Estimation model acquisition device, river flow rate estimation device, method for manufacturing an estimation model, river flow rate estimation method, and program

Abstract

[Problem] There is a need to enable conventional estimation model generating devices to be capable of estimating information relating to a flow rate of a target river. [Solution] Information relating to the flow rate of a river of interest can be estimated by means of an estimation model generating device 101 comprising: a drainage basin data acquiring unit 142 for acquiring drainage basin data of a drainage basin of a river of interest, from an image including the drainage basin; a candidate model acquiring unit 143 for acquiring one or more candidate models for estimating information relating to the flow rate of the river, on the basis of the drainage basin data; a verifying unit 145 for verifying the validity of at least one candidate model from among the one or more candidate models, using verification data; an estimation model acquiring unit 147 for acquiring one candidate model from among the one or more candidate models as an estimation model, on the basis of the results of the verification; and a model accumulating unit 149 for accumulating the acquired estimation model.

Description

Technical Field

[0001] The present invention relates to an estimation model acquisition device for acquiring an estimation model for estimating the flow rate of a river, a river flow rate estimation device, a method for manufacturing an estimation model, a river flow rate estimation method, and a program.

Background Art

[0002] For various purposes, there is a need to grasp information regarding the flow rate of a river. For example, in recent years, as awareness of sustainable water resource utilization has been increasing, information regarding the flow rate of a river has become necessary for appropriately grasping and analyzing information regarding water resources.

[0003] Regarding a method for obtaining information regarding the flow rate of a river, for example, Patent Document 1 below describes calculating a basin gradient from elevation and area obtained from a commercially available topographic map, creating a regression equation between the gradient and the flow rate, and obtaining a regression function for estimating the flow rate of a mountain river basin where flow rate measurement has not been performed.

[0004] In addition, Patent Document 2 below describes calculating a vegetation index (NDVI) from satellite data, obtaining a soil moisture content distribution, and assisting in determining a boring point for groundwater exploration.

[0005] In addition, Non-Patent Document 1 below describes a rainfall runoff inundation model (Rainfall Runoff Inundation Model: RRI model) for integrally analyzing rainfall runoff and flood inundation using rainfall data.

Prior Art Documents

Patent Documents

[0006]

Patent Document 1

Patent Document 2

Non-Patent Documents

[0007]

Non-Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0008] Conventionally, in order to accurately grasp information on the flow rate of a river, it has been necessary to visit the site (the basin of the target river) and conduct flow rate measurements. However, such on-site visits and flow rate measurements require various costs. That is, there is a demand to be able to estimate information on the flow rate of the target river.

[0009] In addition, although the method described in Patent Document 1 above can estimate the flow rate using a regression function, a method different from this is also required.

[0010] An object of this invention is to provide an estimation model acquisition device, a river flow rate estimation device, a method for manufacturing an estimation model, a river flow rate estimation method, and a program that are useful for estimating information on the flow rate of a target river.

Means for Solving the Problems

[0011] The apparatus for generating an estimation model of the first invention includes a basin data acquisition unit that acquires basin data of a basin included in an image of a target river basin, a candidate model acquisition unit that acquires one or more candidate models for estimating information related to the flow rate of the river based on the basin data, a verification unit that verifies the validity of at least one of the one or more candidate models using verification data, an estimation model acquisition unit that acquires one of the one or more candidate models as an estimation model based on the verification result, and a model storage unit that stores the acquired estimation model. It is an apparatus for generating an estimation model.

[0012] With such a configuration, it is possible to generate an estimation model capable of estimating information related to the flow rate of the target river.

[0013] Further, in the apparatus for generating an estimation model of the second invention, with respect to the first invention, the verification unit performs verification of the candidate model using information related to the flow rate obtained using the candidate model to be verified and information obtained by applying the value of a parameter corresponding to the basin data to one or more predetermined expressions related to the flow rate of the river. It is an apparatus for generating an estimation model.

[0014] With such a configuration, it is possible to generate an estimation model capable of estimating information related to the flow rate of the target river with high accuracy using information that can be obtained even from a remote location.

[0015] Further, in the apparatus for generating an estimation model of the third invention, with respect to the second invention, the one or more predetermined expressions include at least one of an isochrone formula representing the relationship between water level and flow rate and a rational formula representing peak flow rate. It is an apparatus for generating an estimation model.

[0016] With such a configuration, it is possible to generate an estimation model capable of estimating information related to the flow rate of the target river with high accuracy.

[0017] Further, in the estimation model generation device of the fourth invention, for the third invention, the verification unit uses information on the flow rate obtained using the candidate model to be verified, information obtained by applying the values of the parameters corresponding to the basin data to the equal flow formula, and information obtained by applying the values of the parameters corresponding to the basin data to the rational formula to verify the candidate model. It is an estimation model generation device.

[0018] With such a configuration, it is possible to generate an estimation model capable of estimating information on the flow rate of the target river with higher accuracy.

[0019] Further, the estimation model generation device of the fifth invention further includes an existing data acquisition unit that acquires information on the flow rate of a river from a hydrological database for any one of the first to fourth inventions. The verification unit uses at least the information acquired by the existing data acquisition unit and the information on the flow rate obtained using the candidate model to be verified to verify the candidate model. It is an estimation model generation device.

[0020] With such a configuration, it is possible to generate an estimation model capable of estimating information on the flow rate of the target river with high accuracy using information that can be acquired even from a remote location.

[0021] Further, in the estimation model generation device of the sixth invention, for the fifth invention, the existing data acquisition unit acquires the minimum flow rate of the river from the hydrological database, and the verification unit uses the base flow rate obtained using the candidate model to be verified and the minimum flow rate acquired by the existing data acquisition unit to verify the candidate model. It is an estimation model generation device.

[0022] With such a configuration, it is possible to generate an estimation model capable of estimating information on the flow rate of the target river with high accuracy.

[0023] Further, the estimation model generation device of the seventh invention repeats the acquisition of one or more candidate models by the candidate model acquisition unit and the verification of one or more candidate models by the verification unit for any one of the first to sixth inventions until it is determined by the estimation model acquisition unit that the verification result satisfies a predetermined condition.

[0024] With such a configuration, it is possible to easily generate an estimation model capable of estimating information on the flow rate of the target river with a certain degree of accuracy.

[0025] Further, the estimation model generation device of the eighth invention is an estimation model generation device in which, for any one of the first to seventh inventions, the estimation model is a rainfall runoff flood model, and the basin data includes land use data.

[0026] With such a configuration, it is possible to generate an estimation model capable of estimating information on the flow rate of the target river with higher accuracy.

[0027] Further, the estimation model generation device of the ninth invention further includes an image acquisition unit that acquires an image to be used based on position information indicating the basin of the target river for any one of the first to eighth inventions, and the basin data acquisition unit acquires basin data of the basin from the acquired image.

[0028] With such a configuration, it is possible to easily estimate information on the flow rate of the target river with high accuracy.

[0029] Further, the river flow rate estimation device of the tenth invention includes any one of the first to ninth estimation model generation devices, a flow rate information acquisition unit that acquires information on the flow rate of the target river using the estimation model stored in the model storage unit, and a flow rate information output unit that outputs information on the flow rate.

[0030] With such a configuration, it is possible to estimate information on the flow rate of the target river.

[0031] Further, the river flow rate estimation device of the eleventh invention further includes a meteorological data acquisition unit that acquires meteorological data of the basin of the target river with respect to the tenth invention, and the flow rate information acquisition unit applies the meteorological data to the estimation model to obtain information regarding the flow rate. It is a river flow rate estimation device.

[0032] With such a configuration, information regarding the flow rate of the target river can be estimated.

[0033] Further, the river flow rate estimation device of the twelfth invention, with respect to the tenth or eleventh invention, the flow rate information acquisition unit calculates and acquires the specific flow rate of each region previously partitioned in the basin using the result output using the estimation model. It is a river flow rate estimation device.

[0034] With such a configuration, the specific flow rate of each basin can be estimated.

[0035] Further, the river flow rate estimation device of the thirteenth invention, with respect to the twelfth invention, further includes a recharge area information acquisition unit that acquires information regarding an area that satisfies a predetermined recharge condition using the specific flow rate of each basin, and an information output unit that outputs information regarding the area. It is a river flow rate estimation device.

[0036] With such a configuration, information regarding an area that satisfies the recharge condition can be output.

Effect of the Invention

[0037] According to the present invention, it is possible to provide an estimation model acquisition device, a river flow rate estimation device, a method for manufacturing an estimation model, a river flow rate estimation method, and a program that are useful for estimating information regarding the flow rate of a target river.

Brief Description of the Drawings

[0038]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Embodiments for Carrying Out the Invention

[0039] Hereinafter, embodiments of an estimation model acquisition device and a river flow rate estimation device including the same will be described with reference to the drawings. In the embodiments, components denoted by the same reference numerals perform the same operations, and thus repeated descriptions may be omitted.

[0040] The terms used hereinafter are generally defined as follows. Note that the semantic meanings of these terms should not always be interpreted as shown here, and for example, when individually described hereinafter, they should be interpreted in consideration of the description.

[0041] Regarding a certain matter, an identifier is a character, symbol, etc. that uniquely indicates the matter. The symbol is, for example, alphanumeric characters, other symbols, etc., but is not limited thereto. An identifier is, for example, a symbol string that does not itself indicate a specific meaning, but any type of information that can identify the corresponding matter is acceptable. That is, the identifier may be the name of the thing it indicates itself, or it may be a combination of symbols that uniquely correspond. A combination of two or more pieces of information (for example, the attribute values of records stored in a database, etc.) may be used as an identifier.

[0042] "Obtaining" may include obtaining a matter input by a user or the like, or may include obtaining information stored in the own device or another device (which may be pre-stored information or information generated by performing information processing in the device). Obtaining information stored in another device may include obtaining information stored in another device via an API or the like, or may include obtaining the content of a document file provided by another device (including the content of a web page, etc.).

[0043] Also, for obtaining information, so-called machine learning techniques may be used. Regarding the use of machine learning techniques, for example, it can be done as follows. That is, learning information with specific types of input information as input and the types of output information to be obtained as output is configured using machine learning techniques. For example, in advance, two or more pairs of input information and output information are prepared, and the two or more pieces of information are given to a module for configuring learning information of machine learning to configure learning information, and the configured learning information is stored in a storage unit. Note that the learning information can also be called a learning device or a classifier. Note that as machine learning techniques, for example, deep learning, random forest, SVR, etc. are acceptable. Also, for machine learning, functions in various machine learning frameworks such as fastText, tinySVM, random forest, TensorFlow, etc., and various existing libraries can be used.

[0044] Note that the learning information is not limited to that obtained by machine learning. The learning information may be, for example, a table showing the correspondence between an input vector based on input information or the like and output information. In this case, the output information corresponding to the feature vector based on the input information may be obtained from the table, or a vector approximating the feature vector based on the input information may be generated using two or more input vectors in the table and parameters such as weighting of each input vector, and the final output information may be obtained using the output information corresponding to each input vector used in the generation and the parameters. Further, the learning information may be, for example, a function or the like representing the relationship between an input vector based on input information or the like and information for generating output information. In this case, for example, the information corresponding to the feature vector based on the input information may be obtained by the function, and the output information may be obtained using the obtained information.

[0045] Outputting information is a concept including display on a display, projection using a projector, printing by a printer, sound output, transmission to an external device, storage in a recording medium, delivery of a processing result to another processing device or another program, and the like. Specifically, for example, it includes enabling display of information on a web page, transmitting it as an e-mail or the like, and outputting information for printing.

[0046] Receiving information is a concept including receiving information input from an input device such as a keyboard, mouse, or touch panel, receiving information transmitted from another device or the like via a wired or wireless communication line, and receiving information read from a recording medium such as an optical disk, magnetic disk, or semiconductor memory.

[0047] (Embodiment)

[0048] In the present embodiment, the estimated model generation device acquires basin data of a target river basin from an image including the basin, and acquires one or more candidate models for estimating the river flow rate based on the basin data. Then, the validity of at least one of the one or more candidate models is verified using verification data, and one candidate model is acquired as an estimated model based on the verification result. The verification of the candidate model is performed, for example, based on information regarding the flow rate obtained using the candidate model and information obtained by applying the values of parameters corresponding to the basin data to one or more predetermined equations regarding the river flow rate. It is preferable that the one or more equations include at least one of the equal flow formula and the rational formula, and the verification is performed using both the information obtained by the equal flow formula and the information obtained by the rational formula.

[0049] Also, the estimated model generation device may be used in a river flow rate estimation device. The river flow rate estimation device may be configured to acquire flow rate information regarding the river flow rate using the generated estimated model and output the flow rate information. The river flow rate estimation device, for example, acquires the meteorological data of the basin and applies the meteorological data to the estimated model to acquire information regarding the river flow rate. Here, the river flow rate estimation device may calculate and acquire the specific flow rate of each region preliminarily partitioned from the output result of the estimated model, and further, acquire information regarding water retention according to the relative water retention amount of each region understood from the specific flow rate. Hereinafter, an example of the estimated model generation device and the river flow rate estimation device configured as described above will be described.

[0050] FIG. 1 is a diagram for explaining the use of the river flow rate estimation device according to the present embodiment. FIG. 2 is a diagram for explaining an example of the estimated model used in the river flow rate estimation device.

[0051] In FIG. 1, a general water cycle is shown by taking one river basin as an example. That is, in nature, water circulates as precipitation (S1) from the air, infiltration (S2) into the ground, outflow (S3) from groundwater to rivers and the ocean, evaporation (S4) from the ground surface, the ocean, etc., and precipitation (S1) from the air. For each aspect, for example, the infiltration amount (the amount of water supplied from the ground surface to the groundwater level), the groundwater level, the river flow rate, the water usage fee associated with human activities, and the evapotranspiration amount, etc. can be values representing the situation of the water cycle.

[0052] In the present embodiment, the river flow rate estimation device is configured to estimate and output information regarding the river flow rate by, for example, an estimation model. The estimation model is information for estimating the flow rate, etc. of the river based on information regarding the river basin. The estimation model is a distributed model that divides the river basin into unit areas. Such an estimation model is generated by an estimation model generation device.

[0053] In the present embodiment, the estimation model is configured using a parameter group including two or more parameters. For each unit area of the model, by setting the values of the parameters, it is possible to estimate information regarding the flow rate of the river in the river basin.

[0054] Here, the unit area may be an area called a mesh that constitutes the model, or may be an area partitioned by other methods. It can be said that the unit area is a pre-partitioned area. For example, an area with a predetermined size specified by longitude and latitude may be set as the unit area. Note that the parameters of each unit area included in one river basin may be set to the same value as each other.

[0055] As shown in FIG. 2, in the present embodiment, the estimation model is, for example, a known rainfall-runoff-flood model (RRI model). That is, in the present embodiment, by using the rainfall-runoff-flood model as the estimation model, the estimation process can be easily performed.

[0056] That is, in the present embodiment, the estimation model is configured to be able to analyze the flow of water such as rainfall runoff and flood inundation in the basin based on input data such as information on topography, hydrogeology, etc. set for each region and meteorological data. Here, as information on the topography of a unit region, for example, information such as elevation, gradient, river topography (river channel cross-section, etc.) may be applicable. Also, as information on hydrogeology, etc., information on roughness coefficient, soil layer thickness, saturated effective porosity, permeability coefficient, groundwater outflow coefficient, etc., that is, information on land use may be applicable. Also, as meteorological data, for example, information on precipitation amount (rainfall distribution, etc.) known from radar-analyzed rainfall, rain gauges, etc. may be applicable.

[0057] The estimation model generated in the present embodiment has a set of constants based on information on topography, hydrogeology, etc. for parameters other than meteorological data so that the amount of water in each unit region can be specified when meteorological data is applied. That is, it is configured so that the amount of water in each unit region can be specified by using meteorological data and the estimation model with the set of constants. In the present embodiment, the estimation model is configured to be able to output information on the amount of water in the basin when information on the precipitation amount in each unit region, that is, the rainfall distribution, is applied as meteorological data. In other words, the output information of the estimation model is, for example, information on rivers such as river flow rate and river water level, and information on the amount of water such as inundation depth and groundwater level. In the present embodiment, by using the estimation model, the river flow rate and river water level in the unit region that is a river, and the inundation depth and groundwater level in other unit regions, etc. can be obtained as information on the river flow rate.

[0058] Note that the estimation model may be other types of distributed models. The distributed model that can be an estimation model may be one in which the state of the water cycle is modeled in a predetermined region including forests, rivers, and other types of land. Here, the predetermined region may be set, for example, in units of factories or in basin units including factories. That is, the predetermined region may be set as a spatial unit having a certain range of spatial extent.

[0059] Note that the parameters are not limited to these and are variously selected depending on the type of distribution model used as the estimation model. For example, those related to elements such as meteorology, the ground surface, shallow layer, and deep layer (e.g., precipitation, temperature, topsoil, sediment, aquifer, etc.) or those related to site-specific properties, etc. can be set. More specifically, for example, parameters such as precipitation amount, evapotranspiration amount, temperature, wind speed, sunshine duration, relative humidity, canopy cover rate, canopy storage amount, litter cover rate, litter storage amount, snow accumulation and melting temperature, albedo, bulk transport coefficient, soil evaporation efficiency, groundwater flow, equivalent roughness coefficient, groundwater flow, permeability coefficient, effective porosity, solid-phase compression rate, relative permeability, capillary pressure, fluid physical properties, fluid density, air density, viscosity coefficient of the fluid, viscosity coefficient of the air, etc. are included.

[0060] Next, the configuration of the river flow rate estimation device according to the present embodiment will be described. In the present embodiment, the river flow rate estimation device is configured to also function as an estimation model generation device that generates an estimation model. It can be said that the river flow rate estimation device has an estimation model generation device. Note that the river flow rate estimation device and the estimation model generation device may be configured as separate devices that are different from each other in terms of hardware. In this case, the river end estimation device may be configured to be able to acquire the estimation model generated by the estimation model generation device. Here, when acquiring the estimation model, it may be configured such that operations and instructions by the user are performed.

[0061] FIG. 3 is a block diagram of the river flow rate estimation device 1 in the present embodiment.

[0062] As shown in FIG. 3, the river flow rate estimation device 1 includes a storage unit 110, a reception unit 120, a reception section 130, a processing unit 140, and a transmission unit 170. The river flow rate estimation device 1 is, for example, a server device.

[0063] In the present embodiment, a user of the river flow rate estimation device 1 can use the river flow rate estimation device 1 by directly operating the river flow rate estimation device 1 or by using a terminal device (not shown) or the like. Note that, as the terminal device, for example, any device such as a portable information terminal device such as a so-called smartphone, a terminal device which is a personal computer (PC) such as a laptop computer, or a tablet-type information terminal device may be used.

[0064] The storage unit 110 includes a learning information storage unit 111, a basin data storage unit 113, and a model storage unit 117. The storage unit 110 is preferably a non-volatile recording medium, but can also be realized by a volatile recording medium. Information such as the acquired information is stored in the storage unit 110, but the process of storing information etc. is not limited to this. For example, information etc. may be stored via a recording medium, information etc. transmitted via a communication line etc. may be stored, or information etc. input via an input device may be stored.

[0065] The learning information storage unit 111 stores learning information acquired in advance. In the present embodiment, the learning information is created such that the information to be input is an image including a basin, and the information to be output is land use data of each part of the image. The learning information is generated using so-called machine learning techniques. The learning information is, for example, generated in a device different from the river flow rate estimation device 1, but may be generated by the processing unit 140 etc. and stored in the learning information storage unit 111.

[0066] Here, the image which is the information to be input is an image acquired by the image acquisition unit 141 as described later. The image is an aerial photograph or a satellite photograph taken by a camera, but is not limited to this. For example, various forms of maps imaged may be used. Also, in the present embodiment, the image is an expression that can also include digitized data that can be configured as an image. For example, the image may be grid data that can be expressed as an image such as a heat map.

[0067] Land use data is data that identifies what kind of land each part of the basin included in the image is. The land use data is, for example, data in which an identifier for specifying the land use is associated with each identifier for specifying each part of the basin. The part of the basin may be each unit area, or may be a part indicated by one or more predetermined numbers of pixels in the image. Further, it may be a predetermined range (for example, an administrative division, a district, etc.) that is divided in advance. Further, the land use includes, for example, paddy fields, other agricultural lands, forests, wastelands, building sites, trunk transportation lands, lakes, rivers, residential areas, urban areas, etc., but is not limited thereto.

[0068] The learning information is generated, for example, as follows by using a machine learning method. That is, the learning information is generated by providing information on a combination of an image including a basin whose use has been confirmed in advance by a survey or the like and land use data for specifying the use of each part of the basin to a module for constructing a classifier of machine learning. The generated learning information is stored in the learning information storage unit 111. As the machine learning method, for example, various methods such as deep learning such as a convolutional neural network (CNN), random forest, SVR, etc. can be used. Further, for machine learning, for example, functions in various machine learning frameworks such as fastText, tinySVM, random forest, TensorFlow, etc. and various existing libraries can be used.

[0069] Note that the learning information may be data that associates a pixel in the image with land use data that can correspond to the pixel. For example, the learning information may be data that associates a specific pixel pattern that appears in the image with land use data corresponding to the pixel pattern. In this case, the basin data acquisition unit 142 described later can acquire land use data related to the unit area by specifying the corresponding land use data with reference to the learning information for each unit area in the image.

[0070] The basin data storage unit 113 stores basin data related to the basin. In the basin data storage unit 113, for example, as will be described later, the basin data acquired by the basin data acquisition unit 142 is accumulated. The basin data is stored, for example, in association with an identifier for identifying the basin. Note that the basin data storage unit 113 may store basin data prepared in advance.

[0071] In the present embodiment, the basin data is, for example, land use data, information on topography, information on hydrogeology, and the like.

[0072] The model storage unit 117 stores the estimation model acquired by the estimation model acquisition unit 147 described later. The estimation model is stored, for example, in association with an identifier for identifying the corresponding basin.

[0073] The receiving unit 120 is usually realized by wireless or wired communication means, but may also be realized by means of receiving broadcasts. The receiving unit 120 receives information transmitted from other devices. The receiving unit 120 stores the received information, for example, in the storage unit 110. In the present embodiment, the user inputs information, etc. using, for example, a terminal device and transmits it to the river flow rate estimation device 1. The receiving unit 120 can store each transmitted information in the storage unit 110.

[0074] The reception unit 130 receives information input using an input means (not shown) connected to the river flow rate estimation device 1 or information input by an input operation (including information read by the device, for example) using a reading device (such as a barcode reader) connected to the information processing device 1. The received information is stored, for example, in the storage unit 110. The input means that can be used for inputting information receivable by the reception unit 130 may be anything, such as a numeric keypad, a keyboard, a mouse, or a menu screen. The reception unit 130 can be realized by a device driver of an input means such as a numeric keypad or a keyboard, or control software of a menu screen. Note that the reception unit 130 may receive information such as voice input by a microphone, for example.

[0075] Incidentally, the reception unit 130 may be regarded as receiving the information received by the reception unit 120 as the information input to the river flow rate estimation device 1. That is, the input of information to the river flow rate estimation device 1 may be interpreted to mean that these pieces of information are indirectly input to the river flow rate estimation device 1 by the user via a terminal device or the like, or may be interpreted to mean that they are directly input to the river flow rate estimation device 1 by the user using the input means. Further, it may be regarded that information is given to the river flow rate estimation device 1 when the user executes a program for automatically generating information or causes various information to function by giving it to the program.

[0076] The processing unit 140 can generally be realized from an MPU, a memory, or the like. The processing procedure of the processing unit 140 is generally realized by software, and the software is recorded on a recording medium such as a ROM. However, it may be realized by hardware (a dedicated circuit). The processing unit 140 performs various processes. The various processes are, for example, the processes performed by each part of the processing unit 140 as follows.

[0077] The processing unit 140 includes an image acquisition unit 141, a basin data acquisition unit 142, a candidate model acquisition unit 143, an existing data acquisition unit 144, a verification unit 145, an estimation model acquisition unit 147, a model storage unit 149, a weather data acquisition unit 151, a flow rate information acquisition unit 153, a water conservation area information acquisition unit 156, a flow rate information output unit 163, and a water conservation area information output unit 166. The processing unit 140 performs various processes. The various processes are, for example, the processes performed by each part of the processing unit 140 as follows.

[0078] The image acquisition unit 141 acquires an image including the basin of a river (hereinafter sometimes referred to as the target river) for which the estimation model is to be generated. The image including the basin refers to an image including an image in which a part of the basin is shown. In the present embodiment, the image including the basin is, for example, a satellite photo, an aerial photo, or an image obtained by constructing an image of the basin based on information of the basin prepared in advance. The image obtained by constructing an image of the basin based on information of the basin prepared in advance is, for example, a map image constructed based on data of a digital elevation model (DEM) prepared in advance and showing changes in color, density, etc. for each elevation, but is not limited thereto. The acquisition of an image includes the concept of acquiring an image stored in the storage unit 110 in advance, acquiring an image from an external server, an image providing service, etc., and acquiring an image taken by instructing a satellite, an aircraft, etc. (for example, a drone may also be used) to photograph the basin.

[0079] Here, in the present embodiment, the image acquisition unit 141 may be configured to acquire an image including the basin based on the position information indicating the basin of the target river. Here, the position information is information capable of specifying longitude and latitude, and is, for example, information indicating longitude and latitude or information indicating a place name associated with longitude and latitude, etc. in advance. The image acquisition unit 141 acquires, for example, an image corresponding to the position information indicating the basin of the target river among the images stored in the storage unit 110 in association with the position information in advance. Further, the image acquisition unit 141 may be configured to transmit the position information indicating the basin of the target river to a satellite, an aircraft, etc. to instruct photographing and acquire the photographed image.

[0080] The catchment data acquisition unit 142 acquires catchment data regarding the catchment of the target river. In the present embodiment, the catchment data acquisition unit 142 acquires catchment data from an image including the catchment. For example, the catchment data acquisition unit 142 acquires land use data of each unit area from the image, and acquires information regarding hydrogeology, etc. of each unit area based on the land use data. Here, the catchment data acquisition unit 142 is configured to acquire catchment data regarding each unit area by using the image and the learning information stored in the learning information storage unit 111. That is, for example, the catchment data acquisition unit 142 applies the acquired image to the learning information, and acquires land use data as the output information thereof. Then, by referring to a table in which information regarding the use of land and hydrogeology, etc. is associated in advance, information regarding the hydrogeology, etc. of the unit area is acquired based on the use of land of the unit area.

[0081] Note that the learning information may be configured to output information regarding hydrogeology, etc. In that case, the catchment data acquisition unit 142 may obtain information regarding the hydrogeology, etc. of each unit area as output information obtained by applying the image to the learning information.

[0082] Note that the catchment data acquisition unit 142 may be configured to acquire information regarding the hydrogeology, etc. of each unit area without using the learning information. For example, literature values regarding the geology of the catchment of the target river may be prepared in advance so as to be acquirable, and information regarding the hydrogeology, etc. of each unit area may be acquired based on the literature values.

[0083] Also, in the present embodiment, the catchment data acquisition unit 142 acquires information regarding the topography of each unit area of the catchment. The catchment data acquisition unit 142 may acquire information regarding the topography of each unit area, for example, by acquiring pre-constructed map information, etc. regarding the catchment of the target river from the storage unit 110, etc. Further, for example, data of a so-called digital elevation model (DEM) may be acquired, and information regarding each unit area may be acquired based on the data.

[0084] In addition, the image acquisition unit 141 may acquire a map image constructed based on the digital elevation model data, and the watershed data acquisition unit 142 may acquire information on the terrain such as the elevation and gradient of each unit area using the map image. In this case, for example, by applying the map image to the learning information that is configured to output information on the terrain of each unit area with the pre-created map image as input information, the information on the terrain of each unit area may be configured to be acquirable as the output information.

[0085] The candidate model acquisition unit 143 acquires one or more candidate models for estimating information on the river flow based on the watershed data. A candidate model is a model that is a candidate for the estimation model. In the present embodiment, the candidate model is a rainfall-runoff-flood model.

[0086] Here, the acquisition of one or more candidate models is performed by applying a constant group of parameters other than the parameters related to the meteorological data among the parameters of the numerical calculation model. In the present embodiment, the candidate model acquisition unit 143 prepares several constant groups to be set (configured) in the parameters of the rainfall-runoff-flood model corresponding to the watershed of the target river, and generates a plurality of candidate models with each constant group set in the parameters.

[0087] Note that the constant values of the parameters set here are, for example, information on the terrain of the unit area and information on hydrogeology and the like. For example, constant groups related to gradient, river terrain (such as movable cross-section), roughness coefficient, soil layer thickness, saturated effective porosity, permeability coefficient, groundwater outflow coefficient, etc. are set.

[0088] The candidate model acquisition unit 143 can prepare the constant groups to be set in two or more candidate models, for example, as follows.

[0089] The candidate model acquisition unit 143 first sets the values of each parameter so as to correspond to the basin data acquired by the basin data acquisition unit 142, and uses them as a set of constant groups. That is, the candidate model acquisition unit 143 sets the values of each parameter to be reasonable values in light of the basin data. For example, for a parameter with corresponding basin data, the value corresponding to the basin data is set. Note that the value obtained by performing an operation by a predetermined operation method based on the basin data may be set as the value of the parameter.

[0090] Further, the candidate model acquisition unit 143 changes the value of at least one parameter among the set of constant groups, and uses this as a set of constants (different set of constants) for applying to different candidate models. In the present embodiment, the candidate model acquisition unit 143 changes the value of at least one parameter by a predetermined rate from the value of the parameter set by the candidate model acquisition unit 143 to obtain different sets of constants. Specifically, for example, for one parameter, the value is increased or decreased by 1% to obtain different sets of constants. Note that the rate of change is not limited to this. Also, different sets of constants may be obtained by adding or subtracting a preset change amount for each parameter to the set value.

[0091] The existing data acquisition unit 144 acquires information on the flow rate of the target river from the hydrological database. The hydrological database is, for example, a database in which information on the flow rate of each past river is accumulated for each river. In the present embodiment, for example, the hydrological database is stored in an external information processing device accessible by the river flow rate estimation device. For example, the hydrological database is available via the Internet. Note that the hydrological database is not limited to such a configuration, and may be stored in the storage unit 110, for example. The existing data acquisition unit 144 acquires information on the flow rate of the target river from the hydrological database using, for example, information for specifying the target basin (for example, an identifier such as the name of the target river) input by the user.

[0092] In this embodiment, for example, HydroATLAS (Linke et al., 2019) can be used as the hydrological database. Such a hydrological database is an organized collection of existing datasets related to the hydrology of river basins and rivers at a resolution of 15 seconds. The original data uses data from the climate database for the period from 1950 to 2000. The hydrological database that can be used is not limited to this, and those composed of observation data related to the target river, etc., may be appropriately used.

[0093] In this embodiment, the existing data acquisition unit 144 is configured to acquire the minimum flow rate of the target river from the hydrological database. Note that the information to be acquired is not limited to the minimum flow rate, and may be other information related to the flow rate, such as the average flow rate of the target river.

[0094] The verification unit 145 verifies the validity of at least one of the one or more candidate models acquired by the candidate model acquisition unit 143 using the verification data.

[0095] Here, the verification data may include, for example, one or more predetermined equations (verification equations) related to the flow rate of the river and the values of the parameters applicable to the verification equations and candidate models. It can be said that the verification unit 145 verifies the validity of the candidate model using the verification equation, at least one candidate model, and the values applicable to the parameters of the equation and the parameters of the candidate model.

[0096] Note that the verification unit 145 uses meteorological data when performing verification. The meteorological data may be based on measured values or may be hypothetically set.

[0097] In the present embodiment, the verification unit 145 verifies the candidate model using information regarding the flow rate obtained using the candidate model to be verified and information obtained by applying the value of the parameter corresponding to the basin data to the verification formula (verification using the verification formula). The value of the parameter corresponding to the basin data may be any value such as a value included in the basin data, a value obtained using a value included in the basin data, a set of constant groups prepared based on the basin data, or a value included in a different set of constant groups. For example, the verification unit 145 applies the value of the parameter to be used among the constant groups used for obtaining the candidate model to be verified to the verification formula. Further, the verification unit 145 performs calculations for each of the candidate model and the verification formula by applying common meteorological data to each of the candidate model and the verification formula. Then, the information obtained by the candidate model and the information obtained by the verification formula are compared, and based on the result, the validity of the candidate model is verified. For example, if the information obtained by the candidate model and the information obtained by the verification formula match each other or the states represented by both are close, it can be verified that the validity of the candidate model is high. On the contrary, if the information obtained by the candidate model and the information obtained by the verification formula are different, it can be said that the validity of the candidate model is not high.

[0098] Here, in the present embodiment, as one or more predetermined verification formulas, an isochrone formula (for example, the so-called Manning's formula) representing the relationship between water level and flow rate and a rational formula representing peak flow rate are used. That is, in the present embodiment, the verification unit 145 verifies the candidate model using information regarding the flow rate obtained using the candidate model to be verified, information obtained by applying the value of the parameter corresponding to the basin data to the isochrone formula, and information obtained by applying the value of the parameter corresponding to the basin data to the rational formula. Note that at least one of the isochrone formula and the rational formula may be used as the verification formula. The isochrone formula may be said to represent the average flow velocity.

[0099] The verification using the isochrone formula and the rational formula can be performed, for example, as follows.

[0100] The verification unit 145 obtains the relationship between the flow rate and the water depth in one or more unit areas that are rivers, using the equal-flow formula. Further, the verification unit 145 obtains the relationship between the flow rate and the water depth obtained by the candidate model. Then, it determines whether or not the correlation between the two satisfies the evaluation conditions. For example, the verification unit 145 determines whether or not the correlation coefficient between the two is equal to or greater than a predetermined value (an example of the evaluation conditions).

[0101] On the other hand, the verification unit 145 obtains the peak flow rate of the target river using a rational formula. Further, the verification unit 145 obtains the maximum value among the water flow rates of each unit area by the candidate model. Then, the two are compared to determine whether or not the result satisfies the evaluation conditions. For example, the verification unit 145 determines whether or not the difference between the peak flow rate obtained from the rational formula and the maximum flow rate obtained by the candidate model is smaller than a predetermined value (an example of the evaluation conditions), or whether it is within a predetermined error range (an example of the evaluation conditions).

[0102] Also, in the present embodiment, in addition to the verification using the verification formula as described above, the verification unit 145 performs verification using the information acquired by the existing data acquisition unit 144. That is, the verification unit 145 performs verification of the candidate model using at least the information acquired by the existing data acquisition unit 144 and the information regarding the flow rate obtained using the candidate model to be verified (verification using existing data). More specifically, for example, the verification unit 145 performs verification of the candidate model using the base flow rate obtained using the candidate model to be verified and the minimum flow rate of the target river acquired by the existing data acquisition unit 144. Note that the information acquired by the existing data acquisition unit 144 may be expressed as verification data.

[0103] Verification using such existing data can be performed, for example, by using a candidate model to be verified, assuming a case where there is no rainfall for a predetermined time, taking the flow rate of the target river obtained as the base flow rate, comparing it with the minimum flow rate obtained from the hydrological database, and determining whether the result satisfies the evaluation conditions. For example, the evaluation condition is that the order of the value of the base flow rate and the minimum flow rate obtained from the hydrological database is within a predetermined range. In addition to the evaluation conditions regarding the order as described above, the evaluation conditions may be that the difference between the base flow rate and the minimum flow rate is smaller than a predetermined value, or that it falls within a predetermined error range.

[0104] In the present embodiment, the verification unit 145 evaluates the validity of the candidate model based on the determination result using the equal flow formula and the determination result using the rational formula as described above. That is, the verification unit 145 evaluates the validity of the candidate model by performing verification using the verification formula. For example, when both the determination result using the equal flow formula and the determination result using the rational formula indicate that the validity of the candidate model is high, the candidate model can be evaluated as being valid. In addition, when any of the determination results indicates that the validity of the candidate model is high, the candidate model may be evaluated as being valid.

[0105] In addition, the verification unit 145 evaluates the validity of the candidate model based on the determination result using the information acquired by the existing data acquisition unit 144. That is, the verification unit 145 evaluates the validity of the candidate model by performing verification using the existing data. For example, when it is determined that the comparison result between the base flow rate and the minimum flow rate obtained from the hydrological database satisfies the evaluation conditions, the candidate model can be evaluated as being valid.

[0106] In the present embodiment, the verification unit 145 is configured to perform verification using existing data on candidate models that have been evaluated as being valid by verification using a verification formula. In this case, the verification unit 145 may perform verification using existing data on the candidate model evaluated as having the highest validity, or may perform verification using existing data on each of the candidate models evaluated as having a certain level of validity. For example, the verification unit 145 may perform verification using existing data on candidate models determined by the estimation model acquisition unit 147 to satisfy the first acquisition condition or candidate models determined to satisfy the second acquisition condition, as described later. And the verification unit 145 is configured to evaluate as having the highest validity the candidate models evaluated as being valid by verification using existing data. Note that the verification method is not limited to this. For example, verification using a verification formula may be performed on all candidate models and verification using existing data may be performed on each, and the results may be combined to evaluate whether each candidate model is valid. Also, verification using existing data may be performed on each of the candidate models, and verification using a verification formula may be performed on the candidate models evaluated as being valid. It is possible to evaluate whether a candidate model is valid by verification using a verification formula or verification using existing data, but when both verification using a verification formula and verification using existing data are performed, it is possible to more appropriately evaluate whether the candidate model is valid. Therefore, when both verification using a verification formula and verification using existing data are performed, it is possible to obtain a candidate model capable of estimating the river flow rate with higher accuracy.

[0107] Based on the verification result of the verification unit 145, the estimation model acquisition unit 147 acquires one of the one or more candidate models as an estimation model. In the present embodiment, the estimation model acquisition unit 147 acquires, as an estimation model, a candidate model evaluated as having higher validity as a result of the verification unit 145 evaluating the validity of each of two or more candidate models.

[0108] For example, the estimated model acquisition unit 147 identifies a candidate model that satisfies both the verification result using the equal flow formula and the verification result using the rational formula satisfying a predetermined level (the first acquisition condition) among the candidate models, and that the verification result using the equal flow formula (for example, the correlation coefficient, etc.) indicates the highest validity (the second acquisition condition). Then, the estimated model acquisition unit 147 can acquire, as the estimated model, a candidate model that satisfies the second acquisition condition and that is evaluated as being appropriate by verification using existing data (the third acquisition condition). Here, the second acquisition condition may be that the verification result using the rational formula indicates the highest validity (the difference between the maximum flow rate and the peak flow rate according to the rational formula is the smallest). Also, for example, the estimated model acquisition unit 147 may score the evaluation result of the candidate model by the verification unit 145 and acquire, as the estimated model, a candidate model that satisfies the condition that the score is the highest (another example of the acquisition condition).

[0109] The model storage unit 149 stores the estimated model acquired by the estimated model acquisition unit 147. The estimated model is stored in the model storage unit 117. The estimated model may be stored, for example, in association with an identifier that identifies the target river and its basin.

[0110] Note that among the river flow rate estimation devices 1, it can be considered that an estimated model generation device that generates an estimated model is configured by each part of the above-described processing unit 140, the storage unit 110, the reception unit 120, and the reception unit 130.

[0111] Also, the river flow rate estimation device 1 can acquire information regarding the flow rate of the target river (hereinafter sometimes referred to as flow rate information) using the generated estimated model, using the following parts of the processing unit 140.

[0112] The meteorological data acquisition unit 151 acquires meteorological data of the basin of the target river. For example, it acquires the rainfall distribution at the time corresponding to the time when the flow rate is estimated as meteorological data. Meteorological data stored in advance in the storage unit 110 or the like in association with an identifier capable of specifying the basin may be acquired, or meteorological data may be acquired from an external server or the like. For example, precipitation data at an observation point close to the basin accumulated in the "Global Satellite Mapping of Precipitation (GSMaP)" can be acquired as meteorological data, but it is not limited to this.

[0113] The flow rate information acquisition unit 153 acquires the flow rate information of the target river using the estimation model stored in the model storage unit 149. The flow rate information acquisition unit 153 acquires the flow rate information of the target river by applying the acquired meteorological data to the estimation model corresponding to the target river. As the flow rate information, for example, the river flow rate of a single unit area may be obtained, but it is not limited to this. For example, the river water level may be acquired. Also, although not directly, information such as the groundwater level and the depth of inundation at a predetermined point may be acquired as flow rate information as information related to the flow rate of the river indirectly. Further, the flow rate information acquisition unit 153 may acquire information such as the groundwater level and the depth of inundation as flow rate information in addition to at least one of the river flow rate and the river water level.

[0114] The flow rate information output unit 163 outputs information, for example, by displaying the information on a display device provided in the river flow rate estimation device 1 in the present embodiment. Note that the flow rate information output unit 163 may be configured to output information by transmitting the information to another device via a network or the like using, for example, the transmission unit 170 or the like. Note that the flow rate information output unit 163 may be considered to include output devices such as a display and a speaker, or may not be considered to include them. The flow rate information output unit 163 can be realized by the driver software of the output device or the driver software of the output device and the output device or the like.

[0115] In the present embodiment, the flow rate information output unit 163 outputs the flow rate information acquired by the flow rate information acquisition unit 153. The flow rate information output unit 163 outputs, for example, by associating the flow rate information with information regarding the position corresponding to the flow rate information, so as to be displayed in a predetermined format on a display, or transmits it to the user's terminal device. Note that only the flow rate information may be output. Also, for example, the output may be performed such that the flow rate information at each point of the river is displayed on a display or the like in a display mode corresponding to the flow rate information. Specifically, it may be displayed such that the higher the river flow rate or the higher the river water level, the higher (or lower) the saturation, density, brightness, etc. becomes, or it may be displayed in a predetermined color. Also, for example, the flow rate information at each point may be mapped and displayed on a map showing the basin or an image including the basin. By doing so, the flow rate information can be presented to the user in a visually easy-to-understand manner.

[0116] Here, in the present embodiment, in addition to or instead of the flow rate information, the specific discharge of each unit area calculated using the flow rate information may be acquired, and information regarding the specific discharge may be output. For example, the flow rate information acquisition unit 153 calculates and acquires the specific discharge of each region previously partitioned in the basin using the result output using the estimation model. The information regarding the specific discharge may also be referred to as the flow rate information.

[0117] In this case, information regarding the water conservation in the basin may be output. That is, the water conservation area information acquisition unit 156 acquires information regarding a unit area that satisfies a predetermined water conservation condition using the specific discharge of each area. Then, the water conservation area information output unit 166 outputs the acquired information, that is, the information regarding the water conservation area. For example, as the water conservation condition, it can be set that the groundwater level is greater than a predetermined value, or the vertical infiltration flow is equal to or greater than a predetermined value, etc. Thereby, the relative water conservation amount of each unit area that can be understood from the specific discharge can be grasped. Further, the water conservation area information output unit 166 outputs, for example, the unit area (water conservation area) that satisfies the water conservation condition in a display mode different from other areas, or displays the value of an index regarding water conservation such as the groundwater level for the water conservation area. Thereby, the user can easily grasp the dominant area regarding the water source conservation function in the basin of the target river, or grasp the area that should be paid attention to compared with other areas regarding the water source conservation function. When considering measures regarding the water source conservation function in the basin of the target river, the possibility of effectively using the necessary labor is increased. Note that information regarding the groundwater conservation area may be output based on the spatial distribution of the specific discharge of each area.

[0118] Note that the specific discharge of each unit area based on two or more different meteorological data may be acquired using an estimation model. And as the water conservation condition, for each unit area, a condition regarding the specific discharge acquired corresponding to a plurality of meteorological data may be set. For example, a condition such as the average value of a plurality of specific discharges corresponding to a plurality of meteorological data being equal to or greater than a predetermined value may be set.

[0119] The transmission unit 170 is usually realized by wireless or wired communication means, but may be realized by a broadcast means. The transmission unit 170 transmits information to another device that is communicably connected to the river flow estimation device 1 via a network. The transmission unit 170 transmits information to, for example, a terminal device operable by the user. In other words, the transmission unit 170 outputs information to, for example, the terminal device.

[0120] In the above description, the estimation model is not limited to simulation models or numerical calculation models such as the above-described rainfall-runoff flood model. For example, it may be learning information (learning device, trained model) of machine learning. That is, the watershed data acquisition unit 142 acquires teacher data, which is a combination of meteorological data, flow rate information (for example, measured values, etc.), land use data, information related to hydrogeology, etc., and information related to topography at a past time point, and the candidate model acquisition unit 143 may construct one or more pieces of learning information by a machine learning method and use it as a candidate model. In this case, the verification data may be the same as the teacher data, that is, past meteorological data and flow rate information, to verify the validity of the candidate model.

[0121] FIG. 4 is a flowchart for explaining the use of the river flow rate estimation device 1.

[0122] Using such a river flow rate estimation device 1, users such as business operators can grasp the flow rate information of the target river and information related to the recharge based thereon. The river flow rate estimation device 1 can be used, for example, in the following manner. Note that the following is an example, and various modifications are possible depending on the configuration and use of the river flow rate estimation device 1.

[0123] (Step S801) First, the user determines the watershed of the target river for which the flow rate information is to be obtained. Specifically, for example, when grasping the flow rate of a river around a factory that uses water resources, the recharge area such as the mountain forest upstream of the factory, the area where the factory withdraws groundwater or river water downstream thereof, and further the area downstream thereof can be selected as the target area. The user inputs information specifying the target watershed to the river flow rate estimation device 1. For example, position information indicating the target watershed is input, but it is not limited thereto. For example, an identifier such as the name of the target river may be input.

[0124] (Step S802) The image acquisition unit 141 acquires an image including the target area. For example, the image acquisition unit 141 acquires the image to be used based on the position information indicating the watershed of the target river.

[0125] (Step S803) The river flow rate estimation device 1 generates an estimation model and stores it in the model storage unit 117. As a result, it becomes possible to use the estimation model to acquire flow rate information. Note that the generation process of the estimation model can be performed, for example, as described later.

[0126] (Step S804) The river flow rate estimation device 1 acquires the assumed meteorological data and applies it to the estimation model. As a result, the flow rate information of the target river can be acquired.

[0127] (Step S805) The river flow rate estimation device 1 calculates the specific flow rate of each unit area of the basin.

[0128] (Step S806) The river flow rate estimation device 1 identifies an area that satisfies a predetermined water retention condition based on the specific flow rate.

[0129] (Step S807) The river flow rate estimation device 1 outputs information regarding the water retention area. As a result, the user can perform activities using the output information regarding the water retention area.

[0130] FIG. 5 is a flowchart showing an example of the estimation model generation process of the same river flow rate estimation device 1.

[0131] (Step S11) The basin data acquisition unit 142 acquires basin data.

[0132] (Step S12) The candidate model acquisition unit 143 generates a plurality of candidate models. In the present embodiment, as described above, a plurality of different candidate models are generated using different constant groups with changed parameter values respectively.

[0133] (Step S13) The verification unit 145 performs verification on the plurality of candidate models using a verification formula. The flow of verification using the verification formula will be described later.

[0134] (Step S14) The verification unit 145 verifies the candidate model to be processed using the existing data. Note that the candidate model to be processed may be the candidate model with the highest accuracy in the evaluation result of step S13, or two or more candidate models with high accuracy in the evaluation result of step S13, or all the generated candidate models. Note that if there is no candidate model determined to be appropriate by verification using the existing data among the candidate models to be processed, the verification unit 145 may return to step S12 or use the next most accurate candidate model in the evaluation result of step S13 as the model to be processed.

[0135] (Step S15) The estimation model acquisition unit 147 acquires, based on the verification result, the single candidate model with the highest accuracy as the estimation model.

[0136] (Step S16) The model storage unit 149 stores the acquired estimation model in the model storage unit 117 in association with the river basin.

[0137] FIG. 6 is a flowchart showing an example of the verification process using the verification formula of the verification unit 145.

[0138] (Step S21) The verification unit 145 sets zero to the counter i.

[0139] (Step S22) The verification unit 145 increments the counter i.

[0140] (Step S23) The verification unit 145 applies the meteorological data, which is the verification data, to the i-th candidate model to be verified. Thereby, the flow rate information is acquired.

[0141] (Step S24) The verification unit 145 acquires the set of constants used for generating the i-th candidate model.

[0142] (Step S25) The verification unit 145 performs an operation according to the verification formula using the set of constants, the verification formula which is the verification data, and the meteorological data.

[0143] (Step S26) The verification unit 145 evaluates the accuracy of the information obtained by the candidate model based on the flow rate information obtained by the candidate model and the calculation result by the verification formula. For example, obtaining a correlation coefficient, comparing the difference between the flow rate information and the calculation result for a specific index with a predetermined value, etc. are performed.

[0144] (Step S27) The verification unit 145 determines whether the counter i matches the total number of the generated candidate models. If they match, the verification ends and returns to the flowchart shown in FIG. 5. If they do not match, it returns to Step S22.

[0145] Note that the verification can be performed as follows, for example.

[0146] FIG. 7 is a first diagram for explaining the verification using the verification formula performed by the verification unit 145. FIG. 8 is a second diagram for explaining the verification using the verification formula performed by the verification unit 145.

[0147] In the examples shown in FIGS. 7 and 8, three estimation models “Pattern 1”, “Pattern 2”, and “Pattern 3” with different constant groups are verified using two types of verification formulas, an equal flow formula and a rational formula.

[0148] In FIG. 7, it is a graph showing the relationship between the flow rate information at a predetermined point obtained from the three estimation models and the flow rate - water level curve (H - Q curve) represented by the flow rate formula at the same point. According to the graph, the flow rate information by “Pattern 1” fits closest to the curve represented by the flow rate formula.

[0149] In FIG. 8, the peak flow rates of the three estimation models and the rational formula when using predetermined meteorological data are shown. For “Pattern 1” among the three estimation models, the error between the peak flow rate obtained by the estimation model and the peak flow rate obtained by the rational formula is the smallest.

[0150] In the above case, it can be seen that "Pattern 1" is the candidate model with the highest accuracy. Therefore, when "Pattern 1" is evaluated as being reasonable through verification using existing data, the estimation model acquisition unit 147 acquires "Pattern 1" as the estimation model.

[0151] FIG. 9 is a diagram for explaining the verification using existing data performed by the verification unit 145.

[0152] In FIG. 9, the vertical axis on the left shows the flow rate on a logarithmic axis, and the horizontal axis shows the day. FIG. 9 shows the transition of the estimated value of the flow rate of a certain target river calculated using one candidate model. Also, using the same axes, the base flow rate value of the target river obtained from the hydrological database (here, HydroATLAS) is shown by a dashed line. In addition, in FIG. 9, the daily rainfall (basin rainfall) in a certain basin of the target river is shown with respect to the vertical axis on the right. The transition of the estimated value of the flow rate is calculated based on this basin rainfall.

[0153] As shown in FIG. 9, it can be said that the minimum value of the estimated flow rate of the target river and the base flow rate obtained from the hydrological database are of approximately the same order. In such a case, the verification unit 145 can determine that this candidate model has high validity. Therefore, when this candidate model is determined to be reasonable through verification using the verification formula, the estimation model acquisition unit 147 acquires this candidate model as the estimation model.

[0154] In this embodiment, an estimation model, which is a distributed model, can be verified using the equal-discharge formula and the rational formula, which are independent verification formulas for different infiltration processes. According to the equal-discharge formula, an H-Q curve approximated by an equation based on in-situ measured values can be obtained, but there is a problem that the river flow rate cannot be calculated by the equal-discharge formula. According to the rational formula, parameters consistent with the flood discharge can be extracted, but there is a problem that only data at a certain point can be obtained. On the other hand, the estimation model can calculate continuous river flow rates. In the estimation model, since all values are estimated values, ensuring accuracy becomes an issue. However, since it has been verified that the accuracy is ensured by these verification formulas, an estimation model that has been verified can be used, so that it is possible to obtain flow rate information with a certain degree of reliability.

[0155] As described above, according to this embodiment, it is possible to obtain an estimation model useful for estimating information on the flow rate of the target river. Then, using the estimation model, information on the flow rate of the target river can be obtained. Since the estimation model is obtained based on the verification results using the verification data, it becomes possible to obtain more accurate flow rate information. In addition, the estimation model can be generated using an image including the basin. Therefore, it is not necessary to conduct on-site surveys such as on-site inspections for constructing the estimation model, and even if it is conducted, the amount of work can be reduced. Therefore, even when the target river is located in a remote area or the like, it is possible to easily and promptly estimate the flow rate information.

[0156] Further, in this embodiment, in the basin of the target river, it is possible to obtain information on the recharge area based on specific-discharge analysis after obtaining the flow rate information without necessarily visiting the site. Therefore, it is possible to easily identify areas that should be noted regarding recharge in the basin of each target river and efficiently engage in activities to maintain water sources and utilize water resources to enhance sustainability.

[0157] In addition, in the present embodiment, for the candidate model, both verification using a verification formula and verification using existing data are performed. Therefore, a candidate model capable of estimating the river flow rate with higher accuracy can be obtained.

[0158] Note that the generation and verification of the candidate model may be performed one by one. In this case, the acquisition and verification of the candidate model may be repeated until a predetermined condition is satisfied. Hereinafter, the estimation model generation process in the case of generating and verifying the candidate model one by one will be described.

[0159] FIG. 10 is a flowchart showing a modification of the estimation model generation process of the river flow rate estimation device 1.

[0160] (Step S211) The catchment data acquisition unit 142 acquires catchment data.

[0161] (Step S212) The candidate model acquisition unit 143 generates one candidate model. In this modification, one candidate model is generated using a set of constants of parameters.

[0162] (Step S213) The verification unit 145 verifies the generated one candidate model. The verification will be described later.

[0163] (Step S214) The estimation model acquisition unit 147 determines whether the candidate model satisfies a predetermined condition based on the verification result. If the candidate model satisfies the predetermined condition, the process proceeds to step S216; otherwise, the process proceeds to step S215. Satisfying the predetermined condition means, for example, that a verification result indicating that the candidate model has a certain degree of validity is obtained based on the verification result (for example, an index representing the accuracy of the candidate model is higher than a predetermined value, etc.).

[0164] (Step S215) The estimation model acquisition unit 147 changes the constant group of the parameters to different values. Note that the estimation model acquisition unit 147 sets a constant group whose content is not the same as the constant group set so far. Then, it returns to step S212. As a result, generation and verification of candidate models using the changed different constant groups are performed.

[0165] (Step S216) The estimation model acquisition unit 147 acquires, as an estimation model, one candidate model determined to satisfy a predetermined condition.

[0166] (Step S217) The model storage unit 149 stores the acquired estimation model in the model storage unit 117 in association with the river basin.

[0167] FIG. 11 is a flowchart showing an example of the verification process of the verification unit 145.

[0168] (Step S221) The verification unit 145 applies meteorological data, which is verification data, to the candidate model. As a result, flow rate information is acquired.

[0169] (Step S222) The verification unit 145 acquires the constant group used for generating the candidate model.

[0170] (Step S223) The verification unit 145 performs an operation according to the verification formula using the constant group, the verification formula which is verification data, and the meteorological data.

[0171] (Step S224) The verification unit 145 evaluates the accuracy of the information acquired by the candidate model based on the flow rate information acquired by the candidate model and the operation result by the verification formula.

[0172] (Step S225) The verification unit 145 verifies the candidate model using existing data. Then, it returns to the flowchart shown in FIG. 10.

[0173] Thus, in the modified example, until it is determined by the estimation model acquisition unit 147 that the verification result satisfies a predetermined condition, the acquisition of one candidate model by the candidate model acquisition unit 143 and the verification of the one candidate model by the verification unit 145 are repeated. Also in this modified example, the same effects as those of the above-described embodiment can be obtained. Further, as in this modified example, by generating and verifying candidate models one by one, when a candidate model with relatively high validity is obtained, it can be obtained as an estimation model relatively quickly.

[0174] Note that in the above-described step S213, verification using a verification formula is performed. When the candidate model satisfies a predetermined condition in step S214, verification using existing data may be performed before proceeding to step S216. In this case, when it is determined that the verification using existing data is appropriate, the process may proceed to step S216, and when not, the process may return to step S212.

[0175] Also, the acquisition of two or more candidate models by the candidate model acquisition unit 143 and the verification of the two or more candidate models by the verification unit 145 may be repeated until it is determined by the estimation model acquisition unit 147 that the verification result satisfies a predetermined condition.

[0176] Note that the processing in this embodiment may be realized by software. And this software may be distributed by software download or the like. Also, this software may be recorded on a recording medium such as a CD-ROM and distributed. The software that realizes the river flow rate estimation device 1 in this embodiment is a program as follows. That is, this program is a program executed by a computer of the river flow rate estimation device 1, and causes the computer of the river flow rate estimation device 1 to function as a flow rate information acquisition unit that acquires information regarding the flow rate of a target river using the estimation model accumulated by the estimation model generation device, and a flow rate information output unit that outputs information regarding the flow rate.

[0177] Also, the software that realizes the estimation model generation process of the river flow rate estimation device 1 in the present embodiment is a program as follows. That is, this program is a program executed by the computer of the river flow rate estimation device 1, and causes the computer of the river flow rate estimation device 1 to function as a watershed data acquisition unit that acquires watershed data of the target river basin from an image including the river basin of the target river, a candidate model acquisition unit that acquires one or more candidate models for estimating information related to the river flow rate based on the watershed data, a verification unit that verifies the validity of at least one of the one or more candidate models using verification data, an estimated model acquisition unit that acquires one of the one or more candidate models as an estimated model based on the verification result, and a model accumulation unit that accumulates the acquired estimated model.

[0178] (Others)

[0179] FIG. 12 is an overview diagram of the computer system 800 in the above embodiment. FIG. 13 is a block diagram of the computer system 800.

[0180] In these figures, the configuration of a computer that executes the program described in this specification and realizes the river flow rate estimation device and the like of the above-described embodiment is shown. The above-described embodiment can be realized by computer hardware and a computer program executed thereon.

[0181] The computer system 800 includes a computer 801 including a CD-ROM drive, a keyboard 802, a mouse 803, and a monitor 804.

[0182] In addition to the CD-ROM drive 8012, the computer 801 includes an MPU 8013, a bus 8014 connected to the CD-ROM drive 8012 and the like, a ROM 8015 for storing programs such as a boot-up program, a RAM 8016 connected to the MPU 8013 for temporarily storing instructions of an application program and providing a temporary storage space, and a hard disk 8017 for storing an application program, a system program, and data. Here, although not shown, the computer 801 may further include a network card for providing connection to a LAN.

[0183] A program for causing the computer system 800 to execute functions such as the information processing apparatus according to the above-described embodiment may be stored in a CD-ROM 8101, inserted into the CD-ROM drive 8012, and further transferred to the hard disk 8017. Alternatively, the program may be transmitted to the computer 801 via a network (not shown) and stored in the hard disk 8017. The program is loaded into the RAM 8016 during execution. The program may be loaded directly from the CD-ROM 8101 or the network.

[0184] The program does not necessarily include an operating system (OS) for causing the computer 801 to execute functions such as the information processing apparatus according to the above-described embodiment, or a third-party program. The program only needs to include a portion of instructions that call appropriate functions (modules) in a controlled manner so as to obtain a desired result. How the computer system 800 operates is well known, and a detailed description thereof will be omitted.

[0185] Note that in the above program, in a transmission step of transmitting information, a reception step of receiving information, and the like, processing performed by hardware, for example, processing performed by a modem or an interface card in the transmission step (processing that can only be performed by hardware) is not included.

[0186] Also, the computer that executes the above program may be singular or plural. That is, centralized processing may be performed, or distributed processing may be performed.

[0187] Also, in the above embodiment, two or more components existing in one device may be physically realized by one medium.

[0188] Also, in the above embodiment, each process (each function) may be realized by being centrally processed by a single device (system), or may be realized by being distributedly processed by a plurality of devices (in this case, the entire system composed of a plurality of devices performing distributed processing can be grasped as one "device").

[0189] Also, in the above embodiment, the transfer of information performed between each component, for example, when the two components that transfer the information are physically different, may be performed by the output of information by one component and the reception of information by the other component, or when the two components that transfer the information are physically the same, it may be performed by moving from the processing phase corresponding to one component to the processing phase corresponding to the other component.

[0190] Also, in the above embodiment, information related to the processing executed by each component, for example, information received, acquired, selected, generated, transmitted, or received by each component, or information such as thresholds, mathematical formulas, addresses, etc. used in the processing by each component, may be temporarily or long-term held in a recording medium not shown even if not specified in the above description. Also, the accumulation of information on the recording medium not shown may be performed by each component or an accumulation unit not shown. Also, the reading of information from the recording medium not shown may be performed by each component or a reading unit not shown.

[0191] In the above-described embodiment, when information used in each component or the like, for example, information such as threshold values, addresses, and various setting values used by each component in processing may be changed by the user, even if not specified in the above description, the user may appropriately change such information, or may not. When the user can change such information, the change may be realized, for example, by an unillustrated reception unit that receives a change instruction from the user and an unillustrated change unit that changes the information in response to the change instruction. The reception of the change instruction by the unillustrated reception unit may be, for example, reception from an input device, reception of information transmitted via a communication line, or reception of information read from a predetermined recording medium.

[0192] The present invention is not limited to the above-described embodiments, and various modifications are possible, and these are also included in the scope of the present invention. Also, among the above-described embodiments, some components and functions may be omitted.

[0193] In the above-described embodiment, the verification unit performs both verification using a verification formula and verification using existing data, but is not limited to this. For example, only verification using a verification formula may be performed, and verification using existing data may not be performed. In this case, information may not be acquired from the hydrological database, and the existing data acquisition unit may not be provided. Also, only verification using existing data may be performed, and verification using a verification formula may not be performed.

Industrial Applicability

[0194] As described above, the river flow rate estimation device according to the present invention has an effect of being able to estimate information regarding the flow rate of the target river, and is useful as a river flow rate estimation device or the like.

Explanation of Signs

[0195] 1 River flow rate estimation device (estimation model generation device) 110 Storage unit 111 Learning information storage unit 113 Watershed data storage unit 117 Model storage unit 120 Receiver 130 Reception unit 140 Processing unit 141 Image acquisition unit 142 Watershed data acquisition unit 143 Candidate model acquisition unit 144 Existing data acquisition unit 145 Verification unit 147 Estimated model acquisition unit 149 Model accumulation unit 151 Meteorological data acquisition unit 153 Discharge information acquisition unit 156 Water conservation area information acquisition unit 163 Discharge information output unit 166 Water conservation area information output unit 170 Transmitter

Claims

1. A basin data acquisition unit that acquires basin data of the basin included in an image of a target river basin; A candidate model acquisition unit that acquires one or more candidate models for estimating information regarding the flow rate of the river based on the basin data; A verification unit that verifies the validity of at least one of the one or more candidate models using verification data; An estimated model acquisition unit that acquires, based on the verification result, one of the one or more candidate models as an estimated model; A model storage unit that stores the acquired estimated model, and The candidate model is a candidate model for estimating the information regarding the flow rate of the river based on input data including meteorological data, The verification unit performs verification of the candidate model using information regarding the flow rate obtained using the candidate model to be verified and information obtained by applying the value of a parameter corresponding to the basin data to one or more predetermined equations regarding the flow rate of the river. An estimated model generation device.

2. The candidate model is a candidate model for estimating the information regarding the flow rate of the river based on input data including, in addition to the meteorological data, information regarding topography including one or more of elevation information, gradient information, and river topography information, or information regarding hydrogeology including one or more of roughness coefficient, soil layer thickness, saturated effective porosity, permeability coefficient, and groundwater outflow coefficient. The estimated model generation device according to Claim 1.

3. The one or more predetermined equations include at least one of an isoflux formula representing the relationship between water level and flow rate and a rational formula representing peak flow rate. The estimated model generation device according to Claim 1.

4. The verification unit performs verification of the candidate model using information regarding the flow rate obtained using the candidate model to be verified, information obtained by applying the value of a parameter corresponding to the basin data to the isoflux formula, and information obtained by applying the value of a parameter corresponding to the basin data to the rational formula. The estimated model generation device according to Claim 3.

5. Further comprising an existing data acquisition unit that acquires information regarding the flow rate of the river from a hydrological database, The verification unit performs verification of the candidate model using at least the information acquired by the existing data acquisition unit and the information regarding the flow rate obtained using the candidate model to be verified. The estimated model generation device according to Claim 1.

6. The existing data acquisition unit acquires the minimum flow rate of the river from a hydrological database, The verification unit verifies the candidate model using the base flow rate obtained using the candidate model to be verified and the minimum flow rate acquired by the existing data acquisition unit. The estimation model generation device according to claim 5.

7. The estimation model generation device according to claim 1, wherein acquisition of one or more candidate models by the candidate model acquisition unit and verification of the one or more candidate models by the verification unit are repeated until it is determined by the estimation model acquisition unit that the verification result satisfies a predetermined condition.

8. The estimation model is a rainfall runoff flood model, The basin data includes land use data. The estimation model generation device according to claim 1.

9. The apparatus further includes an image acquisition unit that acquires an image to be used based on position information indicating a basin of a target river, The basin data acquisition unit acquires basin data of the basin from the acquired image. The estimation model generation device according to claim 1.

10. An estimation model generation device according to claim 1, A flow rate information acquisition unit that acquires information regarding the flow rate of a target river using the estimation model stored in the model storage unit, A river flow rate estimation device including a flow rate information output unit that outputs the information regarding the flow rate.

11. The apparatus further includes a meteorological data acquisition unit that acquires meteorological data of a basin of a target river, The flow rate information acquisition unit applies the meteorological data to the estimation model to acquire the information regarding the flow rate. The river flow rate estimation device according to claim 10.

12. The flow rate information acquisition unit calculates and acquires the specific flow rate of each region previously partitioned in the basin using the result output using the estimation model. The river flow rate estimation device according to claim 10.

13. A recharge area information acquisition unit that acquires information regarding an area that satisfies a predetermined recharge condition using the specific flow rate of each basin, The river flow rate estimation device according to claim 12, further including a recharge area information output unit that outputs the information regarding the area.

14. A method for manufacturing an estimation model, comprising all steps performed by the estimation model generation device according to any one of claims 1 to 9.

15. A river flow rate estimation method, comprising all steps performed by the river flow rate estimation device according to any one of claims 10 to 13.

16. A computer, A program for causing a computer to function as the estimation model generation device according to any one of claims 1 to 9.

17. A computer, A program for causing a computer to function as the river flow rate estimation device according to any one of claims 10 to 13.

Images