Prediction device

Abstract

A prediction device (100) according to the present disclosure is provided with: a prediction unit (130) that predicts the probability of each of a plurality of structures in a designated region being damaged by inputting information on the position of each of the plurality of structures and surrounding environment information indicating the surrounding environment of each of the plurality of structures into a prediction model that predicts the probability of each of a plurality of structures being damaged by a disaster caused by rainfall, the prediction model being constructed by learning of training data that associates the position of each of a plurality of structures, whether the structure is damaged at the time of occurrence of a disaster caused by rainfall, and the surrounding environment of the structure; and a proposal unit (140) that outputs a countermeasure against a disaster caused by rainfall in the designated region on the basis of the probability of each of the plurality of structures being damaged.

Description

Prediction Device

[0001] The present disclosure relates to a prediction device.

[0002] As a system for predicting damage caused by wind and floods to ground-based structures such as utility poles, a system (RAMP-T: Risk Assessment and Management system for Power lifeline for Typhoon) has been developed that calculates wind speed for typhoon wind damage, predicts damage to structures based on the calculated wind speed, and outputs the results (see Non-Patent Document 1). Also, a system (RAMP-Er: Risk Assessment and Management system for Power lifeline for Earthquake) has been developed that predicts damage caused by earthquakes, and the prediction results of these systems can be displayed together on a map (see Non-Patent Document 2).

[0003] On the other hand, there is no system that can predict damage from rainfall-induced landslides at the structural level, such as utility poles. The Japan Meteorological Agency provides a risk map called "Dossa Kikikuru." Landslide warning information is issued based on the soil moisture content predicted using a tank model that uses a perforated tank to model the flow of rain through the soil, and the accumulated rainfall over a 60-minute period. Here, the parameters of the tank model are standardized nationwide, and regional topographical information is not taken into account (see Non-Patent Document 3).

[0004] The National Institute for Land and Infrastructure Management has created a proposed probability map that calculates the probability of landslides based on topography and predisposing factors. This probability map calculates the probability of landslides occurring in three stages based on three types of maps: landslide warning areas, deep-seated collapse estimated frequency maps, and landslide topography distribution maps (see Non-Patent Document 4).

[0005] Regarding road collapses due to earthquakes, methods are being considered to calculate the slope collapse rate and the probability that road networks will be interrupted and settlements will be isolated (see Non-Patent Documents 5 and 6), as well as a method of monitoring and acquiring data using sensors (Non-Patent Document 7).

[0006] Sugimoto et al., "Introduction of a Typhoon Damage Estimation System and Weather Simulation," IEEJ Journal Vol. 138 No. 3 pp. 141-144, 2018. T. Tadokoro "Automatic Generation of Input Data for Distribution System Simulation Programs," 2018 IEEE Innovative Smart Grid Technologies - Asia (ISGT Asia), IEEE, 2018. Japan Meteorological Agency, Landslide Warning Information / Dosabikikuru (Risk Distribution of Heavy Rain Warnings (Landslides)), [Retrieved July 4, 2023], Internet<URL:https: / / www.jma.go.jp / jma / kishou / know / bosai / doshakeikai.html> Matsuda et al., "Considerations on Estimating the Risk of Landslides Nationwide Using Topographical and Geological Thematic Maps," National Institute for Land and Infrastructure Management Document No. 1120, 2020, [Retrieved July 4, 2023], Internet<URL: http: / / www.nilim.go.jp / lab / bcg / siryou / tnn / tnn1120.htm> Sakai et al., "Basic Study on Simple Reliability Assessment Method for Slopes During Earthquakes," Transactions of the Japan Society of Civil Engineers, No. 69, No. 4, pp. I_142-I_147, 2013, [Retrieved July 4, 2023], Internet<URL: https: / / www.jstage.jst.go.jp / article / jscejseee / 69 / 4 / 69_I_142 / _pdf / -char / ja> Ono et al., "An attempt to quantify the possibility of isolation in mountainous areas based on the probability of slope collapse during an earthquake," [Retrieved July 4, 2023], Internet<URL: https: / / www.jstage.jst.go.jp / article / jscejseee / 74 / 4 / 74_I_313 / _pdf / -char / ja> Maekawa et al., "Proposal of a disaster prevention management method for roadside slopes using IoT sensor devices," Journal of the Society for Regional Safety Studies No. 41, 2022, [Retrieved July 4, 2023], Internet<URL: https: / / www.jstage.jst.go.jp / article / jisss / 41 / 0 / 41_197 / _pdf / -char / ja>

[0007] As mentioned above, research is underway on methods for predicting damage to structures caused by typhoons or earthquakes and methods for predicting the probability of landslides. However, there has been insufficient research on methods for predicting damage to individual structures caused by rainfall-related disasters and proposing disaster countermeasures based on the predicted damage.

[0008] The purpose of the present disclosure, made in consideration of the above-mentioned problems, is to provide a prediction device that can predict damage to each structure caused by a disaster resulting from rainfall and propose countermeasures against the disaster based on the predicted damage.

[0009] In order to solve the above problem, the prediction device disclosed herein is a prediction device that predicts damage to each of a plurality of structures located in a target area due to a disaster caused by rainfall, and is equipped with: a prediction unit that inputs location information of each of the plurality of structures in the target area and surrounding environment information that indicates the surrounding environment of each of the plurality of structures into a prediction model that predicts the probability of damage to each of the plurality of structures due to a disaster caused by rainfall, the prediction model being constructed by learning learning data that corresponds, for each of the plurality of structures, the position of the structure, whether or not the structure will be damaged when a disaster caused by rainfall occurs, and the surrounding environment of the structure; and a proposal unit that outputs measures to deal with disasters caused by rainfall in the target area based on the probability of damage to each of the plurality of structures.

[0010] According to the prediction device of the present disclosure, it is possible to predict damage to each of a plurality of structures due to a disaster caused by rainfall, and to propose countermeasures against the disaster based on the predicted damage.

[0011] FIG. 1 is a diagram illustrating an example configuration of a prediction system according to an embodiment of the present disclosure. FIG. 1 is a diagram for explaining the operation of a data construction unit shown in FIG. 1. FIG. 2 is a diagram illustrating an example configuration of facility information shown in FIG. 2. FIG. 2 is a diagram illustrating an example configuration of surrounding environment information shown in FIG. 2. FIG. 3 is a diagram illustrating the configuration of prediction target data shown in FIG. 2. FIG. 1 is a diagram illustrating an example configuration of a model construction unit shown in FIG. 1. FIG. 4 is a diagram illustrating an example configuration of structure damage information shown in FIG. 4. FIG. 4 is a diagram illustrating an example configuration of disaster occurrence rainfall information shown in FIG. 4. FIG. 5 is a diagram illustrating the configuration of learning target structure information shown in FIG. 4. FIG. 4 is a diagram for explaining the operation of a selection unit shown in FIG. 4. A flowchart illustrating an example operation of the selection unit shown in FIG. 4. A flowchart illustrating an example operation of the learning unit shown in FIG. 4. FIG. 5 is a diagram illustrating an example prediction model constructed by the model construction unit shown in FIG. 1. A flowchart illustrating an example operation of the prediction unit shown in FIG. 1. A flowchart illustrating an example operation of the reserve quantity proposal unit shown in FIG. 1. A diagram for explaining the operation of the reserve quantity proposal unit shown in FIG. 1. A diagram illustrating an example definition of materials required for compensation for a structure. A diagram illustrating an example of a reserve quantity proposal result output by the reserve quantity proposal unit shown in FIG. 1. A flowchart illustrating an example operation of the route proposal unit shown in FIG. 1. A diagram illustrating an example of a route proposal result output by the route proposal unit shown in FIG. 1. A diagram illustrating an example of importance set for each structure. A flowchart illustrating another example operation of the route proposal unit shown in FIG. 1. Fig. 2 is a diagram for explaining the operation of a route suggestion unit shown in Fig. 1. Fig. 3 is a flowchart showing an example of the operation of the prediction device shown in Fig. 1. Fig. 4 is a diagram showing an example of the configuration of a computer that functions as the prediction device shown in Fig. 1.

[0012] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.

[0013] 1 is a diagram illustrating an example configuration of a prediction device 100 according to an embodiment of the present disclosure. The prediction device 100 according to the present disclosure predicts damage to each of a plurality of structures located in a target area due to a disaster caused by rainfall. The structures to be subjected to damage prediction include, for example, utility poles, bridges, pipelines, base stations, and the like, which are installed on the ground, but are not limited to these, and may be any structures installed in a dispersed manner in the target area.

[0014] As shown in FIG. 1, a prediction device 100 according to this embodiment includes a data construction unit 110, a model construction unit 120, a prediction unit 130, and a proposal unit 140.

[0015] The data constructing unit 110 constructs prediction target data that is used by the prediction unit 130, described below, to predict damage to each of multiple structures located in a target area due to a disaster caused by rainfall. The prediction target data constructed by the data constructing unit 110 is data related to structures of the same type. That is, as described above, structures may be various types of equipment such as utility poles, bridges, pipelines, or base stations, but the data constructing unit 110 constructs prediction target data for structures of the same type for which damage is predicted. Therefore, for example, if the structure for which damage is predicted is a utility pole, the data constructing unit 110 constructs prediction target data for the utility pole.

[0016] As shown in FIG. 2, the data construction unit 110 receives input of facility information and surrounding environment information.

[0017] The equipment information is information that indicates the position of each of a plurality of structures arranged in a target area, the structure's configuration, the material of the structure, etc. FIG. 3A is a diagram showing an example of the configuration of equipment information. FIG. 3A shows an example of the configuration of equipment information when the structure is a utility pole. As shown in FIG. 3A, the equipment information is information that associates, for example, the position, pole length, material, and structure name of the structure. In the equipment information, the position of the structure is specified, for example, by point coordinates.

[0018] The surrounding environment information indicates the topography of the structure (e.g., elevation, slope, maximum slope within a certain range, microtopography, soil type, road on which the structure is installed, etc.) and the land use status (e.g., land use classification) around the structure. FIG. 3B is a diagram showing an example of the configuration of the surrounding environment information. As shown in FIG. 3B, the surrounding environment information is information that associates, for example, the location of the structure, the elevation of the land around the structure, the slope of the land around the structure, the maximum slope angle within a certain range of the land around the structure, and the land use classification of the land around the structure. In the surrounding environment information, the location of the structure is specified, for example, by the coordinates of a polygon that includes the structure. For example, data provided by the Ministry of Land, Infrastructure, Transport and Tourism can be used as the land use classification. Alternatively, for example, data obtained from aerial photographs or satellite photographs can be used as the land use classification. In this case, for example, image analysis of the aerial photographs or satellite photographs can be performed to calculate the area of ​​the river around the structure and the distance between the structure and the river and use these as variables, or area determination can be performed based on the color tone of the image.

[0019] The data constructing unit 110 combines the above-mentioned facility information and surrounding environment information based on the location of the structure, and constructs prediction target data.

[0020] 3C is a diagram showing an example of the configuration of prediction target data. Fig. 3C shows an example of the configuration of prediction target data when the structure is a utility pole. As shown in Fig. 3C, the prediction target data is data in which the structure name, the structure location, the pole length, the structure material, the elevation of the land surrounding the structure, the slope of the land surrounding the structure, the maximum slope angle of the land surrounding the structure, and the land use classification of the area surrounding the structure are associated with each other.

[0021] Note that the items included in the prediction target data must match the items of the data used to build the prediction model described below. For example, if the elevation of a structure is used to build the prediction model, the prediction target data must also include the elevation of the structure. However, the method of calculating the elevation may be different. For example, when building the prediction model, the elevation determined based on a 250m mesh (a square area with 250m sides) may be used, and the prediction target data may use the elevation actually measured by surveying. Furthermore, the prediction target data must include the location of the structure, and there must be a one-to-one correspondence between the location information of the structure and the equipment information and surrounding environment information related to the structure.

[0022] The surrounding environment information may be mesh data for each predetermined range (for example, 1 km mesh (a square area of ​​1 km square)) as shown in FIG. 3B, or may be point data.

[0023] When the surrounding environment information is point data, the data constructing unit 110 may, for example, associate the position of a structure with the point data of the point closest to the structure. Furthermore, when the surrounding environment information is point data, the data constructing unit 110 may, for example, associate the position of a structure with the average value of the point data within a certain range around the structure. Furthermore, when the surrounding environment information is point data, the data constructing unit 110 may, for example, associate the position of a structure with the most frequently occurring value of the point data within a certain range around the structure.

[0024] 1 again, the model construction unit 120 constructs a prediction model for predicting the probability of damage caused by rainfall to each of a plurality of structures located in a target area. FIG. 4 is a diagram showing an example of the configuration of the model construction unit 120.

[0025] As shown in FIG. 4, the model construction unit 120 includes matching units 121 and 123, a selection unit 122, and a learning unit .

[0026] The collation unit 121 receives input of structural damage information and disaster-causing rainfall information.

[0027] Structure damage information is information indicating whether or not structures have been damaged in the past when disasters caused by rainfall occurred. Fig. 5A is a diagram showing an example of the configuration of structure damage information. As shown in Fig. 5A, structure damage information is information in which, for example, the location of a structure, the name of the structure, and whether or not it has been damaged are associated with each other. In the structure damage information, the location of a structure is specified, for example, by point coordinates.

[0028] Disaster-causing rainfall information indicates the amount of rainfall that occurred when past rainfall-related disasters occurred. FIG. 5B is a diagram showing an example of the configuration of disaster-causing rainfall information. As shown in FIG. 5B, disaster-causing rainfall information associates the location of each fixed range with the amount of rainfall that occurred within that fixed range when a rainfall-caused disaster occurred. The fixed range is specified, for example, by the coordinates of a polygon. The rainfall information included in the disaster-causing rainfall information includes the 72-hour predicted rainfall, the 6-hour predicted soil moisture content, the rainfall obtained from Kikikuru, and the actually measured rainfall.

[0029] The matching unit 121 matches the locations included in the structure damage information with the locations included in the disaster occurrence rainfall information, generates data that associates the structure location, structure name, whether or not the structure is damaged, and the rainfall at the structure location, and outputs the data to the selection unit 122.

[0030] The selection unit 122 selects data on rainfall amounts equal to or greater than a certain level from the data output from the matching unit 121, and generates learning target structure information. FIG. 6 is a flowchart showing an example of the operation of the selection unit 122.

[0031] The selection unit 122 sets a reference damage rate α (step S11). The damage rate is the ratio of structures damaged by a rainfall-induced disaster to structures not damaged at a certain rainfall amount.

[0032] Based on the data generated by the collation unit 121, the selection unit 122 generates a histogram showing the frequency of whether or not structures have been damaged by disasters caused by rainfall for each rainfall amount, as shown in FIG.

[0033] The selection unit 122 calculates the damage rate for each rainfall amount based on the generated histogram, and then sets the rainfall amount at which the calculated damage rate is equal to or greater than the set damage rate α as a threshold value R (step S13).

[0034] The selection unit 122 generates learning target structure information based on data generated by the matching unit 121 in which the rainfall is equal to or greater than the threshold R. FIG. 5C is a diagram showing an example of the configuration of learning target structure information. As shown in FIG. 5C, the learning target structure information is information in which the location of a structure, the name of the structure, and whether or not it has been damaged are associated with each other. In other words, the learning target structure information is information extracted from the structure damage information shown in FIG. 5A that covers locations where the rainfall is equal to or greater than the threshold R.

[0035] Referring again to FIG. 4 , the matching unit 123 receives input of the structure information to be learned and the surrounding environment information. The surrounding environment information input to the matching unit 123 is, for example, the same as the surrounding environment information described with reference to FIG. 3B . The matching unit 123 matches the positions included in the structure information to be learned with the positions included in the surrounding environment information to generate learning data in which the positions of the structures, the names of the structures, whether or not the structures are damaged, and the surrounding environment of the structures (such as the topography around the structures and the land use classification around the structures) are associated with each other. Here, the matching unit 123 includes all items of the surrounding environment information in the learning data.

[0036] The learning unit 124 constructs a prediction model by learning the training data generated by the matching unit 123. The learning unit 124 constructs the prediction model using a supervised learning method such as a gradient boosting decision tree or a random forest. Fig. 8 is a flowchart showing an example of the operation of the learning unit 124 when constructing a prediction model.

[0037] The learning unit 124 performs supervised learning using a method such as the gradient boosting decision tree or random forest described above (step S21) to create a prediction model.

[0038] The learning unit 124 evaluates which features of the training data contribute to the prediction of the prediction model using a method such as permutation importance (step S22).The learning unit 124 then removes variables (features) that do not affect the prediction of the prediction model from the training data, and re-learns the removed training data to create the prediction model (step S23).The learning unit 124 repeats this process to build the training model.The prediction target data built by the data building unit 110 described above needs to include the variables used to build the training model.

[0039] In this way, the model construction unit 120 constructs a predictive model that predicts the probability of damage to each of multiple structures due to rainfall by learning learning data that corresponds, for each of multiple structures, the location of the structure, whether the structure will be damaged in the event of a disaster caused by rainfall, and the surrounding environment of the structure.

[0040] As shown in Fig. 9, the model construction unit 120 constructs a prediction model for each type of rainfall information and each type of structure for which the probability of damage is predicted. In the example shown in Fig. 9, the rainfall information includes 24-hour rainfall, soil rainfall index, 60-minute accumulated rainfall, etc. The structures for which the probability of damage is predicted include utility poles and bridges.

[0041] Referring back to FIG. 1 , the prediction unit 130 receives rainfall information and prediction target data. The rainfall information indicates rainfall at multiple locations in a target area. Specifically, the rainfall information indicates rainfall at, for example, a 250-meter square, a 1-kilometer square, a polygon, or an arbitrary point. When the rainfall information indicates rainfall at an arbitrary point, the entire target area must be covered by space filling or other methods. The rainfall included in the rainfall information may be, for example, 24-hour rainfall, soil rainfall index, or 60-minute accumulated rainfall. These values ​​may be actual measurements or predicted values ​​for 72 hours or 6 hours in advance. However, a prediction model corresponding to the rainfall included in the rainfall information must be constructed.

[0042] FIG. 10 is a flowchart showing an example of the operation of the prediction unit 130. The prediction unit 130 reads out a prediction model from among the multiple prediction models constructed by the model construction unit 120, corresponding to the type of rainfall included in the rainfall information and the type of structure included in the prediction target data (the type of structure for which the damage probability is predicted) (step S31). The prediction unit 130 inputs the prediction target data into the read prediction model and obtains the prediction results of the damage probability for each of the multiple structures in the target area using the prediction model (step S32). As described above, the prediction target data includes location information for each of the multiple structures in the target area and surrounding environment information indicating the surrounding environment of each of the multiple structures. Therefore, the prediction unit 130 inputs the location information for each of the multiple structures in the target area and surrounding environment information indicating the surrounding environment of each of the multiple structures into the prediction model to predict the damage probability for each of the multiple structures. If there is no prediction model corresponding to the type of rainfall included in the rainfall information or the type of structure included in the prediction target data among the multiple prediction models constructed by the model construction unit 120, the prediction unit 130 does not perform a prediction.

[0043] As described above, the prediction model is constructed using data in which the rainfall is equal to or greater than the threshold value R, which is selected by the selection unit 122. Therefore, for example, when the rainfall information indicates that rainfall equal to or greater than the threshold value R is predicted or measured in the target area, it is possible to predict the probability of damage to each of multiple structures located in the target area.

[0044] 1 again, the proposal unit 140 outputs measures to deal with disasters caused by rainfall in a target area based on the prediction results (probability of damage to each of a plurality of structures) of the prediction unit 130. The proposal unit 140 outputs the disaster measures, for example, by displaying them on a display device not shown in FIG. 1. As shown in FIG. 1, the proposal unit 140 includes a reserve amount proposal unit 141 and a route proposal unit 142.

[0045] The reserve quantity proposal unit 141 receives road information and base information as input. The road information is information about roads established in the target area (e.g., road connections, road type (whether it is an emergency transportation road or not), road width, etc.). Also, typically, multiple bases are established to stockpile materials for repairing structures such as utility poles. The base information is information indicating the location of each of the multiple bases established in or around the target area to stockpile materials for repairing structures.

[0046] The stockpile quantity proposal unit 141 predicts the amount of supplies to be stockpiled at each of a plurality of base stations based on road information and base station information, and outputs the predicted amount as a supply quantity proposal result. Specifically, the stockpile quantity proposal unit 141 associates each of a plurality of structures located in a target area with one of a plurality of base stations based on the distance between the structure and each of the plurality of base stations. FIG. 11 is a flowchart showing an example of the operation of the stockpile quantity proposal unit 141 when associating structures with base stations. In the following description, it is assumed that five structures (structures a to e) are located around two base stations (base stations A and B) as shown in FIG. 12.

[0047] The reserve amount proposal unit 141 identifies one structure that is not assigned to a base (hereinafter referred to as an “unassigned structure”) among multiple structures located in the target area. In the following, structure b is assumed to be the unassigned structure.

[0048] The reserve quantity proposal unit 141 calculates the distance between the unassigned structure and each base (step S41). In the example shown in FIG. 12, the reserve quantity proposal unit 141 calculates the distance between structure b, which is an unassigned structure, and each of bases A and B. The reserve quantity proposal unit 141 may calculate the road distance as the distance between the structure and the base, as shown in FIG. 12. The road distance is the distance along the road connecting the structure and the base. In the example shown in FIG. 12, the reserve quantity proposal unit 141 calculates the road distance l between base A and structure b. Ａ and the road distance l between base B and structure b Ｂ Calculate.

[0049] The reserve amount proposing unit 141 associates the unassigned structure with the base station with the shortest calculated distance (step S42).Ａ <Road distance l Ｂ In this case, the reserve quantity proposal unit 141 associates the structure b with the base A.

[0050] Next, the reserve amount proposing unit 141 determines whether or not there are any unassigned structures remaining among the structures located in the target area (step S43).

[0051] If it is determined that there are any unassigned structures remaining (step S43: Yes), the reserve amount proposal unit 141 returns to the processing of step S41.

[0052] If it is determined that there are no unassigned structures remaining (step S43: No), the reserve quantity proposal unit 141 ends associating structures with bases. In Fig. 12, it is assumed that structures a, b, and d are associated with base A, and structures c and e are associated with base B. For each base, the reserve quantity proposal unit 141 creates a database that includes the structures associated with that base and the disaster probability of that structure.

[0053] The reserve amount proposing unit 141 may calculate the road distance by weighting according to the type (reliability) of the road. For example, as a weighting coefficient n according to the reliability of the road, n is set for an emergency transportation road that is likely to be passable even in the event of a disaster. １ (0<n １ < 1 is set, and for national highways with a width of Lm or more, n2 (0 < n １ <n ２ <1) is set. Also, assume that base A and structure b are connected by an emergency transportation road, and base B and structure b are connected by a national highway with a road width of L m or more. In this case, the reserve amount proposal unit 141 calculates the road distance l between base A and structure b. Ａ Coefficient n １ and the road distance l between base B and structure b Ｂ Coefficient n ２ The base to which structure b is associated may be determined by comparing the value multiplied by .

[0054] The reserve quantity proposing unit 141 associates each of the multiple structures with one of the bases, and then predicts the number of structures that will be affected by a disaster (the number of affected structures) for each of the multiple bases based on the probability of damage to the structures associated with that base. The reserve quantity proposing unit 141 may, for example, predict the number of affected structures based on the probability of damage to the structures associated with the base. Furthermore, the reserve quantity proposing unit 141 may, for example, determine the number of affected structures as the number of structures that have a probability of damage equal to or greater than a predetermined value, among the structures associated with the bases. Furthermore, the reserve quantity proposing unit 141 may, as the number of affected structures, add, for example, several percent as a safety factor to the number of affected structures predicted based on the probability of damage.

[0055] The preparedness quantity proposal unit 141 calculates the supplies required for each of multiple base stations based on the predicted number of affected areas. The amount of supplies required to repair each structure is defined in advance for each supply (item) required for repair, for example, as shown in FIG. 13A. The preparedness quantity proposal unit 141 calculates the supplies required for each base station based on this definition and the predicted number of affected areas for each base station. Then, as shown in FIG. 13B, the preparedness quantity proposal unit 141 outputs the predicted number of affected areas for each base station in association with the required amount of each supply.

[0056] Referring again to FIG. 1 , road information and route information are input to the route suggestion unit 142. The road information is information about roads established in the target area (e.g., road connections, road types (whether they are emergency transportation roads or not), road widths, etc.). Also, typically, multiple bases are established for inspecting structures such as utility poles. The base information is information indicating the location of each of the multiple bases for inspecting structures established in or around the target area.

[0057] The route proposal unit 142 determines a route for inspecting structures that have suffered damage due to a rainfall-induced disaster based on road information and base information, and outputs the determined route proposal result. Specifically, similar to the reserve quantity proposal unit 141, the route proposal unit 142 associates each of a plurality of structures located in a target area with one of a plurality of bases based on the distance between the structure and each of the plurality of bases. Then, for each of the plurality of bases, the route proposal unit 142 determines a route that travels around structures that have a damage probability equal to or greater than a predetermined threshold among the structures associated with the base, and outputs the determined route. FIG. 14 is a flowchart showing an example of the operation of the route proposal unit 142 when determining a route.

[0058] The route suggestion unit 142 sets a threshold value of the probability of damage to determine whether or not a structure should be inspected (step S51).

[0059] The route suggestion unit 142 sets a structure assigned to one base that has a disaster probability equal to or greater than a threshold as a target structure (step S52). The target structure is a structure to be inspected.

[0060] The route suggestion unit 142 sets the structure farthest from the base as the final destination structure among the destination structures that have not been set as either the final destination structure or an intermediate destination structure (described later) (step S53).

[0061] The route proposing unit 142 calculates the distance (first distance) from the destination structure most recently set as the final destination structure or intermediate destination structure (hereinafter referred to as the "structure of interest") to the closest destination structure (hereinafter referred to as the "nearest structure") among the destination structures connected to the structure of interest by a road. The route proposing unit 142 also calculates the distance (second distance) from the structure of interest to the base. The route proposing unit 142 determines whether the first distance is smaller than the second distance (step S54). That is, the route proposing unit 142 determines whether the nearest structure is closer to the structure of interest than the base.

[0062] If it is determined that the first distance is smaller than the second distance (step S54: Yes), the route proposing unit 142 sets the nearest structure as the intermediate destination structure (step S55) and returns to the processing of step S54.

[0063] If it is determined that the first distance is equal to or greater than the second distance (step S54: No), the route proposing unit 142 outputs, as a countermeasure, a route from the base to the intermediate and final destination structures in the reverse order of the order in which the intermediate and final destination structures were set (step S56). In this way, the route proposing unit 142 repeats the search for a route until the base is closer to the target structure than the nearest structure.

[0064] After outputting the route, the route proposing unit 142 determines whether there is a destination structure that has not been set as a final destination structure or an intermediate destination structure (an unset destination structure) (step S57).

[0065] If it is determined that there is an unset destination structure (step S57: Yes), the route proposing unit 142 returns to the process of step S53. If it is determined that there is no unset destination structure (step S57: No), the route proposing unit 142 ends the process.

[0066] Depending on the location of the destination structure, multiple routes (two routes (route A and B) in the example shown in FIG. 15) may be output for one base, as shown in FIG. 15. In the example shown in FIG. 15, route A is a route from the base to intermediate destination structure A. １ , intermediate purpose structure A ２ Route B is a route from the base to intermediate destination structure B. １ , intermediate purpose structure B ２ and the final destination structure B in order.

[0067] In this way, when outputting multiple routes for one base station, the route proposing unit 142 may determine the importance of each of the multiple routes based on the importance of destination structures included in the route, and may further output the determined importance for each route. For example, the route proposing unit 142 may store in advance the importance set for each structure, as shown in FIG. 16 , and determine the importance of the route based on the importance of structures included in the route.

[0068] For example, the route proposing unit 142 may determine the highest importance level among the importance levels of structures included in a route as the importance level of the route. Alternatively, the route proposing unit 142 may determine the average importance level of structures included in a route as the importance level of the route.

[0069] When outputting multiple routes for one base, the importance of each route can also be output, making it easier to determine inspection priorities.In addition, after a disaster occurs, when human resources are limited, inspection work on structures can be carried out in order of priority, making inspections more efficient.

[0070] Although Fig. 14 has been described using an example in which a route is proposed for each structure, the present disclosure is not limited to this. Fig. 17 is a flowchart showing another example of the operation of the route proposal unit 142. In Fig. 17, the same processes as those in Fig. 14 are denoted by the same reference numerals, and descriptions thereof will be omitted.

[0071] After setting the target structure (step S52), the path suggestion unit 142 divides the inspection area that can be inspected by one or more inspectors within a certain time into multiple parts and creates a mesh (step S61). The shape of the mesh is, for example, a rectangle obtained by dividing the inspection area in two directions (X direction and Y direction) as shown in Fig. 18, but is not limited to this and may be a polygon other than a rectangle or a circle.

[0072] The route proposing unit 142 sets, among the meshes obtained by dividing the inspection range, a mesh that includes the target structure as an inspection mesh (step S62).

[0073] The route proposing unit 142 calculates the shortest route from the base to the inspection mesh (step S63) and outputs it as a route. Here, if there are multiple inspection meshes, the route proposing unit 142 calculates the shortest route for each inspection mesh and outputs it as a route.

[0074] If multiple routes exist, the route proposing unit 142 sets a mesh importance for each inspection mesh (step S64). The route proposing unit 142 may, for example, set the maximum value of the importance of the target structures included in the inspection mesh as the mesh importance of the inspection mesh. Alternatively, the route proposing unit 142 may, for example, set the average value of the importance of the target structures included in the inspection mesh as the mesh importance of the inspection mesh.

[0075] The route proposing unit 142 sets and outputs the priorities of the multiple routes in the order of the mesh importance that has been set (step S65).

[0076] The route suggestion unit 142 may update the connection relationships between structures and change the route suggestion results by reading, for example, road conditions when a disaster occurs.

[0077] In addition, in this embodiment, the structure is described as a structure whose position can be expressed by point coordinates, such as a utility pole, but the present invention is not limited to this. For example, if the structure is a continuous structure such as a road or a levee, damage prediction can be performed in a similar manner by dividing the structure into multiple sections and setting the center of each section as the position of the structure. Furthermore, the prediction results of the prediction unit 130, the reserve quantity proposal results of the reserve quantity proposal unit 141, and the route proposal results of the route proposal unit 142 may be displayed on a map of the target area, for example.

[0078] Next, the operation of the prediction device 100 according to this embodiment will be described.

[0079] FIG. 19 is a flowchart showing an example of the operation of the prediction device 100 according to this embodiment, and is a diagram for explaining the prediction method performed by the prediction device 100.

[0080] The prediction unit 130 inputs the location information of each of the multiple structures in the target area and the surrounding environment information indicating the surrounding environment of each of the multiple structures into the prediction model, and predicts the probability of damage to each of the multiple structures (step S110). Here, the prediction model is constructed by learning from training data in which the location of each of the multiple structures, whether or not the structure will be damaged in the event of a rainfall-related disaster, and the surrounding environment of the structure are associated with each other.

[0081] The proposing unit 140 outputs countermeasures against disasters caused by rainfall in the target area based on the disaster probability of each of the multiple structures (step S120).

[0082] As described above, the prediction device 100 according to this embodiment includes a prediction unit 130 and a proposal unit 140. The prediction unit 130 inputs location information of each of the multiple structures in a target area and surrounding environment information indicating the surrounding environment of each of the multiple structures into a prediction model that predicts the probability of damage to each of the multiple structures due to rainfall, the prediction model being constructed by learning learning data that associates, for each of the multiple structures, the location of the structure, whether or not the structure will be damaged in the event of a rainfall-related disaster, and the surrounding environment of the structure. The proposal unit 140 outputs measures to deal with rainfall-related disasters in the target area based on the damage probability of each of the multiple structures.

[0083] By predicting the probability of damage to each structure using a predictive model constructed by learning from the above-mentioned training data, it is possible to extract structures that are predicted to be damaged in a disaster and propose countermeasures according to the damage to the structures.

[0084] The prediction device 100 described above can be realized by a computer 10 shown in FIG. 20 . A program for causing the computer 10 to function as the prediction device 100 may be provided. The program may be stored in a storage medium or provided via a network. FIG. 20 is a block diagram showing a schematic configuration of a computer 10 functioning as the prediction device 100. The computer 10 may be a general-purpose computer, a dedicated computer, a workstation, a PC (Personal Computer), an electronic notepad, or the like. The program instructions may be program code, code segments, or the like for performing necessary tasks.

[0085] 20, a computer 10 includes a processor 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, a storage 14, an input unit 15, a display unit 16, and a communication interface (I / F) 17. Each component is communicably connected to one another via a bus 19. The processor 11 is specifically a CPU (Central Processing Unit), an MPU (Micro Processing Unit), a GPU (Graphics Processing Unit), a DSP (Digital Signal Processor), a SoC (System on a Chip), or the like, and may be configured with multiple processors of the same or different types.

[0086] The processor 11 is a control unit that controls each component and performs various arithmetic operations. That is, the processor 11 reads a program from the ROM 12 or the storage 14 and executes the program using the RAM 13 as a work area. The processor 11 controls each component and performs various arithmetic operations in accordance with the program stored in the ROM 12 or the storage 14. In this embodiment, the ROM 12 or the storage 14 stores a program for causing the computer 10 to operate as the prediction device 100 according to the present disclosure. The program is read and executed by the processor 11 to realize each component of the prediction device 100, i.e., the data construction unit 110, the model construction unit 120, the prediction unit 130, and the proposal unit 140.

[0087] The program may be provided in a form stored on a non-transitory storage medium such as a CD-ROM (Compact Disk Read Only Memory), a DVD-ROM (Digital Versatile Disk Read Only Memory), a USB (Universal Serial Bus) memory, etc. The program may also be provided in a form downloaded from an external device via a network.

[0088] The ROM 12 stores various programs and various data. The RAM 13 temporarily stores programs or data as a working area. The storage 14 is configured with an HDD (Hard Disk Drive) or an SSD (Solid State Drive) and stores various programs including the operating system and various data.

[0089] The input unit 15 includes a pointing device such as a mouse and a keyboard, and is used to input various types of information.

[0090] The display unit 16 is, for example, a liquid crystal display, and displays various types of information. The display unit 16 may be a touch panel type and function as the input unit 15. The display unit 16 displays, for example, the prediction result of the prediction unit 130, the preparation amount proposal result of the preparation amount proposal unit 141, and the route proposal result of the route proposal unit 142.

[0091] The communication interface 17 is an interface for communicating with other devices, for example, an interface for a LAN.

[0092] The following additional notes are provided regarding the above-described embodiments.

[0093] [Supplementary Item 1] A prediction device that predicts damage to each of a plurality of structures located in a target area due to a disaster caused by rainfall, comprising a control unit, wherein the control unit inputs location information of each of the plurality of structures in the target area and surrounding environment information that indicates the surrounding environment of each of the plurality of structures into a prediction model that predicts the probability of damage to each of the plurality of structures due to a disaster caused by rainfall, the prediction model being constructed by learning learning data that associates, for each of the plurality of structures, the position of the structure, whether or not the structure will be damaged when a disaster caused by rainfall occurs, and the surrounding environment of the structure, to predict the probability of damage to each of the plurality of structures, and outputs measures to deal with disasters caused by rainfall in the target area based on the probability of damage to each of the plurality of structures.

[0094] [Supplementary Item 2] In the prediction device described in Supplementary Item 1, the control unit constructs the prediction model for each type of rainfall information and for each type of structure for which the damage probability is predicted by learning learning data that associates the positions of structures set at points where the rainfall indicated in the rainfall information is equal to or greater than a predetermined value, the presence or absence of damage to the structures, and the surrounding environment of the structures, and the prediction device reads out a prediction model corresponding to the type of rainfall information in a target area and the type of structure for which the damage probability is predicted, and predicts the damage probability of each of the plurality of structures.

[0095] [Supplementary Item 3] In the prediction device described in Supplementary Item 1 or 2, a plurality of bases are provided for stockpiling materials for repairing the structures, and the control unit associates each of the plurality of structures with one of the plurality of bases based on the distance between the structure and each of the plurality of bases, predicts the number of structures affected by the disaster for each of the plurality of bases based on the probability of damage to the structure associated with that base, and calculates the supplies required for the base based on the predicted number of affected structures.

[0096] [Supplementary Item 4] In the prediction device described in Supplementary Item 3, the control unit calculates, for each of the plurality of structures, a road distance that is a distance along a road connecting the structure and each of the plurality of bases, and associates the base with the structure having the smallest calculated road distance.

[0097] [Supplementary Item 5] The prediction device according to Supplementary Item 4, wherein the control unit calculates the road distance by applying a weight according to the type of the road.

[0098] [Supplementary Item 6] The prediction device according to any one of Supplementary Items 1 to 5, wherein a plurality of bases are provided for inspecting the structures, and the control unit associates each of the plurality of structures with one of the plurality of bases based on the distance between the structure and each of the plurality of bases, and outputs, for each of the plurality of bases, a route for inspecting structures whose damage probability is equal to or greater than a predetermined threshold, as the countermeasure.

[0099] [Supplementary Item 7] In the prediction device described in Supplementary Item 6, when multiple routes are output for one base station, the control unit determines the importance of each of the multiple routes based on the importance of a destination structure included in the route, and further outputs the determined importance for each route.

[0100] [Supplementary Item 8] In the prediction device according to any one of Supplementary Items 1 to 7, the surrounding environment of the structure includes at least one of a topography around the structure and a land use classification around the structure.

[0101] [Supplementary Item 9] A prediction method using a prediction device that predicts damage to each of a plurality of structures located in a target area due to a disaster caused by rainfall, the prediction method comprising: inputting location information of each of the plurality of structures in the target area and surrounding environment information indicating the surrounding environment of each of the plurality of structures into a prediction model that predicts the probability of damage to each of the plurality of structures due to a disaster caused by rainfall, the prediction model being constructed by learning learning data that associates, for each of the plurality of structures, the position of the structure, whether or not the structure will be damaged when a disaster caused by rainfall occurs, and the surrounding environment of the structure; predicting the probability of damage to each of the plurality of structures; and outputting measures to deal with a disaster caused by rainfall in the target area based on the probability of damage to each of the plurality of structures.

[0102] [Supplementary Item 10] A non-transitory storage medium storing a program executable by a computer, the non-transitory storage medium storing the program causing the computer to operate as the prediction device according to any one of Supplementary Items 1 to 8.

[0103] Although the above-described embodiments have been described as typical examples, it will be apparent to those skilled in the art that many modifications and substitutions can be made within the spirit and scope of the present disclosure. Therefore, the present invention should not be interpreted as being limited by the above-described embodiments, and various modifications and alterations are possible without departing from the scope of the claims. For example, multiple building blocks shown in the block diagrams of the embodiments can be combined into one, or one building block can be divided.

[0104] REFERENCE SIGNS LIST 10 Computer 11 Processor 12 ROM 13 RAM 14 Storage 15 Input unit 16 Display unit 17 Communication I / F 19 Path 100 Prediction device 110 Data construction unit 120 Model construction unit 121, 123 Collation unit 122 Selection unit 124 Learning unit 130 Prediction unit 140 Proposal unit 141 Reserve amount proposal unit 142 Route proposal unit

Claims

1. A prediction device that predicts damage to each of a plurality of structures located in a target area due to a disaster caused by rainfall, comprising: a prediction unit that inputs location information of each of the plurality of structures in the target area and surrounding environment information indicating the surrounding environment of each of the plurality of structures into a prediction model that predicts the probability of damage to each of the plurality of structures due to a disaster caused by rainfall, the prediction model being constructed by learning learning data that corresponds, for each of the plurality of structures, the position of the structure, the presence or absence of damage to the structure at the time of the occurrence of a disaster caused by rainfall, and the surrounding environment of the structure, and predicts the probability of damage to each of the plurality of structures; and a proposal unit that outputs measures against disasters caused by rainfall in the target area based on the probability of damage to each of the plurality of structures.

2. A prediction device as described in claim 1, further comprising a model construction unit which constructs the prediction model by learning learning data in which the positions of structures set at points where the amount of rainfall indicated in rainfall information is equal to or greater than a predetermined value, the presence or absence of damage to the structures, and the surrounding environment of the structures are associated with each other, the model construction unit constructs the prediction model for each type of rainfall information and for each type of structure for which the damage probability is predicted, and the prediction unit reads out a prediction model corresponding to the type of rainfall information in a target area and the type of structure for which the damage probability is predicted, and predicts the damage probability of each of the plurality of structures.

3. A prediction device as described in claim 1, wherein a plurality of bases are provided for stockpiling materials for repairing the structures, and the proposal unit associates each of the plurality of structures with one of the plurality of bases based on the distance between the structure and each of the plurality of bases, predicts for each of the plurality of bases the number of structures that will be affected by the disaster based on the probability of damage to the structures associated with that base, and calculates the supplies required for the base based on the predicted number of damages.

4. A prediction device as described in claim 1, wherein a plurality of bases are provided for inspecting the structures, and the proposal unit associates each of the plurality of structures with one of the plurality of bases based on the distance between the structure and each of the plurality of bases, and outputs, for each of the plurality of bases, a route for inspecting structures among the structures associated with the base, the structures having a damage probability equal to or greater than a predetermined threshold, as the countermeasure.

Images