Prediction device

Abstract

This prediction device comprises a control unit. The control unit constructs a first prediction model from a plurality of pieces of terrain data for first learning and information regarding the presence or absence of damage at a plurality of first points on land where landslide occurred in the past. The plurality of pieces of terrain data for first learning are pieces of data respectively indicating the terrain of a plurality of predetermined regions respectively including the plurality of first points. The information regarding the presence or absence of damage at the plurality of first points is information indicating whether or not landslide has occurred at each of the plurality of first points. The first prediction model outputs the risk of landslide when terrain data is inputted. The control unit inputs, into the first prediction model, terrain data for first prediction indicating the terrain of a prescribed region including a first target point on land subject to prediction, and thereby predicts the risk of landslide at the first target point.

Description

Prediction device

[0001] This disclosure relates to a prediction device.

[0002] Conventionally, various technologies have been known for predicting the risk of landslides during heavy rainfall.

[0003] For example, the Central Research Institute of Electric Power Industry has developed a typhoon wind speed prediction system (RAMP-T: Risk Assessment and Management system for Power lifeline - Typhoon) that predicts wind speed during typhoons (Non-Patent Literature 1). They have also developed an earthquake damage estimation system (RAMP-Er: Risk Assessment and Management system for Power lifeline - Earthquake real time). A technology has also been developed to display the combined analysis results of this earthquake damage estimation system and the typhoon wind speed prediction system on a map (Non-Patent Literature 2). Furthermore, a system that predicts and displays the number of buildings damaged by typhoons, heavy rains, and earthquakes on a map is also known (Non-Patent Literature 3).

[0004] Regarding landslides, a technology is known for issuing landslide warnings based on a predicted soil moisture index using a tank model and 60-minute rainfall (Non-Patent Literature 4). In addition, a draft landslide occurrence probability map based on topographic and geological factors, created by the National Institute for Land and Infrastructure Management, is known (Non-Patent Literature 5). This draft landslide occurrence probability map calculates the probability of landslide occurrence in three stages, from probability category 1 to 3, based on landslide warning areas, deep-seated collapse frequency maps, and landslide topographic distribution maps. Furthermore, a landslide occurrence prediction support system has been developed by the National Research Institute for Earth Science and Disaster Resilience (Non-Patent Literature 6).

[0005] Soichiro Sugimoto et al., "Introduction of a System for Estimating Damage to Facilities Caused by Typhoons and Meteorological Simulation," IEEJ Journal, Vol. 138, No. 3, pp. 141-144, 2018, T. Tadokoro et al., "Automatic Generation of Input Data for Distribution System Simulation Programs", 2018 IEEE Innovative Smart Grid Technologies - Asia (ISGT Asia), IEEE, 2018 "Real-time damage prediction website / app cmap", [online], [October 21, 2024], Internet <URL: https: / / aioinissaydowa.co.jp / corporate / service / cmap / > "Landslide disaster warning information / Landslide Kikikuru (Landslide risk distribution for heavy rain warnings)", [online], [October 21, 2024], Internet <URL: https: / / www.jma.go.jp / jma / kishou / know / bosai / doshakeikai.html> Matsuda Masayuki et al., "A Study on a Method for Estimating the Risk of Sediment-Related Disasters Nationwide Using Thematic Maps Related to Topography and Geology," National Institute for Land and Infrastructure Management Document, No. 1120, 2020, [online], [October 21, 2024], Internet <URL: https: / / www.nilim.go.jp / lab / bcg / siryou / tnn / tnn1120.htm>, "Predicting the Risk of Sediment-Related Disasters and Disseminating Information with Map Images - Development of a 'Sediment-Related Disaster Occurrence Prediction Support System'," [online], [October 21, 2024], Internet <URL: https: / / www.engineering-eye.com / interview / user / 14 / >

[0006] As described above, various technologies are known for predicting the risk of landslides during heavy rainfall. However, there is no technology that predicts the risk of landslides occurring in mountainous areas during heavy rainfall while taking into account the topography of individual plots of land in those areas. For example, the parameters of the tank model described in Non-Patent Document 4 are standardized nationwide. Therefore, this tank model does not reflect information about the topography of individual plots of land. If the topography of individual plots of land is taken into consideration, the risk of landslides occurring in mountainous areas during heavy rainfall can be predicted with high accuracy.

[0007] In light of these points, the purpose of this disclosure is to accurately predict the risk of landslides occurring in mountainous areas during heavy rainfall.

[0008] A prediction device according to one embodiment of the present disclosure includes a control unit that constructs a first prediction model using a plurality of first learning terrain data and information on whether or not damage occurred at a plurality of first points on land where landslides have occurred in the past, wherein the plurality of first learning terrain data are data that each represents the terrain of a plurality of predetermined areas including each of the plurality of first points, the information on whether or not damage occurred at the plurality of first points is information that each represents whether or not a landslide occurred at each of the plurality of first points, the first prediction model outputs the risk of landslides when terrain data is input, and the control unit predicts the risk of landslides at the first target point by inputting first prediction terrain data, which represents the terrain of a predetermined area including a first target point on the land to be predicted, into the first prediction model.

[0009] According to one embodiment of this disclosure, the risk of landslides occurring in mountainous areas during heavy rainfall can be predicted with high accuracy.

[0010] This is a block diagram showing an example of a prediction device according to one embodiment of the present disclosure. This is a flowchart showing an example of the overall flow of the sediment disaster risk prediction process. This is a diagram for explaining the first target point, the sediment runoff route, and the second target point. This is a flowchart showing an example of the flow of the sediment collapse risk prediction process at the first target point. This is a diagram showing an example of elevation information. This is a diagram for explaining example 1 of setting a predetermined area. This is a diagram for explaining example 2 of setting a predetermined area. This is a diagram for explaining example 3 of setting a predetermined area. This is a diagram showing an example of a model formula. This is a diagram showing an example of a model formula. This is a flowchart showing an example of the sediment runoff route prediction process. This is a diagram showing an example of a collection. This is a diagram showing an example of a collection. This is a diagram showing an example of a collection. This is a flowchart showing an example of the sediment runoff route prediction process. This is a flowchart showing an example of the sediment disaster risk prediction process. This is a flowchart showing an example of the sediment disaster risk prediction process. This is a diagram for explaining a pattern for calculating the sediment disaster risk.

[0011] The embodiments relating to this disclosure will be described below with reference to the drawings.

[0012] A prediction device 10 according to one embodiment of the present disclosure, as shown in Figure 1, can predict the risk of sediment-related disasters in the land to be predicted. By predicting the risk of sediment-related disasters, countermeasures against sediment-related disasters can be taken in advance. For example, evacuation routes can be planned in advance when a sediment-related disaster occurs. Also, when clearing mountainous areas to install structures, predicting the risk of sediment-related disasters allows for the installation of structures that incorporate countermeasures against sediment-related disasters.

[0013] The prediction device 10 comprises an input unit 11, an output unit 12, a storage unit 13, and a control unit 14.

[0014] The input unit 11 is capable of receiving input from the user. The input unit 11 is configured to include at least one input interface capable of receiving input from the user. The input interface is, for example, a physical key, a capacitive key, a pointing device, a touchscreen integrated with a display, or a microphone.

[0015] The output unit 12 is capable of outputting data. The output unit 12 is configured to include at least one output interface capable of outputting data. The output interface is, for example, a display or a speaker. The display is, for example, an LCD (Liquid Crystal Display) or an organic EL (Electronic Luminescence) display.

[0016] The storage unit 13 is configured to include at least one semiconductor memory, at least one magnetic memory, at least one optical memory, or at least two combinations thereof. The storage unit 13 may function as a main memory, an auxiliary memory, or a cache memory. The storage unit 13 stores data used for the operation of the prediction device 10 and data obtained by the operation of the prediction device 10. The storage unit 13 may store a program executed by the control unit 14.

[0017] The control unit 14 is configured to include at least one processor, at least one dedicated circuit, or a combination thereof. The processor is, for example, a general-purpose processor such as a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit), or a dedicated processor specialized for a specific process. The dedicated circuit is, for example, an FPGA (Field-Programmable Gate Array) or an ASIC (Application Specific Integrated Circuit). The control unit 14 controls each part of the prediction device 10 and executes processes related to the operation of the prediction device 10.

[0018] [Prediction Process for Sediment-Related Disaster Risk] Figure 2 is a flowchart showing an example of the overall flow of the sediment-related disaster risk prediction process. When the control unit 14 receives an instruction from the user to execute the sediment-related disaster risk prediction process via the input unit 11, it starts the process of step S1. However, the control unit 14 may execute the process of step S1, the process of step S2, and the process of step S3 in parallel.

[0019] In step S1, the control unit 14 predicts the risk of landslide at the first target point. The first target point is a location on the land to be predicted where the risk of landslide is to be predicted, as shown in Figure 3. The land to be predicted is, for example, a mountain slope. Figure 3 shows one first target point. However, it is possible to predict the risk of landslide at each of multiple first target points. Details of the process in step S1 will be described later with reference to Figure 4.

[0020] In step S2, the control unit 14 predicts a sediment runoff route as shown in Figure 3 by extracting valley topography from the land to be predicted. The sediment runoff route is the route along which sediment flows. When sediment flows along the sediment runoff route, the starting point of the sediment runoff route becomes the first target point, and the ending point of the sediment runoff route becomes the second target point, which will be described later. Details of the process in step S2 will be described later with reference to Figure 12.

[0021] In step S3, the control unit 14 predicts the risk of soil and sediment reaching the second target point. The second target point is a point on the land to be predicted, as shown in Figure 3, where the risk of soil and sediment reaching the target point is to be predicted. Figure 3 shows one second target point. However, it is possible to predict the risk of soil and sediment reaching each of multiple second target points. Details of the process in step S3 will be described later with reference to Figure 17.

[0022] In step S4, the control unit 14 predicts the risk of landslides in the target land based on the risk of landslides at the first target point, the route of soil runoff, and the risk of soil reaching the second target point. Details of the process in step S4 will be described later with reference to Figure 18.

[0023] [Prediction process for the risk of landslides at the first target point] Figure 4 is a flowchart showing an example of the flow of the prediction process for the risk of landslides at the first target point. The flow shown in Figure 4 corresponds to the process of step S1 shown in Figure 2. The flow shown in Figure 4 includes the construction process of the first prediction model and the prediction process. The construction process of the first prediction model includes the processes of steps S11, S12, S13, and S14. The prediction process includes the processes of steps S21, S22, S23, and S24. The control unit 14 may execute the processes of steps S11 to S14 and the processes of steps S21 to S23 in parallel.

[0024] <Step S11> In the process of step S11, the control unit 14 acquires data on past landslides and multiple elevation information. The control unit 14 may acquire the data on past landslides and multiple elevation information by receiving them from the user via the input unit 11.

[0025] The past landslide data includes location information for multiple first points on land where landslides have occurred in the past, and information on whether or not damage occurred at these multiple first points. Land where landslides have occurred in the past is, for example, a mountain slope. The multiple first points are, for example, points at predetermined intervals on land where landslides have occurred in the past. The multiple first points include points where landslides have occurred and points where landslides have not occurred. The information on whether or not damage occurred at a first point indicates whether or not a landslide occurred at that first point. Location information is, for example, longitude and latitude information. The past landslide data may also include information on the type of landslide disaster, which will be the dependent variable described later.

[0026] Past landslide data may include rainfall data from when landslides occurred in the past. Rainfall data from when landslides occurred in the past may include, for example, hourly or 24-hour rainfall data. However, the rainfall data can be any data from when landslides occurred in the past.

[0027] Multiple elevation data points each represent the elevation of multiple points on land where landslides have occurred in the past. The elevation data is associated with the location information of the point whose elevation is indicated by that data. In addition to elevation data, the control unit 14 may acquire any other information related to past landslides. This information may include geological information associated with location information and vegetation information associated with location information. Location information may include, for example, longitude and latitude information.

[0028] Elevation information may be obtained by any means or method. For example, elevation information may be obtained from publicly available map information such as mesh information. Elevation information obtained as map information is associated with location information. As another example, elevation information may be measured values ​​obtained by surveying. In this case, location information can be associated with elevation information by obtaining location information of the place where the elevation information was measured while surveying the elevation information. As yet another example, elevation information may be obtained by analyzing images taken by a camera or depth images obtained by a ToF (Time of Flight) camera. In this case, location information can be obtained along with elevation information by analyzing the images or depth images. As yet another example, elevation information may be obtained from point cloud data obtained by laser measurement. In this case, location information can be obtained along with elevation information by analyzing the point cloud data. As yet another example, elevation information may be obtained from satellite information. Location information can be obtained along with elevation information from satellite information.

[0029] Elevation information may be obtained as mesh information. The mesh size may be arbitrary. For example, elevation information as shown in Figure 5 may be obtained as mesh information. In Figure 5, elevation information is associated with the position information of the center of a rectangular mesh and the position information of each vertex of the rectangular mesh. Specifically, in Figure 5, elevation information is E ０ Elevation E １ , ..., Elevation E ８ These are the elevations E ０, elevation E １ , …, elevation E ８ For elevation E, positions (X ０ , Y ０ ), positions (X １ , Y １ ), …, positions (X ８ , Y ８ ) are associated. Also, for these elevations E ０ , elevation E １ , …, elevation E ８ , positions (x ０１ y ０１ , x ０２ y ０２ , x ０３ y ０３ , x ０４ y ０４ ), …, positions (x ８１ y ８１ , x ８２ y ８２ , x ８３ y ８３ , x ８４ y ８４ ) are associated.

[0030] In the process of step S11, the control unit 14 may acquire data of a plurality of past landslides. Some or all of this data of a plurality of past landslides may be, respectively, data of landslides that occurred in different lands, or data of landslides that occurred at different times on the same land. When the control unit 14 acquires data of past landslides that occurred in different lands, the control unit 14 acquires a plurality of elevation information for each of the different lands.

[0031] <Step S12> In the process of step S12, the control unit 14 compares the location information of the first point with the location information associated with the elevation information and obtains the elevation information of a predetermined area including the first point. The predetermined area may be an area including the area of ​​the first point and the surrounding area of ​​the first point. For example, the control unit 14 uses a GIS (Geographic Information System) system to compare the location information of the first point with the location information associated with the elevation information and obtains the elevation information of a predetermined area including the first point. For each of the multiple first points, the control unit 14 obtains the elevation information of the predetermined area including that first point. Here, the control unit 14 may arbitrarily set the predetermined area including the first point. Examples 1 to 3 of setting the predetermined area including the first point will be described below with reference to Figures 6, 7 and 8.

[0032] [Setting Example 1] As shown in Figure 6, the control unit 14 may set a mesh containing the first point and meshes adjacent to the mesh containing the first point in a predetermined area. The meshes adjacent to the mesh containing the first point may be meshes adjacent to the mesh containing the first point in the vertical, horizontal, and diagonal directions.

[0033] [Setting Example 2] The control unit 14 may set a predetermined area to be within a predetermined distance from the first point, as shown in Figure 7. The predetermined distance may be set according to the scale of the landslide predicted by the prediction device 10. The control unit 14 may also set the mesh included within a predetermined distance from the first point to be the predetermined area.

[0034] [Setting Example 3] The control unit 14 may set a predetermined area to be a range determined from the first point according to a certain policy, as shown in Figure 8. The control unit 14 may set a predetermined area to be a mesh included in the range determined from the first point according to a certain policy. The policy may be set according to the scale of the landslide predicted by the prediction device 10.

[0035] <Step S13> In the process of step S13, the control unit 14 generates first learning terrain data that shows the topography of a predetermined area including the first point, based on the elevation information of that predetermined area. The control unit 14 generates first learning terrain data for each of the multiple first points. The control unit 14 associates the generated first learning terrain data with the first point from which the data was used to generate the first learning terrain data. The first learning terrain data may be data that shows topographic features that could be a cause of landslides. An example of first learning terrain data is described below. However, the first learning terrain data may be any data. Also, all or any combination of the data described below may be used as first learning terrain data.

[0036] The first training terrain data may include at least one of the maximum elevation and minimum elevation in a predetermined region. The control unit 14 obtains the maximum elevation and minimum elevation in a predetermined region from the elevation information of the predetermined region. Hereinafter, the maximum elevation in a predetermined region is referred to as "Maximum Elevation E ｍａｘ It is also stated that the minimum elevation in a given area is "Minimum elevation E ｍｉｎ It is also stated that the elevation of the first location is "Elevation E Ｔ It is also written as ". Elevation E Ｔ The maximum elevation E in a predetermined region ｍａｘ If (E Ｔ = E ｍａｘ ), the control unit 14 controls the altitude E Ｔ Maximum elevation E ｍａｘ It may be associated with the first point. Also, elevation E Ｔ The minimum elevation E in a predetermined region ｍｉｎ If (E Ｔ = E ｍｉｎ ), the control unit 14 controls the altitude E Ｔ Minimum elevation E ｍｉｎ It may be associated with the first point.

[0037] The first training terrain data may include at least one of the maximum elevation difference and the minimum elevation difference in a predetermined region. For example, the control unit 14 may include the maximum elevation E ｍａｘ From minimum elevation E ｍｉｎBy subtracting this, the maximum value of the elevation difference in the specified area is calculated.

[0038] The terrain data for the first training is the maximum value a of the slope angle from the first point in a predetermined region. ｍａｘ , minimum value a ｍｉｎ and the difference in inclination angle (a ｍａｘ -a ｍｉｎ ) may include at least one of the following. For example, the control unit 14 may include the maximum elevation E ｍａｘ and the elevation E of point 1 Ｔ Based on this, the maximum value of the slope angle a ｍａｘ The control unit 14 calculates the elevation E of the first point. Ｔ and minimum elevation E ｍｉｎ Based on this, the minimum value of the inclination angle a ｍｉｎ The control unit 14 calculates the maximum value a. ｍａｘ From the minimum value a ｍｉｎ By subtracting this, the difference in inclination angle (a ｍａｘ -a ｍｉｎ Calculate the result.

[0039] If the terrain data for the first training includes elevation information for n points (where n is an integer satisfying 1 ≤ n) within a predetermined region, then the slope angle a between the first point and each of the n points is... １ ~a ｎ The mean and standard deviation of the data may be included.

[0040] The first training terrain data may include at least one of the following: the distance from a valley to a first point, the distance from a river to a first point, the distance from a forest to a first point, the distance from a breakpoint to a first point, and the distance from a breakline to a first point. The valleys, rivers, forests, breakpoints, and breaklines represent terrain features. A breakpoint is a point where the slope becomes steep. A breakline is a line connecting multiple breakpoints. The control unit 14 may detect valleys, rivers, forests, breakpoints, and breaklines based on elevation information of a predetermined area.

[0041] The first training terrain data may include at least one of the following: a point in a predetermined region with a high standard deviation of plane curvature, and the distance from that point to the first point. The control unit 14 may calculate the standard deviation of plane curvature in the predetermined region based on the elevation information of the surrounding region.

[0042] The first training terrain data may include data on the curvature of a predetermined area, slope classification, and water system. The control unit 14 may calculate the curvature of the predetermined area based on elevation information of the surrounding area. Slope classification may be pre-set based on features such as the slope angle. The control unit 14 may classify the slopes of the predetermined area into one of a plurality of pre-set slope classifications based on the elevation information of the predetermined area. The control unit 14 may detect water systems from the predetermined area based on the elevation information of the predetermined area.

[0043] <Step S14> In step S14, the control unit 14 constructs a first prediction model based on multiple first training terrain data and information on whether or not damage occurred at multiple first locations. As described above, the information on whether or not damage occurred at multiple first locations is information indicating whether or not a landslide occurred at each of the multiple first locations. The constructed first prediction model outputs the risk of landslides when terrain data is input. The risk of landslides is, for example, the probability that a landslide will occur. The control unit 14 may also construct a first prediction model based on multiple first training terrain data, information on whether or not damage occurred at multiple first locations, and multiple rainfall data. These multiple rainfall data are the rainfall data included in each of the multiple past landslide data obtained in step S11.

[0044] The control unit 14 constructs a first prediction model using a machine learning or statistical method, for example, with first training terrain data as explanatory variables and whether or not a landslide occurred at the first location as the objective variable. In addition to or instead of the first training terrain data, the control unit 14 may use the types of landslides included in the past landslide data obtained in step S11 as the objective variable. Examples of landslides that can be used as the objective variable are debris flows, landslides, and mudslides. The control unit 14 may also use three stages—debris flows, landslides, and mudslides—as objective variables. Furthermore, in addition to or instead of the first training terrain data, the control unit 14 may use at least one of the geological information and vegetation information obtained in step S11 as the objective variable. Because location information is associated with the geological information and vegetation information obtained in step S11, the control unit 14 can identify the geological information and vegetation information for a predetermined range including the first location. In other words, the control unit 14 can identify at least one of geological information and vegetation information to be used in addition to, or instead of, the topographic data for first learning, along with information on whether or not there is damage at the first location. However, in constructing the first prediction model, the control unit 14 uses the same type of target variable obtained from past landslide data. The control unit 14 may construct a first prediction model for each different topographic data set for first learning that may be a cause of landslides. In this case, multiple first prediction models are constructed.

[0045] The machine learning methods used to construct the first predictive model are, for example, supervised machine learning methods such as gradient boosting decision trees, random forests, or neural networks.

[0046] The statistical method used to construct the first prediction model is, for example, a method that uses a model equation given by a function such as a linear or sigmoid curve. In this case, the first prediction model is given in the form of a model equation. Furthermore, this model equation is constructed for each variable, including the explanatory variable and the dependent variable. For example, the control unit 14 constructs a model equation as shown in Figures 9 to 10.

[0047] In Figure 9, the control unit 14 uses the function f(E) as the first prediction model. Ｔ Construct the function f(E).Ｔ ) is the elevation E of the first point. Ｔ The variable is used, and the risk of landslides is shown as the damage rate, which is the probability of a landslide occurring. In Figure 9, the horizontal axis is elevation. The vertical axis on the left shows the number of landslides that occurred. The vertical axis on the right shows the damage rate. "None" indicates that no landslides occurred. "Yes" indicates that a landslide occurred.

[0048] In Figure 10, the control unit 14 uses the function f(a) as the first prediction model. ｍａｘ Construct the function f(a ｍａｘ ) is the maximum value of the slope angle a ｍａｘ The variable is used, and the risk of sediment-related disasters is shown as the damage rate, which is the probability of a landslide occurring. In Figure 10, the horizontal axis shows the maximum value of the slope angle. The left vertical axis shows the number of landslides that occurred. The right vertical axis shows the damage rate. "None" indicates that no landslides occurred. "Yes" indicates that a landslide occurred.

[0049] In Figure 11, the control unit 14 uses the function f(d) as the first prediction model. Ｓ Construct the function f(d). Ｓ ) is the distance d from the transition point to the first point. Ｓ The variable is used, and the risk of sediment-related disasters is shown as the damage rate, which is the probability of a landslide occurring. In Figure 11, the horizontal axis shows the distance from the point of inclination to the first point. The vertical axis on the left shows the number of landslides that occurred. The vertical axis on the right shows the damage rate. "None" indicates that no landslide occurred. "Yes" indicates that a landslide occurred.

[0050] The control unit 14 may calculate the damage rate, which is the probability of a landslide occurring, by integrating multiple functions given as a model equation. For example, the control unit 14 calculates the damage rate by integrating multiple functions as shown in equation (1). Damage rate = f(E Ｔ ) + f(a ｍａｘ ) + f(d Ｓ ) + ... (1)

[0051] <Step S21> In the process of step S21, the control unit 14 acquires the location information of at least one first target point and the elevation information of multiple points. The control unit 14 may acquire the location information of the first target point and the elevation information of multiple points by receiving them from the user via the input unit 11.

[0052] The first target point is, as described above with reference to Figure 3, a point on the land to be predicted where the risk of landslides is to be predicted. For example, if the risk of landslides is to be predicted at 10m intervals on the land to be predicted, location information for multiple first target points may be obtained. In this case, the location information for multiple first target points will be location information at 10m intervals.

[0053] Multiple elevation information points each represent the elevation of multiple points on the land to be predicted. Elevation information is associated with the location information of the point whose elevation is indicated by that elevation information. In addition to elevation information, the control unit 14 may acquire any information related to landslides. Information related to landslides may include geological information associated with location information and vegetation information associated with location information. Location information is, for example, longitude and latitude information. Elevation information may be acquired in the processing of step S11 by any means or method as described above.

[0054] Here, the position of the first target point obtained in step S21 does not have to match the position of the first point obtained in step S11, or it may match. Also, the land to be predicted may be different from or the same as land where landslides have occurred in the past. However, the resolution of the multiple elevation information obtained in step S21 must match the resolution of the multiple elevation information obtained in step S11.

[0055] <Step S22> In the process of step S22, the control unit 14 compares the position information of the first target point with the position information associated with the elevation information and obtains the elevation information of a predetermined area including the first target point. For example, the control unit 14 compares the position information of the first target point with the position information associated with the elevation information using a GIS system, in the same or similar manner as in the process of step S12, and obtains the elevation information of a predetermined area including the first target point. Here, the control unit 14 sets the predetermined area including the first target point in the same way as in the process of step S12. For example, if the control unit 14 sets the predetermined area according to setting example 1 in the process of step S12, as shown in Figure 6, it sets the mesh including the first target point and the mesh adjacent to the mesh including the first target point as the predetermined area. Also, if the control unit 14 sets the predetermined area according to setting example 2 in the process of step S12, as shown in Figure 7, it sets the range within a predetermined distance from the first target point as the predetermined area. Furthermore, if the control unit 14 sets a predetermined area according to the setting example 3 in the process of step S12, it sets the predetermined area to a range determined from the first target point according to a certain policy, as shown in Figure 8.

[0056] <Step S23> In the process of step S23, the control unit 14 generates first prediction terrain data that shows the terrain of a predetermined area based on the elevation information of the predetermined area including the first target point. The control unit 14 generates first prediction terrain data of the same type as the first learning terrain data generated in the process of step S13.

[0057] <Step S24> In step S24, the control unit 14 predicts the risk of landslide at the first target point. The control unit 14 predicts the risk of landslide at the first target point by inputting the first prediction terrain data generated in step S23 into the first prediction model generated in step S14.

[0058] If the control unit 14 has constructed a first prediction model in step S14 based on multiple first training data and multiple rainfall data, it may accept rainfall data from the user via the input unit 11. This rainfall data is, for example, rainfall data predicted for a day when the risk of landslides is to be predicted. The control unit 14 predicts the risk of landslides at the first target point by inputting the first prediction terrain data generated in step S23 and the accepted rainfall data into the first prediction model.

[0059] If the control unit 14 has constructed multiple first prediction models in the process of step S14, it may use the multiple first prediction models to predict multiple landslide risks for a single first target point.

[0060] The control unit 14 may display the risk of landslides at the first target point on the display of the output unit 12, along with a map showing the location information of the first target point.

[0061] [Prediction process for sediment runoff routes] Figure 12 is a flowchart showing an example of the flow of the sediment runoff route prediction process. The flow shown in Figure 12 corresponds to the process in step S2 shown in Figure 2.

[0062] <Step S31> In the process of step S31, the control unit 14 acquires location information and elevation information for multiple candidate locations. The control unit 14 may acquire the location information and elevation information for multiple candidate locations by receiving them from the user via the input unit 11.

[0063] Candidate locations are potential valley locations on the land being predicted. However, the candidate locations obtained in step S31 may be any location on the land being predicted.

[0064] Multiple elevation information points each represent the elevation of multiple points on the land to be predicted. Elevation information is associated with the location information of the point whose elevation is indicated by that elevation information. In addition to elevation information, the control unit 14 may acquire any information related to landslides. Information related to landslides may include geological information associated with location information and vegetation information associated with location information. Location information is, for example, longitude and latitude information. Elevation information may be acquired in the processing of step S11 by any means or method as described above.

[0065] <Step S32> In the process of step S32, the position information of the candidate point is compared with the position information associated with the elevation information, and the elevation information of a predetermined area including the candidate point is obtained. For example, the control unit 14, in the same or similar manner as in the process of step S21, compares the position information of the candidate point with the position information associated with the elevation information using a GIS system, and obtains the elevation information of a predetermined area including the candidate point. The control unit 14 may set the predetermined area by any of the setting examples 1 to 3 described above, referring to Figures 6 to 8.

[0066] <Step S33> In the process of step S33, the control unit 14 generates topographic data that shows the characteristics of valley terrain in a predetermined area, based on the elevation information of the predetermined area including the candidate points. The control unit 14 generates topographic data that shows the characteristics of valley terrain for each of the multiple candidate points. The control unit 14 associates the generated topographic data with the candidate points whose data was used to generate the topographic data. The topographic data that shows the characteristics of valley terrain may be any data as long as it shows the characteristics of valley terrain. As an example, the topographic data may include data on the slope angle, curvature, slope classification, and water system of the predetermined area. The control unit 14 may calculate the slope angle and curvature of the predetermined area based on the elevation information of the predetermined area. The slope classification may be set in advance based on the characteristics of the slope, such as the slope angle. The control unit 14 may classify the slopes of the predetermined area into one of a plurality of pre-set slope classifications based on the elevation information of the predetermined area. The control unit 14 may identify the water system of the predetermined area based on the elevation information of the predetermined area.

[0067] <Step S34> In step S34, the control unit 14 extracts valley topography from the land to be predicted based on the terrain data generated in step S33. As an example, the control unit 14 detects depressions in the terrain based on at least one of the slope angle, curvature, slope classification, and water system of the terrain data. The control unit 14 extracts valley topography by grouping the detected depressions into a single set. In extracting valley topography, the control unit 14 may group the depressions into a set of lines as shown in Figure 13, or into a set of meshes as shown in Figure 14, or into a set of polygons as shown in Figure 15. Also, as shown in Figure 16, if the control unit 14 detects depressions as point data, it may group the depressions into a single set by providing a buffer for the data of each point. For example, the control unit 14 may provide a buffer for the data of each point using the buffer function of the GIS system.

[0068] The control unit 14 uses the extracted valley topography as a route for soil runoff.

[0069] [Prediction process for the risk of sediment reaching the second target point] Figure 17 is a flowchart showing an example of the flow of the prediction process for the risk of sediment reaching the second target point. The flow shown in Figure 4 corresponds to the process of step S3 shown in Figure 2. The flow shown in Figure 17 includes the construction process of the second prediction model and the prediction process. The construction process of the second prediction model includes the processes of steps S41, S42, S43, and S44. The prediction process includes the processes of steps S51, S52, S53, and S54. The control unit 14 may execute the processes of steps S41 to S44 and the processes of steps S51 to S53 in parallel.

[0070] <Step S41> In the process of step S41, the control unit 14 acquires past sediment arrival data and multiple elevation information. The control unit 14 may acquire past sediment arrival data and multiple elevation information by receiving them from the user via the input unit 11.

[0071] Past landslide arrival data consists of location information for multiple secondary points on land where landslides have occurred in the past, and information on whether or not damage occurred at these secondary points. These secondary points are, for example, points at predetermined intervals on land where landslides have occurred in the past. These secondary points include both points where landslides occurred and points where they did not. The information on whether or not damage occurred at a secondary point indicates whether or not landslides reached that point. Location information is, for example, longitude and latitude. Past landslide arrival data may also include information on the type of landslide disaster, which will be the dependent variable described later.

[0072] Past landslide arrival data may include rainfall data from past landslide occurrences. Rainfall data from past landslide occurrences may include, for example, hourly or 24-hour rainfall data. However, the rainfall data can be any data from past landslide occurrences.

[0073] The past data on sediment arrival may be the same land data as the past landslide data obtained in step S11, or it may be different land data from the past landslide data obtained in step S11.

[0074] Each of the multiple elevation data points represents the elevation of multiple points on land where sediment intrusion has occurred in the past. The elevation data is associated with the location information of the point whose elevation is indicated by that elevation data. In addition to the elevation data, the control unit 14 may acquire any information related to past sediment intrusion. The information related to past sediment intrusion may include geological information associated with location information and vegetation information associated with location information. Location information is, for example, longitude and latitude information. The elevation data may be acquired in the processing of step S11 by any means or method as described above.

[0075] In step S41, the control unit 14 may acquire data on multiple past landslides. Some or all of this data on multiple past landslides may be data on landslides that occurred in different locations, or data on landslides that occurred in the same location at different times. When the control unit 14 acquires data on past landslides that occurred in different locations, it acquires multiple elevation information for each of the different locations.

[0076] <Step S42> In the process of step S42, the control unit 14 compares the location information of the second point with the location information associated with the elevation information and obtains the elevation information of a predetermined area including the second point. For example, the control unit 14 compares the location information of the second point with the location information associated with the elevation information using a GIS system, in the same or similar manner as in the process of step S12, and obtains the elevation information of a predetermined area including the second point. The control unit 14 may set the predetermined area according to the setting examples 1 to 3 described above, with reference to Figures 6 to 8.

[0077] <Step S43> In the process of step S43, the control unit 14 generates second learning terrain data that shows the topography of a predetermined area including the second point, based on the elevation information of that predetermined area. The control unit 14 generates second learning terrain data for each of the multiple second points. The control unit 14 associates the generated second learning terrain data with the second point whose data was used to generate the second learning terrain data. The second learning terrain data may be data that shows topographic features that can be factors that cause soil and sediment to reach the area. The control unit 14 may generate the second learning terrain data in the same or similar manner as in the process of step S13.

[0078] <Step S44> In step S44, the control unit 14 constructs a second prediction model based on multiple second learning terrain data and information on whether or not damage occurred at multiple second locations. As described above, the information on whether or not damage occurred at multiple second locations is information indicating whether or not sediment has reached each of the multiple second locations. The constructed second prediction model outputs the risk of sediment arrival when terrain data is input. The risk of sediment arrival is, for example, the probability that sediment will arrive. The control unit 14 may also construct a second prediction model based on multiple second learning terrain data, information on whether or not damage occurred at multiple second locations, and multiple rainfall data. These multiple rainfall data are the rainfall data included in each of the multiple past sediment arrival data obtained in step S41.

[0079] The control unit 14 constructs a second prediction model in the same or similar manner as the process in step S14. The control unit 14 may construct a second prediction model for each different second training terrain data that may be factors in the arrival of sediment. In this case, multiple second prediction models are constructed. In the same or similar manner as the process in step S14, in addition to or instead of the second training terrain data, the control unit 14 may use the type of sediment disaster included in the past sediment arrival data obtained in the process of step S41 as the target variable. Alternatively, in addition to or instead of the second training terrain data, the control unit 14 may use at least one of the geological information and vegetation information obtained in the process of step S41 as the target variable. Because location information is associated with the geological information and vegetation information obtained in the process of step S41, the control unit 14 can identify the geological information and vegetation information for a predetermined range including the second point.

[0080] <Step S51> In the process of step S51, the control unit 14 acquires the location information of at least one second target point and the elevation information of multiple points. The control unit 14 may acquire the location information of the second target point and the elevation information of multiple points by receiving them from the user via the input unit 11.

[0081] The second target point, as described above with reference to Figure 3, is a point on the land subject to prediction where we want to predict the risk of soil and landslide arrival. For example, if we want to predict the risk of soil and landslide arrival at 10m intervals on the land subject to prediction, location information for multiple second target points may be obtained. In this case, the location information for multiple second target points will be location information at 10m intervals.

[0082] Each of the multiple elevation data points represents the elevation of a specific point on the land being predicted. The elevation data is associated with the location information of the point whose elevation is indicated by that elevation data. In addition to the elevation data, the control unit 14 may acquire any information related to sediment arrival. The information related to sediment arrival may include geological information associated with location information and vegetation information associated with location information. Location information may be, for example, longitude and latitude information. The elevation data may be acquired in the processing of step S11 by any means or method as described above.

[0083] Here, the position of the second target point obtained in step S51 does not have to match the position of the second point obtained in step S41, or it may match. Also, the land to be predicted may be different from or the same as land where sediment intrusion has occurred in the past. However, the resolution of the multiple elevation information obtained in step S51 must match the resolution of the multiple elevation information obtained in step S41.

[0084] <Step S52> In the process of step S52, the control unit 14 compares the position information of the second target point with the position information associated with the elevation information and obtains the elevation information of a predetermined area including the second target point. For example, the control unit 14 compares the position information of the second target point with the position information associated with the elevation information using a GIS system, in the same or similar manner as in the process of step S12, and obtains the elevation information of a predetermined area including the second target point. Here, the control unit 14 sets the predetermined area including the second target point in the same way as in the process of step S42. For example, if the control unit 14 sets the predetermined area according to setting example 1 in the process of step S42, as shown in Figure 6, it sets the mesh including the second target point and the mesh adjacent to the mesh including the second target point as the predetermined area. Also, if the control unit 14 sets the predetermined area according to setting example 2 in the process of step S42, as shown in Figure 7, it sets the range within a predetermined distance from the second target point as the predetermined area. Furthermore, if the control unit 14 sets a predetermined area according to the setting example 3 in the process of step S42, it sets the range determined from the second target point in accordance with a certain policy as the predetermined area, as shown in Figure 8.

[0085] <Step S53> In the process of step S53, the control unit 14 generates second prediction terrain data that shows the terrain of a predetermined area based on the elevation information of the predetermined area including the second target point. The control unit 14 generates second prediction terrain data of the same type as the second learning terrain data generated in the process of step S43.

[0086] <Step S54> In step S54, the control unit 14 predicts the risk of sediment reaching the second target point. The control unit 14 predicts the risk of sediment reaching the second target point by inputting the second prediction terrain data generated in step S53 into the second prediction model generated in step S44.

[0087] If the control unit 14 has constructed a second prediction model in step S44 based on multiple second training data and multiple rainfall data, it may accept rainfall data from the user via the input unit 11. This rainfall data is, for example, rainfall data predicted for a day when the risk of landslide arrival is to be predicted. The control unit 14 predicts the risk of landslide arrival at the second target point by inputting the second prediction terrain data generated in step S53 and the accepted rainfall data into the second prediction model.

[0088] If the control unit 14 constructs multiple second prediction models in the process of step S44, it may use the multiple second prediction models to predict the risk of multiple soil erosion reaching a single second target point.

[0089] The control unit 14 may display on the output unit 12's display the risk of soil and sand reaching the second target point, along with a map showing the location information of the second target point.

[0090] [Prediction Process for Sediment-Related Disaster Risk] Figure 18 is a flowchart showing an example of the flow of the prediction process for sediment-related disaster risk. The flow shown in Figure 18 corresponds to the process in step S4 shown in Figure 2. The sediment-related disaster risk calculated by the following process is the risk calculated based on the risk of landslide and the risk of sediment reaching the area.

[0091] <Step S61> In the process of step S61, the control unit 14 identifies the calculation unit for calculating the risk of sediment-related disasters.

[0092] First, the control unit 14 detects a first target point connected to the upper part of the sediment runoff route and a second target point connected to the lower part of the sediment runoff route. The control unit 14 assumes that the first target point is connected to the upper part of the sediment runoff route if the position of the upper part of the sediment runoff route coincides with the position of the first target point. However, even if the position of the upper part of the sediment runoff route and the position of the first target point do not coincide, the control unit 14 may consider the upper part of the sediment runoff route and the first target point to be connected if the first target point is located within a first distance from the upper part of the sediment runoff route. Furthermore, the control unit 14 assumes that the second target point is connected to the lower part of the sediment runoff route if the position of the lower part of the sediment runoff route coincides with the position of the second target point. However, even if the position of the lower part of the sediment outflow route and the position of the second target point do not coincide, the control unit 14 may consider the lower part of the sediment outflow route and the second target point to be connected if the second target point is located within the second distance from the lower part of the sediment outflow route. The first distance and the second distance may be set based on data from past sediment disasters.

[0093] The control unit 14 uses a pattern of one first target point and one second target point connected by one sediment runoff route as the basic unit for calculating sediment disaster risk. However, the control unit 14 may use multiple first target points and one second target point connected by one sediment runoff route as the unit for calculating sediment disaster risk. Alternatively, the control unit 14 may use one first target point and multiple second target points connected by one sediment runoff route as the unit for calculating sediment disaster risk. Alternatively, the control unit 14 may use multiple first target points and multiple second target points connected by one sediment runoff route as the unit for calculating sediment disaster risk.

[0094] The control unit 14 identifies, for example, calculation unit 1, calculation unit 2, calculation unit 3, and calculation unit 4 as shown in Figure 19.

[0095] Calculation unit 1 is a pattern of one first target point and one second target point connected by one sediment runoff route.

[0096] The calculation unit 2 is a pattern of one first target point connected by one soil runoff route and N second target points (where N is an integer satisfying 2 ≤ N).

[0097] Calculation unit 3 is a pattern of M first target points (where M is an integer satisfying 2 ≤ M) and one second target point connected by a single sediment runoff route.

[0098] The calculation unit 4 is a pattern of P first target points (where P is an integer satisfying 2 ≤ P) and Q second target points (where Q is an integer satisfying 2 ≤ Q) connected by a single sediment runoff route.

[0099] <Step S62> In step S62, the control unit 14 calculates the risk of landslides based on the calculation units identified in step S61. As described below, the control unit 14 may calculate the risk of landslides using calculation methods corresponding to calculation unit 1, calculation unit 2, calculation unit 3, and calculation unit 4.

[0100] Here, if the control unit 14 predicts multiple landslide risks for one first target point in the process of step S24, it may use one landslide risk for one first target point in the following calculation process based on those multiple landslide risks. For example, if the control unit 14 predicts multiple landslide risks for one first target point in the process of step S24, it may use the maximum value, average value, or median value of the multiple landslide risks.

[0101] Furthermore, if the control unit 14 predicts multiple risks of soil and debris reaching a single second target point in the processing of step S54, it may use one soil and debris reaching risk for a single second target point in the following calculation processing, based on the multiple soil and debris reaching risks. For example, if the control unit 14 predicts multiple risks of soil and debris reaching a single second target point in the processing of step S54, it may use the maximum value, average value, or median value among the multiple soil and debris reaching risks.

[0102] [Calculation Unit 1] As described above, the calculation unit 1 is a pattern of one first target point and one second target point connected by one sediment outflow route. Hereinafter, the risk of sediment collapse at the first target point is denoted as "sediment collapse risk R １ ". Also, the risk of sediment reaching the second target point is denoted as "sediment reaching risk R ２ ".

[0103] In the calculation unit 1, the control unit 14 calculates the risk R of sediment disaster according to Equation (2). R = aR １ + bR ２ (2) In Equation (2), the coefficient a is the weight coefficient for the sediment collapse risk R １ . The coefficient b is the weight coefficient for the sediment reaching risk R ２ . The coefficients a and b may be set based on the type of sediment disaster. However, the coefficients a and b may also be set to 1 (a = b = 1).

[0104] As another example, the control unit 14 may calculate the risk R of sediment disaster as the maximum value of the sediment collapse risk R １ and the sediment reaching risk R ２ (R = MAX(R １ , R ２ )). As yet another example, the control unit 14 may calculate the risk R of sediment disaster as the average value of the sediment collapse risk R １ and the sediment reaching risk R ２ (R = AVERAGE(R １ , R ２ )).

[0105] [Calculation Unit 2] As described above, the calculation unit 2 is a pattern of one first target point and N second target points connected by one sediment outflow route. Hereinafter, among the sediment reaching risks R ２ to the N second target points, the sediment reaching risk R ２ to the i-th (i is an integer satisfying 1 ≤ i ≤ N) second target point is denoted as "sediment reaching risk R ２ｉ ".

[0106] In the calculation unit 2, there is one first target point where the risk of landslide is predicted, while there are multiple second target points where the risk of sediment arrival is predicted. Therefore, when a landslide occurs at one first target point, the control unit 14 calculates the sediment arrival risk R that sediment reaches the i-th second target point ３ｉ as the risk R of sediment disaster. In this case, the control unit 14 calculates, as the risk R of sediment disaster, the sediment arrival risk R ３１ ~R ３Ｎ . With such a configuration, when a landslide occurs at one first target point, the risk that sediment reaches each of the N second target points can be evaluated.

[0107] The control unit 14 calculates the sediment arrival risk R that sediment reaches the i-th second target point when a landslide occurs at one first target point ３ｉ by the formula (3). R ３ｉ = cR １ + dR ２ｉ (3) In the formula (3), the coefficient c is a weighting coefficient for the landslide risk R １ . The coefficient d is a weighting coefficient for the sediment arrival risk R ２ｉ . The coefficients c and d may be set based on the type of sediment disaster. However, the coefficients c and d may be set to 1 (c = d = 1).

[0108] As another example, the control unit 14 may calculate the risk R of sediment disaster as the sum of the landslide risk R １ and the sediment arrival risks R ２１ ~R ２Ｎ (R = SUM(R １ , R ２１ ~R ２Ｎ )). As yet another example, the control unit 14 may calculate the risk R of sediment disaster as the sum of the landslide risk R １ and the maximum value of the sediment arrival risks R ２１ ~R ２Ｎ (R = R １ + MAX(R ２１ ~R ２Ｎ )). As yet another example, the control unit 14 may calculate the risk R of sediment disaster as the landslide risk R １And, the risk of sediment intrusion R ２１ ~R ２Ｎ It may also be calculated as the sum of the average values ​​(R = R １ +AVERAGE(R ２１ ~R ２Ｎ )).

[0109] [Calculation Unit 3] As described above, Calculation Unit 3 is a pattern of M first target points and one second target point connected by one sediment runoff route. Below, the sediment collapse risk R of the M first target points is calculated. １ Of these, the landslide risk R of the j-th (where j is an integer satisfying 1 ≤ j ≤ M) first target point １ "Landslide risk R １ｊ It is written as follows:

[0110] In calculation unit 3, there are multiple first target points where the risk of landslides is predicted, while there is only one second target point where the risk of soil arrival is predicted. Therefore, the control unit 14 calculates the soil arrival risk R, which is the risk of soil reaching one second target point if a landslide occurs at the j-th first target point. ４ｊ The risk R of a sediment-related disaster may be calculated as the risk R of a sediment-related disaster. In this case, the control unit 14 calculates the risk R of sediment-related disaster as the risk R of sediment-related disaster. ４１ ~R ４Ｎ This calculates the risk of soil reaching one second target point if a landslide occurs at each of the M first target points.

[0111] The control unit 14 calculates the risk R of soil reaching one second target point if a landslide occurs at the j-th first target point. ４ｊ This is calculated using equation (4). R ４ｊ = eR １ｊ +fR ２ (4) In equation (4), the coefficient e is the risk of landslides R １ｊ This is a weighting coefficient for R. The coefficient f is the risk of sediment intrusion R. ２ These are weighting coefficients. Coefficients e and f may be set based on the type of landslide. However, coefficients e and f may also be set to 1 (e = f = 1).

[0112] As another example, the control unit 14 determines the risk R of landslides and the risk R of soil collapses. １１ ~R １Ｍ and risk of sediment intrusion R ２ It can also be calculated as the sum of (R = SUM(R １１ ~R １Ｍ , R ２ )). As another example, the control unit 14 determines the risk R of sediment-related disasters, and the risk R of soil collapse. １１ ~R １Ｍ The maximum value among them, and the risk of sediment intrusion R ２ It can also be calculated as the sum of (R = MAX(R) １１ ~R １Ｍ ) + R ２ ). As another example, the control unit 14 determines the risk R of sediment-related disasters, and the risk R of soil collapse. １１ ~R １Ｍ The average value and the risk of sediment intrusion R ２ It may also be calculated as the sum of (R = AVERAGE(R) １１ ~R １Ｍ ) + R ２ ).

[0113] [Calculation Unit 4] As described above, Calculation Unit 4 is a pattern of P first target points and Q second target points connected by a single sediment runoff route. Below, the sediment collapse risk R of the P first target points is calculated. １ Of these, the landslide risk R of the first target point, which is the kth (where k is an integer satisfying 1 ≤ k ≤ P) １ "Landslide risk R １ｋ It is stated that the risk of soil and debris reaching Q second target points is R. ２ Of these, the landslide risk R of the m-th (where m is an integer satisfying 1 ≤ m ≤ Q) second target point. ２ "Landslide risk R ２ｍ It is written as follows:

[0114] In calculation unit 4, there are multiple first target points where the risk of landslides is predicted, and multiple second target points where the risk of soil arrival is predicted. Therefore, the control unit 14 calculates the soil arrival risk R of soil reaching the mth second target point if a landslide occurs at the kth first target point. ５ｋｍThe risk R of a sediment-related disaster may be calculated as the risk R of a sediment-related disaster. In this case, the control unit 14 calculates the risk R of sediment-related disaster as the risk R of sediment-related disaster. ５ｋｍ In other words, the risk of sediment intrusion R ５１１ ~R ５ＰＱ This calculates the risk of soil reaching the m-th second target point if a landslide occurs at the k-th first target point.

[0115] The control unit 14 determines the risk R of soil reaching the m-th second target point if a landslide occurs at the k-th first target point. ５ｋｍ This is calculated using equation (5). R ５ｋｍ = gR １ｋ +hR ２ｍ (5) In equation (5), the coefficient g is the risk of landslides R １ｋ This is a weighting coefficient for R. The coefficient h is the risk of sediment intrusion R. ２ｍ These are weighting coefficients. Coefficients g and h may be set based on the type of sediment-related disaster. However, coefficients g and h may also be set to 1 (g = h = 1).

[0116] As another example, the control unit 14 determines the risk R of landslides and the risk R of soil collapses. １１ ~R １Ｐ and risk of sediment intrusion R ２１ ~R ２Ｑ It can also be calculated as the sum of (R = SUM(R １１ ~R １Ｐ , R ２１ ~R ２Ｑ )). As another example, the control unit 14 determines the risk R of sediment-related disasters, and the risk R of soil collapse. １１ ~R １Ｐ The maximum value among them, and the risk of sediment intrusion R ２１ ~R ２Ｑ It can also be calculated as the sum of the maximum value among them (R = MAX(R １１ ~R １Ｐ ) + MAX (R ２１ ~R ２Ｑ )). As another example, the control unit 14 determines the risk R of sediment-related disasters, and the risk R of soil collapse. １１ ~R １Ｐ The average value and the risk of sediment intrusion R ２１ ~R２Ｑ It may also be calculated as the sum of (R = AVERAGE(R) １１ ~R １Ｐ )+AVERAGE(R ２１ ~R ２Ｑ )).

[0117] In step S62, the control unit 14 may display the calculated risk of landslides on the display of the output unit 12.

[0118] <Step S63> In the process of step S63, the control unit 14 processes the standardized landslide risk R １ and risk of sediment intrusion R ２ The safety level of the sediment runoff route extracted as a valley topography is calculated. As described below, the control unit 14 calculates the safety level of the sediment runoff route according to the calculation methods corresponding to calculation unit 1, calculation unit 2, calculation unit 3, and calculation unit 4 as shown in Figure 19.

[0119] [Calculation Unit 1] The control unit 14 calculates the landslide risk R of the first target point. １ The risk R of soil and debris reaching the second target point. ２ The values ​​are standardized so that each falls within the range of 0 to 1. The landslide risk R after standardization. １ "Landslide risk r １ It is stated as follows: Standardized soil and sediment arrival risk R ２ "Risk of sediment arrival r" ２ It is written as follows:

[0120] The control unit 14 calculates the safety level S of the soil runoff route using equation (6). S = (1 - r １ ) × (1 - r ２ ) (6)

[0121] [Calculation Unit 2] The control unit 14 calculates the landslide risk R of the first target point. １ And the risk R of soil and debris reaching each of the N second target points. ２１ ~R ２Ｎ The values ​​are standardized so that each falls within the range of 0 to 1. The landslide risk R after standardization. １ As mentioned above, "landslide risk r １It is stated as follows: The normalized soil and sediment reach risk R to the i-th (where i is an integer satisfying 1 ≤ i ≤ N) second target point out of N second target points. ２ｉ "Risk of sediment arrival r" ２ｉ It is written as follows:

[0122] The control unit 14 calculates the safety level S of the soil runoff route using equation (7). S = (1 - r １ ) × [1 - { (1 - r ２１ ) × (1 - r ２２ ) × ... × (1 - r ２Ｎ )}] (7)

[0123] [Calculation Unit 3] The control unit 14 calculates the landslide risk R of each of the M first target points. １１ ~R １Ｍ The risk R of soil and debris reaching the second target point. ２ The values ​​are normalized so that each falls within the range of 0 to 1. The normalized landslide risk R of the j-th first target point (where j is an integer satisfying 1 ≤ j ≤ M) out of M first target points. １ｊ "Landslide risk r １ｊ It is stated as follows: Standardized soil and sediment arrival risk R ２ As mentioned above, "the risk of sediment intrusion r ２ It is written as follows:

[0124] The control unit 14 calculates the safety level S of the soil runoff route using equation (8). S = [1 - {(1 - r １１ ) × (1 - r １２ ) × ... × (1 - r １Ｍ )}] × (1-r ２ ) (8)

[0125] [Calculation Unit 4] The control unit 14 calculates the landslide risk R of each of the P first target points. １１ ~R １Ｐ And the risk R of soil and debris reaching each of the Q second target points. ２１ ~R ２Ｑ The values ​​are normalized so that each falls within the range of 0 to 1. The normalized landslide risk R of the k-th (where k is an integer satisfying 1 ≤ k ≤ P) first target point out of P first target points. １ｋ "Landslide risk r １ｋThe normalized soil and sediment reach risk R of the m-th (where m is an integer satisfying 1 ≤ m ≤ Q) second target point out of Q second target points. ２ｍ "Risk of sediment arrival r" ２ｍ It is written as follows:

[0126] The control unit 14 calculates the safety level S of the soil runoff route using equation (9). S = [1 - {(1 - r １１ ) × (1 - r １２ ) × ... × (1 - r １Ｐ )}] ×[1-{(1-r ２１ ) × (1 - r ２２ ) × ... × (1 - r ２Ｑ )}] (9)

[0127] In step S63, the control unit 14 may display the calculated safety level of the soil runoff route on the display of the output unit 12.

[0128] Here, the information used in the flows shown in Figure 4, Figure 12, and Figure 17 may be independent or the same. For example, the means or direction of acquiring elevation information may differ between the flows shown in Figure 4, Figure 12, and Figure 17. Also, the resolution of the various types of information may differ or be the same between the flows shown in Figure 4, Figure 12, and Figure 17. For example, in the flows shown in Figure 4 and Figure 17, measured elevation information with a mesh size of 10m may be used, while in the flow shown in Figure 12, elevation information with a mesh size of 1m obtained from satellite information may be used. Even with such a configuration, the risk of landslides on the target land can be predicted using the flow shown in Figure 18.

[0129] In the prediction device 10 according to this embodiment, the control unit 14 constructs a first prediction model using a plurality of first learning terrain data and information on the presence or absence of damage at a plurality of first locations. Furthermore, the control unit 14 predicts the risk of landslides at a first target point by inputting the first prediction terrain data into the first prediction model. The first learning terrain data represents the terrain of a predetermined area including a first location on land where landslides have occurred in the past. The first prediction terrain data represents the terrain of a predetermined area including a first target point on the land to be predicted. By using the first learning terrain data and the first prediction terrain data that represent the terrain in this way, the risk of landslides can be predicted while taking into account the terrain of each individual piece of land. As a result, the risk of landslides can be predicted with high accuracy. Therefore, according to this embodiment, the risk of landslides occurring in mountainous areas during heavy rainfall can be predicted with high accuracy.

[0130] Furthermore, in this embodiment, the control unit 14 may predict the risk of landslides at each of the multiple first target points. By predicting the risk of landslides at each of the multiple first target points, it is possible to identify the first target points with a high risk of landslides among the multiple first target points. With this configuration, it is possible to identify areas with a high risk of landslides in the land being predicted. In addition, based on the level of landslide risk, it is possible to prioritize which of the multiple first target points should have landslide countermeasures. This makes it possible to take countermeasures against landslides in advance.

[0131] Furthermore, in this embodiment, the control unit 14 may construct a second prediction model using a plurality of second learning terrain data and information on the presence or absence of damage at a plurality of second locations. The control unit 14 may predict the risk of sediment reaching the second target point by inputting the second prediction terrain data into the second prediction model. The second learning terrain data represents the terrain of a predetermined area including a second location on land where sediment has occurred in the past. The second prediction terrain data represents the terrain of a predetermined area including the second target point on the land to be predicted. By using the second learning terrain data and the second prediction terrain data that represent the terrain in this way, the risk of sediment reaching the land can be predicted while taking into account the terrain of each individual piece of land.

[0132] Furthermore, in this embodiment, the control unit 14 may predict the risk of soil and debris reaching each of the multiple second target points. By predicting the risk of soil and debris reaching each of the multiple second target points, it is possible to identify the second target points with a high risk of soil and debris reaching among the multiple second target points. With this configuration, it is possible to identify areas in the land being predicted that have a high risk of soil and debris reaching. In addition, based on the level of risk of soil and debris reaching, it is possible to prioritize the measures that should be taken against soil and debris reaching at the multiple second target points. This makes it possible to take measures against soil and debris reaching in advance.

[0133] Furthermore, in this embodiment, the control unit 14 may predict the risk of landslides in the target land based on the risk of landslides at the first target point, the route of soil runoff, and the risk of soil reaching the second target point. By combining these, the risk of landslides can be predicted with greater accuracy.

[0134] This disclosure is not limited to the embodiments described above. For example, two or more blocks described in the block diagram may be combined, or one block may be divided. Instead of executing two or more steps described in the flowchart in chronological order as described, they may be executed in parallel or in a different order, depending on the processing capacity of the device performing each step or as necessary. Other modifications are possible without departing from the spirit of this disclosure.

[0135] For example, in the processes of step S13 and step S23 shown in Figure 4, the control unit 14 is described as generating first learning terrain data and first prediction terrain data based on elevation information. Also, in the process of step S33 shown in Figure 12, the control unit 14 is described as generating terrain data that shows the characteristics of valley terrain based on elevation information. Also, in the processes of steps S43 and S53 shown in Figure 17, the control unit 14 is described as generating second learning terrain data and second prediction terrain data based on elevation information. However, the control unit 14 may generate the first learning terrain data, the first prediction terrain data, the terrain data showing the characteristics of valley terrain, the second learning terrain data, and the second prediction terrain data based on arbitrary information other than elevation information. For example, the control unit 14 may generate this terrain data from publicly available data or image analysis, etc.

[0136] The prediction device described herein can also be implemented using a computer and a program, and the program can be recorded on a recording medium or provided via a network.

[0137] For example, an embodiment is also possible in which a general-purpose computer functions as the prediction device 10 according to the above embodiment. Specifically, a program describing the processing content that realizes each function of the prediction device 10 according to the above embodiment is stored in the memory of the general-purpose computer, and the processor reads and executes the program. Therefore, this disclosure can also be realized as a program that can be executed by a processor, or as a non-temporary computer-readable medium that stores the program.

[0138] 10: Prediction device, 11: Input unit, 12: Output unit, 13: Storage unit, 14: Control unit

Claims

1. A prediction device comprising a control unit that constructs a first prediction model using multiple first learning terrain data and information on the presence or absence of damage at multiple first points on land where landslides have occurred in the past, wherein the multiple first learning terrain data are data that each represents the terrain of a plurality of predetermined areas including each of the plurality of first points, the information on the presence or absence of damage at the plurality of first points is information that each represents whether or not a landslide has occurred at each of the plurality of first points, the first prediction model outputs the risk of landslides when terrain data is input, and the control unit predicts the risk of landslides at the first target point by inputting first prediction terrain data, which represents the terrain of a predetermined area including a first target point on the land to be predicted, into the first prediction model.

2. The control unit constructs a second prediction model using a plurality of second learning terrain data and information on whether or not damage occurred at a plurality of second points on land where sediment has occurred in the past, the plurality of second learning terrain data each represent the terrain of a plurality of predetermined areas including each of the plurality of second points, the information on whether or not damage occurred at the plurality of second points each represent information on whether or not sediment has reached each of the plurality of second points, the second prediction model outputs the risk of sediment arrival when terrain data is input, and the control unit predicts the risk of sediment arrival at the second target point by inputting second prediction terrain data showing the terrain of a predetermined area including a second target point on the land to be predicted into the second prediction model, the prediction device according to claim 1.

3. The prediction device according to claim 2, wherein the control unit predicts the route of soil runoff by extracting valley topography from the land to be predicted.

4. The prediction device according to claim 3, wherein the control unit predicts the risk of a landslide in the land to be predicted based on the risk of a landslide at the first target point, the route of the soil runoff, and the risk of the soil reaching the second target point.

Images