will be described later.

The turbidity estimating device 2 displays information indicating the estimated turbidity on a predetermined display device or transmits it to predetermined devices (including the management device 3) such as computers installed at various locations. Further, the turbidity estimating device 2 generates an alarm and notifies a predetermined device, for example, when the estimated turbidity increases and enters a predetermined turbidity range for which an alarm is to be generated.

The water treatment facility 300 is, for example, a water purification plant, and includes a management device 3, a pond (mixing pond, etc.) 301, a chemical injection device (flocculant injection device, etc.) 302, an agitation device 303, a chemical control unit 31, and the like. The management device 3, the drug control unit 31, and the turbidity estimating device 2 described above constitute a drug injection control system.

The management device 3 receives information indicating turbidity transmitted from the turbidity estimating device 2, and transmits it to the drug control unit 31. The chemical control unit 31 controls the injection of a chemical (for example, a flocculant) in the water treatment facility and the stirring operation of the stirring device 303 according to the turbidity estimated by the turbidity estimation device 2. . For example, when the turbidity estimated by the turbidity estimation device 2 exceeds a certain threshold, the injection amount or injection rate of the drug performed by the drug injection device 302 is increased, or the injection operation that has been stopped is started. control.

(Photography method for creating database)
FIGS. 2A and 2B outline two types of imaging methods for creating a database.

FIG. 2A shows a method in which multiple containers (for example, buckets, beakers, etc.) are used to photograph multiple waters at different depths using the surveillance camera 11. In this case, each container shall have a predetermined color (for example, white). On the other hand, FIG. 2B shows a method in which water is photographed by the surveillance camera 11 using one container equipped with members that realize a plurality of types of water depths. Note that the container shown in FIG. 2B may be similar to the container shown in FIG. 1, but a different container may be used.

Although only two containers 101A and 101B are shown in FIG. 2A, more containers may be provided to photograph water at greater depths. Furthermore, by changing the depth of water in one container in increments of 1 (cm), for example, water at a plurality of different depths may be photographed.

FIG. 2B shows a container 101 that includes a member 102 having a plurality of step-like steps inside as described above. The member 102 has, for example, a plurality of (for example, four) horizontal surfaces (surfaces on the side to be photographed) having different heights from the water bottom, and these surfaces realize a plurality of types of water depths from the water surface. The surveillance camera 11 photographs the water contained in the container 101 including such a member 102. The photographed image obtained from the surveillance camera 11 shows a plurality of (for example, four) regions having different water depths. Since each region has a different water depth, the color of the water surface appears different in each region. That is, the hue of each area is different.

In addition, in the example of FIG. 2B, multiple types of water depths are realized in one container by using the member 102 having a plurality of steps, but the present invention is not limited to this example. For example, without using the member 102, a single container 101 having a bottom with a plurality of steps may be used to realize a plurality of types of water depths.

In both the method shown in FIG. 2A and the method shown in FIG. 2B, when creating a database, water with a plurality of types of turbidity is first prepared. In order to prepare water with multiple types of turbidity, water may be obtained not only at one location (eg, a certain river) but also at multiple locations (eg, multiple rivers). Furthermore, the present invention is not limited to acquiring local water such as from rivers or dams, but may also employ, for example, simulated water that artificially reproduces local water. Alternatively, water having a specific turbidity may be produced by diluting raw water, for example. In this example, it is assumed that water with one or more types of turbidity is acquired for one river. This type of water acquisition will be carried out from multiple rivers. During photography, images taken at multiple types of water depth are obtained for each type of water with one type of turbidity. Such captured images are acquired for water with multiple types of turbidity.

The number of times of imaging is different between the method shown in FIG. 2A and the method shown in FIG. 2B. The method shown in FIG. 2B requires fewer shots than the method shown in FIG. 2A.

In the method shown in FIG. 2A, a container 101A or 101B is used for water having one type of turbidity, the water is placed in the container, and images are taken at different water depths. In this case, the number of times M of photographing is performed is M=(number of rivers)×(number of turbidities)×(number of water depths).

In the method shown in FIG. 2B, one container 101 equipped with a member 102 is used for water with one type of turbidity, the water is placed in the container 101, and water at multiple types of depths is photographed at once. In this case, the number of times N of photographing is performed is N=(number of rivers)×(number of turbidities).

For example, when the number of water depths is four, N=M/4, and the method shown in FIG. 2B can reduce the number of photographing operations to 1/4 compared to the method shown in FIG. 2A.

By doing so, in both the method shown in FIG. 2A and the method shown in FIG. 2B, "a plurality of photographed images of different water depths according to turbidity" can be obtained by photographing. From the photographed image, "hue for each water depth" for each turbidity can be obtained.

(Hardware configuration of turbidity estimation device 2)
FIG. 3 shows an example of the hardware configuration of the turbidity estimation device 2. However, the configuration shown in FIG. 3 is not limited to this example, and may be modified and implemented as appropriate.

As shown in FIG. 3, the turbidity estimating device 2 includes a receiving section 21, a processing section 22, an input section 23, a display section 24, a storage section 25, and a transmitting section 26, which are connected to the bus BL. The storage section 25 includes a main storage section 251 and an auxiliary storage section 252.

The receiving unit 21 corresponds to a receiving device that receives data transmitted from the outside (for example, "a plurality of captured images of different water depths" transmitted from the management device 1, etc.).

The processing unit 22 corresponds to a processor and controls each unit connected to the bus BL. The processing unit 22 executes programs and databases stored in the storage unit 25.

The input unit 23 corresponds to an input device such as a keyboard, touch panel, or pointing device.

The display unit 24 corresponds to a display device that displays various information (for example, information indicating estimated turbidity, etc.) on a screen.

The main storage section 251 of the storage section 25 corresponds to a storage device that stores programs executed by a processor, and the auxiliary storage section 252 corresponds to a storage device that stores various data.

The transmitter 26 corresponds to a transmitter that transmits various types of information (for example, information indicating estimated turbidity) to an external device (for example, the management device 3, etc.).

(Configuration of functions related to turbidity estimation)
FIG. 4 shows an example of the configuration of a turbidity estimation function (a function related to turbidity estimation) realized by a program executed by the processing unit 22. However, the configuration shown in FIG. 4 is not limited to this example, and may be modified and implemented as appropriate.

The turbidity estimation function according to the present embodiment includes a first information creation unit (water depth-hue information creation unit for each turbidity) 221, and a second information creation unit (water depth-hue information creation unit for turbidity estimation). 222 , and a turbidity estimation section 223 .

The first information creation unit 221 captures "a plurality of photographed images at different water depths by turbidity" (for example, turbidity It has a function of creating first information D11 indicating "the relationship between water depth and hue" for each turbidity based on M types x N types of water depth photographed images). The created individual pieces of first information D11 are stored as a database 220 in a predetermined area within the storage unit 25.

The second information creation unit 222 is based on "a plurality of photographed images at different water depths" (for example, photographed images at N types of water depths) obtained by photographing the water to be estimated for turbidity from above the water surface at different water depths. It has a function of creating second information D12 indicating "the relationship between water depth and hue."

The turbidity estimating unit 223 has a function of estimating the turbidity of water as a turbidity estimation target using first information D11 and second information D12 for each turbidity included in the database 220.

Specifically, the turbidity estimation unit 223 selects the turbidity shown in the second information D12 from among the "relationship between water depth and hue" for each turbidity shown in the first information D11 for each turbidity. Select the first information D11 (= information D13) that is closest to the "relationship between water depth and hue" of the water to be estimated, and calculate the turbidity of the water to be estimated from the "relationship between water depth and hue" of the selected information D13. Find the turbidity of water.

(Relationship between turbidity and hue)
FIG. 5 shows the relationship between turbidity and hue.

FIG. 5 shows an example of a plurality of captured images when a plurality of waters with different turbidities at the same water depth are captured using a container similar to the container 101A or 101B described in FIG. 2A. .

For example, for water with a turbidity of 11 (degrees), the hue H obtained from the photographed image is recognized as 70 (degrees). For water with a turbidity of 30 (degrees), the hue H obtained from the photographed image is recognized as 52 (degrees). For water with a turbidity of 57 (degrees), the hue H obtained from the photographed image is recognized as 40 (degrees). For water with a turbidity of 232 (degrees), the hue H obtained from the photographed image is recognized to be in the range of 28 to 40 (degrees).

That is, the higher the turbidity of water, the smaller the hue H becomes, and the resolution thereof decreases, making it difficult to distinguish small differences in turbidity. The higher the turbidity of water, the shallower the water depth suitable for estimating turbidity, and the lower the turbidity of water, the deeper the water depth suitable for estimating turbidity.

(Extraction of information from captured images)
FIG. 6 shows a method for extracting "hue for each water depth" from a photographed image.

FIG. 6 shows an example of a photographed image G11 obtained when a container similar to the container 101A or 101B described in FIG. 2A is filled with water having a certain turbidity prepared for database creation and photographed. There is. Such photographed images are acquired at different water depths.

For example, fill a container with water, change the depth of the water in steps of (cm), and acquire multiple images taken at different depths (for example, images in the RGB color system). Convert to HSV color space to obtain images for each water depth.

Next, in order to obtain the hue of each water surface from each image, a region of interest (ROI) is set at the center of the image. Then, for each image, a histogram G12 of hue H is generated for all pixels within the ROI, and a hue H peak (°) indicating the mode of hue H is determined and summarized in a table G13. For example, if the hue H peak of an image taken at a water depth of 1 (cm) is 39 (°), 39 (°) is stored in the table G13 as the hue H peak corresponding to the water depth of 1 (cm).

The table G13 created in this way corresponds to the first information D11 indicating the "relationship between water depth and hue" of water having a certain turbidity. This first information D11 includes the turbidity as attribute information.

(proper histogram)
The histogram described in FIG. 6 may be multimodal or unimodal.

FIG. 7A shows an example of a multimodal histogram, and FIG. 7B shows an example of a unimodal histogram.

In a multimodal histogram as shown in FIG. 7A, it is difficult to determine the peak of hue H. In a histogram with a single peak as shown in FIG. 7B, the peak of hue H can be easily determined. If the histogram has multiple peaks, it is likely that the water depth is not appropriate, and it means that the value of hue H is not determined due to the color of the bottom of the container, sunlight reflection on the water surface, etc. In such a case, after resetting an appropriate water depth according to the turbidity of the water, the photographed image is reacquired and the histogram is regenerated.

(Plot on graph)
Table G13 explained in FIG. 6 shows the "relationship between water depth and hue" of water having a certain turbidity, and such information is created for each turbidity.

FIG. 8 shows an example in which the information of table G13 created for each turbidity is plotted on a graph.

Graph G16 shown in FIG. 8 is a graph in which the information in table G13 created for each turbidity is plotted as individual points on the graph. A graph G17 shown in FIG. 8 is an enlarged view of the range 40 in which significant data exists in the graph G16. In each graph, the horizontal axis indicates water depth (cm), and the vertical axis indicates hue (°).

In graph G17, the following seven types of data are plotted for each turbidity.

・Water sampling location: Area A, turbidity: 32 (degrees), Container used: Beaker ・Water sampling location: Area A, Turbidity: 32 (degrees), Container used: White bucket ・Water sampling location : B area, turbidity: 41 (degrees), container used: beaker ・Water sampling location: B area, turbidity: 41 (degrees), container used: white bucket ・Water sampling location: C area, turbidity Degree: 89 (degrees), Container used: Beaker ・Water sampled location: C area, Turbidity: 550 (degrees), Container used: Beaker ・Water sampled location: C area, Turbidity: 550 (degrees) , Container used: White bucket However, the example shown in FIG. 8 is one example, and the present invention is not limited to this example.

For example, the white bucket is not limited to a white bucket, but may be a bucket of a different color from white. The beaker may also have a predetermined color.

Furthermore, in order to obtain more data, water may be sampled not only from the three districts A to C but also from more districts. Furthermore, if the area of water targeted for turbidity estimation is, for example, only Area A, water may be sampled only from Area A. Further, the container used is not limited to a beaker or a white bucket, but other types of containers may be used to collect water. Further, the types of containers used may be unified to one type. In this case, for example, the container 101 provided with the member 102 described in FIG. 2B inside or other types of containers may be employed.

(Creation of water depth-hue curve)
The first information creation unit 221 described above creates, for example, first information D11 for each turbidity indicating the "relationship between water depth and hue" based on a group of plots for each turbidity on the graph G17. Each of the first information D11 may be the plot group by turbidity on the graph G17 itself, but it is desirable to adjust it as necessary. In this embodiment, a group of plots for each turbidity is expressed as a curve for each turbidity (hereinafter referred to as a "water depth-hue curve"). At this time, a method such as the least squares method may be applied. In addition, a predetermined correction or deletion is performed on plot groups that cause a decrease in the accuracy of turbidity estimation. Further, the averaging may be performed after removing outliers exceeding a predetermined threshold value in a range where the hue varies by a certain amount or more.

By expressing the plot group for each turbidity as a "water depth-hue curve" for each turbidity, the "water depth-hue curve" of the water that is the target of turbidity estimation can be expressed as a "water depth-hue curve" for each turbidity. This makes it easier to determine which of the curves it falls under.

Table G18 shown in FIG. 8 roughly shows the relationship between turbidity (and hue) and water depth (cm) suitable for estimating turbidity.

As described above, the higher the turbidity of water, the smaller the hue H becomes, and the resolution thereof decreases, making it difficult to distinguish small differences in turbidity. The higher the turbidity of water, the shallower the water depth suitable for estimating turbidity, and the lower the turbidity of water, the deeper the water depth suitable for estimating turbidity.

For example, ``water with a turbidity of 32 (degrees) sampled in area A (beaker)'' and ``water with a turbidity of 32 degrees (degrees) sampled in area A (white)'' have the lowest turbidity among the seven types of water. If you look at the plot on graph G17 for "Bucket", you can see that there is variation in hue in the low water depth range (for example, less than 10 (cm)), but in the high water depth range of code 41 (for example, 10 to 30 (cm) ), the variation in hue is suppressed and a stable hue (for example, around 55 (°)) can be seen.

The next lowest turbidity is ``water with a turbidity of 41 (degrees) sampled in area B (beaker)'' and ``water with a turbidity of 41 (degrees) sampled in area B (white bucket)''. In the range 42 where the water depth is high, variation in hue is suppressed and a stable hue (for example, around 38 (°)) is observed.

"Water with a turbidity of 89 (degrees) sampled in area C (beaker)" has a shallow water depth in the range of code 43 (for example, around 10 (cm) or less) and has a hue (for example, 32 to 40 (degrees)). ) can be seen.

"Water with a turbidity of 550 (degrees) sampled in area D (beaker)" and "water with a turbidity of 550 (degrees) sampled in area D (white bucket)" with the highest turbidity are respectively water depths. A hue (for example, 25 to 30 (°)) can be seen in the range of code 44 (for example, around 10 (cm) or less), and a range of code 45 (for example, around 10 (cm) or less). , the hue (for example, 10 to 20 (°)) can be seen.

In this way, there is a water depth range in which the hue is stable for each turbidity. Therefore, when creating the "relationship between water depth and hue" for each turbidity, leave a group of plots in the water depth range where the hue is stable for each turbidity, and apply the prescribed correction to the plot groups in other ranges. Alternatively, it may be deleted.

(Example of operation of turbidity estimation device 2)
Next, with reference to FIG. 9, an example of a processing procedure for creating the database 220 shown in FIG. 4 will be described.

For example, multiple types of river water with different M types of turbidity are acquired (step S11), the river water is put into containers for each M type of turbidity, and photographed images of N types of water depth (for example, in RGB color system) are obtained. (step S12).

Thereafter, various processes are performed by the first information creation unit 221 on the acquired captured images of "M types of turbidity x N types of water depth".

When the first information creation unit 221 obtains the photographed images of "M types of turbidity x N types of water depth", the first information creation unit 221 converts the color space of each photographed image from RGB to HSV (step S13).

Next, the first information creation unit 221 sets a region of interest (ROI) in each captured image (each H image after HSV conversion) (step S14).

Next, the first information creation unit 221 performs a process of extracting hue from pixels within the ROI in each H image (including a process of generating a histogram of each pixel value and selecting the mode) (step S15).

Finally, the first information creation unit 221 creates a "water depth-hue curve" for each turbidity (corresponding to the first information D11) using the hue for each water depth extracted for each turbidity, and The database 220 is constructed using the above information (step S16).

Next, an example of a processing procedure for estimating the turbidity of water whose turbidity is to be estimated will be described with reference to FIG.

For example, at a site 100 where there is a water source, a predetermined container is installed through which the water from the water source flows, and the surveillance camera 11 constantly photographs the water surface in the container to generate photographed images of "N types of water depths". (Step S21). The generated captured images of "N types of water depths" are transmitted from the management device 1 to the turbidity estimation device 2.

Thereafter, the second information creation unit 222 performs various processes on the acquired captured images of “N types of water depths”, and finally the turbidity estimation unit 223 performs turbidity estimation processing.

When the second information creation unit 222 obtains the photographed images of "N types of water depths", it converts the color space of each photographed image from RGB to HSV (step S22).

Next, the second information creation unit 222 sets a region of interest (ROI) in each captured image (each H image after HSV conversion) (step S23).

Next, the second information creation unit 222 performs a process of extracting hue from pixels within the ROI in each H image (including a process of generating a histogram of each pixel value and selecting the mode) (step S24).

Next, the second information creation unit 222 creates "water depth-hue information" (corresponding to the second information D12) of the water whose turbidity is to be estimated, using the extracted hue for each water depth (step S25 ).

Finally, the turbidity estimation unit 223 plots the "water depth-hue information" of the water to be estimated for turbidity on the "water depth-hue curve" for each turbidity, and plots the "water depth-hue curve" with the maximum overlap. ” (corresponding to information D13), and from the attribute information of the identified “water depth-hue curve”, the turbidity of the water to be estimated is outputted to a display device or the like (step S26).

According to the first embodiment, it is possible to estimate the turbidity of water at the site based on information on hue and water depth obtained from images captured by a surveillance camera installed in a dam or river. Therefore, it becomes possible to grasp the turbidity of water at the site using remote sensing, and there is no need for people to go to the site and measure turbidity using a turbidity meter. In addition, based on information on hue and water depth obtained from captured images, it is possible to timely grasp turbidity changes in water sources, making it possible to early detect excessive turbidity that may occur in water sources, for example, downstream. Efficient water purification operations at water treatment plants become possible.

<Second embodiment>
Next, a second embodiment will be described. Below, the explanation will focus on parts that are different from the first embodiment.

(Entire system configuration)
The configuration of the entire system including the turbidity estimating device according to the second embodiment is similar to that shown in FIG. 1, but the structure inside the container 101 and the processing of the turbidity estimating device 2 are different.

In the first embodiment, in the case where a container 101 is used, "hues for each water depth" can be obtained from images taken by the surveillance camera 11 by arranging members 102 that realize multiple types of water depths therein. An example of how to do this is shown below.

In contrast, in the second embodiment, when using the container 101, for example, plates (members) 103, which will be described later, are placed on each of a plurality of surfaces of the member 102 in the container 101 having different heights. , "average brightness of the plate for each water depth" or "edge strength of the contour of the plate for each water depth" can be obtained from the photographed image of the surveillance camera 11. In this case, since the depths from the water surface to each plate 103 (hereinafter referred to as "water depths") are different, it can be said that multiple types of water depths are realized.

However, the installation of each plate 103 is not necessarily required. for example,
If a structure is provided in the container 101 that allows obtaining the "average brightness of a certain part for each water depth" or "the edge strength of the contour part of a certain part for each water depth" from images captured by the surveillance camera 11, each plate The installation of 103 may be omitted.

Further, when arranging the plate 103, the structure for arranging the plate 103 and the container used are not limited to the above-mentioned example. Multiple types of water depths may be realized in other forms.

For example, without using the member 102, a plurality of containers may be prepared and a plate 103 may be placed at the bottom of each container to realize a plurality of different water depths. Furthermore, without using the member 102, multiple types of water depths may be realized by providing a single container with a plurality of bottoms of different heights and arranging the plate 103 on each bottom.

The turbidity estimating device 2 according to the second embodiment calculates, from the “multiple captured images at different water depths” transmitted from the management device 1, “the hue of the plate surface for each water depth” rather than “the hue for each water depth”. Obtain the "average brightness" or "edge intensity of the contour of the plate for each water depth."

Specifically, the turbidity estimation device 2 calculates the “average brightness of the plate surface for each water depth” or “the contour of the plate for each water depth” from the “multiple captured images at different water depths” transmitted from the management device 1. information indicating the "average brightness of the surface of the plate for each water depth" or "edge strength of the contour of the plate for each water depth" and each information in the database for turbidity estimation described later. By comparing the values, the turbidity of water in the container 101 is estimated. The database consists of "average brightness of the surface of the plate for each water depth" or " It includes information indicating the edge strength of the contour of the plate at each water depth. A more specific turbidity estimation method will be described later.

(Photography method for creating database)
Various imaging methods according to the second embodiment are similar to those described in FIGS. 2A and 2B. However, in the example of FIG. 2A, a plate 103 is placed at the bottom of each container, and in the example of FIG. 2B, a plate (described later) ( member) 103 shall be placed.

In both the method shown in FIG. 2A and the method shown in FIG. 2B, "a plurality of photographed images of different water depths according to turbidity" are obtained by photographing. From the photographed image, "average brightness of the surface of the plate for each water depth" for each turbidity and "edge intensity of the contour of the plate for each water depth" for each turbidity are obtained.

(Extraction of information from captured images)
FIG. 11 shows a method for extracting "average brightness of the plate surface for each water depth" and "edge strength of the contour of the plate for each water depth" from the photographed image.

FIG. 11 shows the results obtained when a white plate (white plate) 103 is placed at the bottom of a container similar to the container 101A or 101B explained in FIG. 2A, and water with a certain turbidity prepared for database creation is photographed. An example of a captured image G21 is shown. Such photographed images are acquired at different water depths.

For example, fill a container with water, change the water depth in cm increments, acquire multiple images taken at different water depths (for example, RGB color system images), and convert these multiple images into monochrome images. and obtain images for each depth.

Next, a region of interest (ROI) is set to include the entire plate 103 in order to obtain the "average brightness of the surface of the plate" and "edge intensity of the contour of the plate" from each image.

Then, the "average brightness of the surface of the plate" is determined from the region within the ROI. For example, the brightness of the area where the white plate 103 is present is determined for each image from within the ROI of the monochrome image G22, which has been subjected to predetermined image processing to make it easier to obtain the brightness of the plate surface as necessary, and the average value Find the average brightness. Furthermore, the "edge strength of the contour part of the plate" is determined from the region within the ROI. For example, the edge strength of the outline of the white plate 103 is determined from within the ROI of the monochrome image G23, which has been subjected to predetermined image processing to facilitate obtaining the edge strength as needed.

The "average brightness of the surface of the plate" and the "edge strength of the contour of the plate" thus determined for each water depth are summarized in table G24.

The table G24 created in this way includes first information D21 indicating the "relationship between water depth and average brightness" of water having a certain turbidity, and first information D31 indicating the "relationship between water depth and edge strength". corresponds to The first information D21 and the first information D31 include the turbidity as attribute information.

Note that details of the first information D21 and the first information D31 will be described later.

(Hardware configuration of turbidity estimation device 2)
The hardware configuration of the turbidity estimating device 2 according to the second embodiment is the same as that described in FIG. 3.

(Configuration of functions related to turbidity estimation)
FIG. 12 shows an example of the configuration of a turbidity estimation function (a function related to turbidity estimation) realized by a program executed by the processing unit 22. However, the configuration shown in FIG. 12 is not limited to this example, and may be modified and implemented as appropriate.

The turbidity estimation function according to the present embodiment includes a first information creation unit (turbidity-based water depth-average brightness information creation unit) 224A and a second information creation unit (turbidity It includes a first information creation section (water depth by turbidity-edge strength information creation section) 224B and a second information creation section as functions related to "edge strength". 225B (water depth-edge strength information creation unit for turbidity estimation target), and further includes a turbidity estimation unit 226 as a function common to both. Note that the first information creation section 224A and the first information creation section 224B may be integrated. Furthermore, the second information creation section 225A and the second information creation section 225B may be integrated.

In this embodiment, an example will be shown in which a function related to "average brightness" and a function related to "edge strength" are selectively used depending on weather conditions (conditions such as weather and sunlight). Processing according to weather conditions will be discussed later.

Note that both the function related to "average brightness" and the function related to "edge intensity" are not necessarily required.

For example, if turbidity estimation is always performed using "average brightness" information without using "edge strength" information, the first information creation section 224B and the second information creation section 225B may be omitted. In addition, the processing related to "edge strength" in the turbidity estimation unit 226 may be omitted.

In addition, when turbidity estimation is always performed using "edge intensity" information without using "average brightness" information, the first information creation section 224A and the second information creation section 225A are omitted. At the same time, the processing related to "average brightness" in the turbidity estimation unit 226 may be omitted.

(Functions related to "average brightness")
The first information creation unit (water depth by turbidity-average brightness information creation unit) 224A is a "turbidity-based information creation unit" that is obtained by photographing multiple types of water with different turbidities from above the water surface at different water depths. A function that creates first information D21 indicating the "relationship between water depth and average brightness" for each turbidity based on "multiple captured images with different water depths" (for example, captured images with M types of turbidity x N types of water depth) has. The created individual pieces of first information D21 are stored as a database 220 in a predetermined area within the storage unit 25.

The second information creation unit (water depth-average brightness information creation unit for which turbidity is estimated) 225A is configured to produce multiple images obtained by photographing the water to be estimated for turbidity from above the water surface at different depths. It has a function of creating second information D22 indicating the "relationship between water depth and average brightness" based on "images" (for example, photographed images of N types of water depths).

The turbidity estimating unit 226 has a function of estimating the turbidity of water as a turbidity estimation target using first information D21 and second information D22 for each turbidity included in the database 220.

Specifically, the turbidity estimation unit 226 selects the turbidity shown in the second information D22 from among the "relationship between water depth and average brightness" for each turbidity shown in the first information D21 for each turbidity. Select the first information D21 (=information D23) that is closest to the "relationship between water depth and average brightness" of the water to be estimated, and calculate the turbidity from the "relationship between water depth and average brightness" of the selected information D23. Find the turbidity of the water to be estimated. In addition, since the sunlight conditions differ depending on environmental conditions such as weather, the average brightness by water depth in "Relationship between water depth and average brightness" is an index (reference value) that represents the degree of attenuation of the average brightness as the water depth changes. Think of it as.

(Function related to "edge strength")
The first information creation unit (water depth by turbidity - edge strength information creation unit) 224B is a "turbidity-based information creation unit" that is obtained by photographing multiple types of water with different turbidities from above the water surface at different water depths. A function that creates first information D31 indicating the "relationship between water depth and edge strength" for each turbidity based on "multiple captured images with different water depths" (for example, captured images with M types of turbidity x N types of water depth) has. The created individual pieces of first information D31 are stored as a database 220 in a predetermined area within the storage unit 25.

The second information creation unit (water depth-edge strength information creation unit for turbidity estimation) 225B is configured to generate multiple images obtained by photographing the water for which turbidity estimation is to be estimated from above the water surface at different depths. It has a function of creating second information D32 indicating the "relationship between water depth and edge strength" based on "images" (for example, photographed images of N types of water depths).

The turbidity estimating unit 226 has a function of estimating the turbidity of water as a turbidity estimation target using first information D31 and second information D32 for each turbidity included in the database 220.

Specifically, the turbidity estimation unit 226 selects the turbidity shown in the second information D32 from among the "relationship between water depth and edge strength" for each turbidity shown in the first information D31 for each turbidity. Select the first information D31 (= information D33) that is closest to the "relationship between water depth and edge strength" of the water to be estimated, and calculate the turbidity from the "relationship between water depth and edge strength" of the selected information D33. Find the turbidity of the water to be estimated. In addition, since the sunlight conditions differ depending on environmental conditions such as weather, the edge strength by water depth in "Relationship between water depth and edge strength" is an index (reference value) that represents the degree of attenuation of edge strength as water depth changes. Think of it as.

(Selection processing function according to weather conditions)
As mentioned above, the various functions shown in Figure 12 include a function related to "average brightness" and a function related to "edge strength." A function related to "average brightness" and a function related to "edge strength" are selectively used.

For example, when the sunlight is strong, the surface and water of the water whose turbidity is to be estimated may be illuminated by strong light, and when it is cloudy, the surface and water of the water whose turbidity is to be estimated may become dark. Therefore, depending on the weather conditions, it may not be possible to accurately obtain the "average brightness" or the "edge strength" from the photographed image. In such a case, it is desirable to select the appropriate one of "average brightness" and "edge intensity" to perform turbidity estimation. In that case, standard data indicating whether "average brightness" or "edge strength" should be used for each type of weather condition is recorded in advance in the database 220, and based on the standard data, the The above selection may be made.

For example, the turbidity estimating unit 226 uses either the water depth-average brightness information D22 or the water depth-edge intensity information D32 in accordance with information on weather conditions supplied from the outside to estimate the turbidity of the water. It may be configured to have a function of estimating turbidity. At this time, it may be determined based on reference data which of the water depth-average brightness information D22 and the water depth-edge strength information D32 to use.

Further, depending on the information on the weather conditions, one of the functions of the second information creation section 225A and the second information creation section 225B may be disabled and the other may be enabled. Further, this process may be performed based on the above reference data.

(Plot on graph)
Table G24 explained in FIG. 11 shows the "relationship between water depth and average brightness" and "relationship between water depth and edge strength" of water with a certain turbidity, but such information is Created.

FIGS. 13A and 13B show examples in which the information of table G24 created for each turbidity ("relationship between water depth and average brightness" and "relationship between water depth and edge strength") is plotted on a graph.

Graph G25 shown in FIG. 13A is a graph in which the information on "relationship between water depth and average brightness" in table G13 created for each turbidity is plotted as individual points on the graph. In the graph, the horizontal axis shows water depth (cm), and the vertical axis shows average brightness.

On the other hand, a graph G26 shown in FIG. 13B is a graph in which the information on the "relationship between water depth and edge strength" of the table G13 created for each turbidity is plotted as individual points on the graph. In the graph, the horizontal axis shows water depth (cm), and the vertical axis shows edge strength.

Each graph plots the following four types of data for each turbidity.

・Water sampling location: Area A, turbidity: 32 (degrees)
・Water sampling location: B area, turbidity: 41 (degrees)
・Water sampling location: C area, turbidity: 89 (degrees)
・Water sampling location: C area, turbidity: 550 (degrees)
However, the example shown in FIGS. 13A and 13B is an example, and the present invention is not limited to this example.

For example, in order to obtain more data, water may be sampled not only in the three districts A to C but also in more districts. Furthermore, if the area of water targeted for turbidity estimation is, for example, only Area A, water may be sampled only from Area A.

(Creation of water depth-average brightness curve)
For example, the first information creation unit 224A described above creates first information D21 for each turbidity indicating the "relationship between water depth and average brightness" based on a group of plots for each turbidity on the graph G25 in FIG. 13A. Create. Each of the first information D21 may be the plot group by turbidity on the graph G25 itself, but it is desirable to adjust it as necessary. In this embodiment, these turbidity-specific plot groups are expressed as turbidity-specific curves (hereinafter referred to as "water depth-average brightness curves"). At this time, a method such as the least squares method may be applied. In addition, a predetermined correction or deletion is performed on plot groups that cause a decrease in the accuracy of turbidity estimation. Alternatively, the averaging may be performed after removing outliers exceeding a predetermined threshold in a range where the average luminance varies by a certain amount or more.

By expressing the plot group for each turbidity as a "water depth-average brightness curve" for each turbidity, the "water depth-average brightness curve" of the water that is the target of turbidity estimation can be expressed as a "water depth-average brightness curve" for each turbidity. - average luminance curve", it becomes easier to judge which one falls under.

(Creation of water depth-edge strength curve)
For example, the first information creation unit 224A described above creates first information D31 for each turbidity indicating the "relationship between water depth and edge strength" based on the plot group for each turbidity on the graph G26 in FIG. 13B. Create. Each of the first information 31 may be the plot group by turbidity on the graph G25 itself, but it is desirable to adjust it as necessary. In this embodiment, these turbidity-specific plot groups are expressed as turbidity-specific curves (hereinafter referred to as "water depth-edge intensity curves"). At this time, a method such as the least squares method may be applied. In addition, a predetermined correction or deletion is performed on plot groups that cause a decrease in the accuracy of turbidity estimation. Alternatively, the averaging may be performed after removing outliers exceeding a predetermined threshold value in a range where the edge strength varies by a certain amount or more.

By expressing the plot group for each turbidity as a "water depth-edge strength curve" for each turbidity, the "water depth-edge strength curve" of the water that is the target of turbidity estimation can be expressed as a "water depth-edge strength curve" for each turbidity. - It becomes easier to determine which edge strength curve it corresponds to.

(Example of operation of turbidity estimation device 2)
Next, an example of a processing procedure for creating the database 220 shown in FIG. 12 will be described with reference to FIG. 14. In the following description, the first information creation section 224A and the first information creation section 224B may be configured to perform various processes in an integrated manner.

For example, multiple types of river water with different M types of turbidity are obtained (step S31), and the river water is poured into a container in which a white plate (white plate) 103 is submerged for each M type of turbidity. A process of acquiring a photographed image (for example, an RGB color system image) is performed (step S32).

Thereafter, various processes are performed by the first information creation unit 224A and the first information creation unit 224B on the acquired captured images of “M types of turbidity×N types of water depth”.

When the first information creation unit 224A and the first information creation unit 224B obtain captured images of “turbidity M types×water depth N types”, they generate a monochrome image from each captured image (step S33).

Next, the first information creation unit 224A and the first information creation unit 224B set a region of interest (ROI) in each captured image (each H image after HSV conversion) (step S34).

Next, the first information creation unit 224A and the first information creation unit 224B perform a process of extracting the average brightness and edge strength from the pixels in the ROI in each H image (step S35).

Finally, the first information creation unit 224A creates a "water depth-average brightness curve" for each turbidity (corresponding to the first information D21) using the average brightness for each water depth extracted for each turbidity, The first information creation unit 224B creates a "water depth-average brightness curve" for each turbidity (corresponding to the first information D31) using the average brightness for each water depth extracted for each turbidity, and uses these The database 220 is constructed (step S36).

Next, with reference to FIG. 15, an example of a processing procedure for estimating the turbidity of water as a turbidity estimation target will be described. In the following description, the second information creation section 225A and the second information creation section 225B may be configured to perform various processes in an integrated manner.

For example, at a site 100 where there is a water source, a predetermined container is installed through which the water from the water source flows, and the surveillance camera 11 constantly photographs the water surface in the container to generate photographed images of "N types of water depths". (Step S41). The generated captured images of "N types of water depths" are transmitted from the management device 1 to the turbidity estimation device 2.

Thereafter, the second information creation unit 225A and the second information creation unit 225B perform various processes on the captured images of “N types of water depths” that have been acquired, and finally the turbidity estimation unit 223 performs turbidity estimation. Processing will take place.

When the second information creation unit 225A and the second information creation unit 225B obtain photographed images of “N types of water depths”, they convert the color space of each photographed image from RGB to HSV (step S42).

Next, the second information creation unit 225A and the second information creation unit 225B set a region of interest (ROI) in each captured image (each H image after HSV conversion) (step S43).

Next, depending on the weather conditions, the second information creation unit 225A or the second information creation unit 225B performs a process of extracting the average brightness or edge strength from pixels within the ROI in each H image (step S44).

Next, the second information creation unit 225A or the second information creation unit 225B uses the extracted average brightness or edge strength for each water depth to generate “water depth-average brightness information” (the (corresponding to the second information D22) or "water depth-edge strength information" (corresponding to the second information D32) of the water whose turbidity is to be estimated (step S45).

Finally, the turbidity estimating unit 226 plots the "water depth-average brightness information" of the water whose turbidity is to be estimated on the "water depth-average brightness curve" for each turbidity, and plots the "water depth-average brightness information" of the water whose turbidity is to be estimated on the "water depth-average brightness curve" for each turbidity. The turbidity of the water to be estimated is determined from the attribute information of the identified "water depth-average brightness curve" (corresponding to information D23), and outputted to a display device, etc.; The turbidity estimating unit 226 plots the "water depth-edge strength information" of the water whose turbidity is to be estimated on the "water depth-edge strength curve" for each turbidity, and determines the "water depth-edge strength curve" with the maximum overlap. ” (corresponding to information D33), and from the attribute information of the specified “water depth-edge strength curve”, the turbidity of the water to be estimated is outputted to a display device or the like (step S46).

According to the second embodiment, the turbidity of water at the site is estimated based on brightness and water depth information obtained from images captured by surveillance cameras installed in dams and rivers, or based on edge strength and water depth information. It becomes possible to do so. Therefore, it becomes possible to grasp the turbidity of water at the site using remote sensing, and there is no need for people to go to the site and measure turbidity using a turbidity meter. In addition, it is possible to timely understand turbidity changes in water sources based on brightness and water depth information obtained from photographed images, or based on edge strength and water depth information, so that excessive turbidity can be detected at an early stage, enabling efficient water purification operations at downstream water treatment plants, for example.

Although several embodiments of the invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, substitutions, and changes can be made without departing from the gist of the invention. These embodiments and their modifications are included within the scope and gist of the invention, as well as within the scope of the invention described in the claims and its equivalents.

1, 3... Management device, 2... Turbidity estimation device, 11... Surveillance camera (imaging device), 21... Receiving section, 22... Processing section, 23... Input section, 24... Display section, 25... Storage section, 26... Transmission unit, 31... Drug control unit, 100... Site, 101, 101A, 101B... Container, 102, 103... Member, 220... Database, 221... First information creation unit (water depth-hue information creation unit by turbidity) ), 222...Second information creation section (water depth-hue information creation section for turbidity estimation), 223...Turbidity estimation section, 224A...first information creation section (water depth-average brightness information creation section by turbidity) section), 225A...second information creation section (water depth-average brightness information creation section for turbidity estimation), 224B...first information creation section (water depth-edge intensity information creation section by turbidity), 225B... 2nd information creation section (water depth-edge strength information creation section for turbidity estimation), 226...Turbidity estimation section, 300...Water treatment facility, 301...Pond (mixing pond, etc.), 302...Medicine injection device (coagulation agent injection device, etc.), 303... Stirring device.

Claims (15)

A first method that shows the relationship between water depth and a predetermined feature amount based on a plurality of captured images obtained by photographing a plurality of waters with different turbidities from above the water surface at different water depths. a first information creation unit that creates information for each turbidity;
A plurality of photographed images at different water depths obtained by photographing the water to be estimated for turbidity from above the water surface at different water depths. Second information indicating the relationship between the water depth and the feature amount is obtained based on the plurality of photographed images. a second information creation unit that creates;
A turbidity estimating device, comprising: a turbidity estimation unit that estimates the turbidity of the water to be estimated, using the first information for each turbidity and the second information. The turbidity estimation device according to claim 1, wherein the feature amount includes hue information. The turbidity estimating device according to claim 1, wherein the feature amount includes information on brightness. The turbidity estimation device according to claim 1, wherein the feature amount includes information on edge strength. The first information creation unit includes:
Obtaining hue information as the feature quantity from the photographed image,
As the first information, water depth-hue information indicating the relationship between water depth and hue is created for each turbidity,
The second information creation unit includes:
creating water depth-hue information indicating a relationship between water depth and hue as the second information;
The turbidity estimation device according to claim 1. The first information creation unit includes:
Obtaining brightness information from the captured image as the feature amount,
As the first information, water depth-brightness information indicating the relationship between water depth and brightness is created for each turbidity,
The second information creation unit includes:
As the second information, water depth-brightness information indicating a relationship between water depth and brightness is created;
The turbidity estimation device according to claim 1. The first information creation unit includes:
Obtaining edge strength information from the captured image as the feature amount,
As the first information, water depth-edge strength information indicating the relationship between water depth and edge strength is created for each turbidity,
The second information creation unit includes:
As the second information, water depth-edge strength information indicating a relationship between water depth and edge strength is created;
The turbidity estimation device according to claim 1. The first information creation unit further includes:
Obtaining edge strength information from the captured image as the feature amount,
As the first information, water depth-edge strength information indicating the relationship between water depth and edge strength is created for each turbidity,
The second information creation unit includes:
As the second information, create water depth-edge strength information indicating the relationship between water depth and edge strength,
The turbidity estimating unit is
Estimating the turbidity of the water of the turbidity estimation target using either water depth-luminance information or water depth-edge strength information according to weather conditions;
The turbidity estimation device according to claim 6. The turbidity estimation device according to any one of claims 1 to 8,
One or more containers containing the water whose turbidity is to be estimated and each having a different water depth from the water surface to the water bottom or a predetermined member;
an imaging device that photographs the water to be estimated for turbidity contained in the container from above the water surface at different water depths, and generates a plurality of photographed images at different water depths;
A turbidity estimation system comprising: the container is one container that includes the predetermined member inside thereof;
The turbidity estimation system according to claim 9. the predetermined member is a member having a horizontal surface;
The turbidity estimation system according to claim 10. The predetermined member is a member having a plurality of horizontal surfaces of different heights,
The turbidity estimation system according to claim 10. The turbidity estimation device according to any one of claims 1 to 8,
A chemical injection control system, comprising: a chemical control unit that controls injection of a chemical in a water treatment facility according to the estimated turbidity. A turbidity estimation method executed by a turbidity estimation device,
Based on multiple captured images obtained by photographing multiple types of water with different turbidities from above the water surface at different water depths, a graph showing the relationship between water depth and a predetermined feature quantity is obtained. Creating the information in 1 for each turbidity,
Creating second information indicating the relationship between the water depth and the feature amount based on a plurality of captured images obtained by photographing the water to be estimated for turbidity from above the water surface at different depths. and,
A turbidity estimation method, comprising: estimating the turbidity of the water of the turbidity estimation target using the turbidity-specific first information and the second information. to the computer,
Based on multiple captured images obtained by photographing multiple types of water with different turbidities from above the water surface at different water depths, a graph showing the relationship between water depth and a predetermined feature quantity is obtained. A function to create information on 1 by turbidity,
a function of creating second information indicating a relationship between water depth and the feature amount based on a plurality of captured images obtained by photographing the water whose turbidity is to be estimated from above the water surface at different water depths; ,
A program for realizing a function of estimating the turbidity of water of the turbidity estimation target using the turbidity-specific first information and the second information.