WO2021199245A1 - 分析装置、分析方法及び記憶媒体 - Google Patents
分析装置、分析方法及び記憶媒体 Download PDFInfo
- Publication number
- WO2021199245A1 WO2021199245A1 PCT/JP2020/014752 JP2020014752W WO2021199245A1 WO 2021199245 A1 WO2021199245 A1 WO 2021199245A1 JP 2020014752 W JP2020014752 W JP 2020014752W WO 2021199245 A1 WO2021199245 A1 WO 2021199245A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- geospatial information
- displacement
- height
- value
- information
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
- G01S13/89—Radar or analogous systems specially adapted for specific applications for mapping or imaging
- G01S13/90—Radar or analogous systems specially adapted for specific applications for mapping or imaging using synthetic aperture techniques, e.g. synthetic aperture radar [SAR] techniques
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
- G01S13/89—Radar or analogous systems specially adapted for specific applications for mapping or imaging
- G01S13/90—Radar or analogous systems specially adapted for specific applications for mapping or imaging using synthetic aperture techniques, e.g. synthetic aperture radar [SAR] techniques
- G01S13/9021—SAR image post-processing techniques
- G01S13/9023—SAR image post-processing techniques combined with interferometric techniques
Definitions
- the present disclosure relates to a technique for analyzing the displacement of the ground surface height, and more particularly to a technique for analyzing the ground surface height measured by SAR (Synthetic Aperture Radar).
- SAR Synthetic Aperture Radar
- Patent Document 1 describes a surface change determination method for determining the presence or absence of surface change based on an interference SAR image using a change determination model generated by machine learning.
- the purpose of the present disclosure is to provide an analyzer that can determine the factors of fluctuations in the height of the ground surface.
- the analyzer is based on a plurality of types of geospatial information representing at least one of a state of the ground surface and an underground state of the ground surface, at each of a plurality of points on the ground surface.
- the combination of the first extraction means for extracting the value of the geospatial information and the geospatial information that the determination model contributes to the displacement of the height is determined based on at least a part of the values of the geospatial information.
- the analyzer is based on a plurality of types of geospatial information representing at least one of a state of the ground surface and an underground state of the ground surface, at each of a plurality of points on the ground surface. Height displacement at the plurality of points so that the first extraction means for extracting the value of the geospatial information and the determination model predict the displacement of the height based on at least a part of the values of the geospatial information.
- the analysis method is based on a plurality of types of geospatial information representing at least one of a state of the ground surface and an underground state of the ground surface, at each of a plurality of points on the ground surface.
- the plurality of geospatial information values are extracted so that the determination model determines the combination of the geospatial information that contributes to the height displacement based on at least a portion of the geospatial information values.
- the determination model is learned based on the displacement of the height at the point and the extracted value of the geospatial information.
- the analysis method is based on a plurality of types of geospatial information representing at least one of a state of the ground surface and an underground state of the ground surface, at each of a plurality of points on the ground surface.
- the determination model is learned based on the value of the geospatial information.
- the storage medium is based on a plurality of types of geospatial information representing at least one of a state of the ground surface and an underground state of the ground surface, at each of a plurality of points on the ground surface.
- the combination of the first extraction process for extracting the value of the geospatial information and the geospatial information in which the determination model contributes to the displacement of the height is determined based on at least a part of the values of the geospatial information.
- a program for causing a computer to execute a learning process for learning the determination model based on the displacement of the height at the plurality of points and the extracted value of the geospatial information is stored.
- the storage medium is based on a plurality of types of geospatial information representing at least one of a state of the ground surface and an underground state of the ground surface, at each of a plurality of points on the ground surface.
- the first extraction process for extracting the value of the geospatial information and the displacement of the height at the plurality of points so that the determination model predicts the displacement of the height based on at least a part of the values of the geospatial information.
- a learning process for learning the determination model based on the extracted values of the geospatial information and a program for causing the computer to execute the learning process are stored.
- One aspect of the present disclosure is also realized by the program stored in the above-mentioned storage medium.
- This disclosure has the effect of being able to determine the factors that cause fluctuations in the height of the ground surface.
- FIG. 1 is a block diagram showing an example of the configuration of the analyzer according to the first and second embodiments of the present disclosure.
- FIG. 2 is a flowchart showing an example of the operation of the analyzer 10 according to the first embodiment of the present disclosure.
- FIG. 3 is a block diagram showing a configuration of an analysis system of a modification of the first and second embodiments of the present disclosure.
- FIG. 4 is a block diagram showing an example of a detailed configuration of a learning device, an analysis device, and a geospatial information storage device included in the analysis system of the modified examples of the first and second embodiments of the present disclosure.
- FIG. 5 is a flowchart showing an example of the operation of the analyzer according to the first embodiment of the present disclosure.
- FIG. 6 is a diagram showing an example of the configuration of the analyzer according to the third and fourth embodiments of the present disclosure.
- FIG. 7 is a flowchart showing an example of the operation of the analyzer according to the third and fourth embodiments of the present disclosure.
- FIG. 8 is a diagram showing an example of the hardware configuration of the computer according to the embodiment of the present disclosure.
- FIG. 9 is a diagram showing an example of the type of embankment land.
- FIG. 10 is a diagram showing an example of surface geology.
- FIG. 11 is a diagram showing a riverbed.
- FIG. 1 is a block diagram showing an example of the configuration of the analyzer 10 according to the first embodiment of the present disclosure.
- the analyzer 10 includes a first receiving unit 111, a first extracting unit 112, a learning unit 113, a second receiving unit 121, a second extracting unit 122, and a determination unit 123. It includes an output unit 124, a model storage unit 125, and a geospatial information storage unit 131.
- the analyzer 10 may be realized as a combination of two or more devices that are communicably connected to each other.
- a terminal device for the user to input data to the analysis device 10 may be communicably connected to the analysis device 10 via, for example, a communication network.
- a communication network for example, a communication network.
- First receiving unit 111 receives data representing the displacement of the height of the ground surface as learning data.
- the user may use the terminal device described above to input data representing the displacement of the height of the ground surface to the first receiving unit 111.
- the first receiving unit 111 receives data representing the displacement of the height of the ground surface from the terminal device.
- the height displacement represents, for example, the transition of height at the same point (or a point considered to be the same) on the ground surface obtained by observations at multiple time points in the past. Height displacement is sometimes referred to as height variation.
- the height is, for example, the height at a point on the ground surface obtained by observation using a radar mounted on a flying object such as an artificial satellite or an aircraft as a synthetic aperture radar (SAR). In the following description, such observations will be referred to as observations by Synthetic Aperture Radar (SAR).
- SAR Synthetic Aperture Radar
- the transition of the height may be represented by data that can specify, for example, a plurality of values representing the height obtained by observations at a plurality of past time points and the order in which the heights are observed.
- the data representing the transition of the height may be, for example, data including a plurality of combinations of a value representing the height and data representing the time point at which the height is obtained by observation.
- the unit of data representing the time point may be appropriately determined.
- the data representing the time point may represent a date, or may represent a date and a time.
- the unit of time may also be set as appropriate.
- the displacement data may include information (eg, latitude and longitude information) representing the position of a point on the ground surface where the displacement of the height represented by the displacement data has been measured.
- the information representing the position may be other information that can identify the position on the ground surface.
- the information indicating the position of the point is referred to as the point information.
- the above-mentioned displacement of the height of the ground surface may represent the displacement of the height at each of a plurality of points on the ground surface.
- the data representing the displacement of the height of the ground surface is referred to as the ground surface displacement data.
- the ground surface displacement data may be a combination of displacement data at a plurality of points.
- the data representing the displacement of the height of the ground surface received by the first receiving unit 111 as learning data is referred to as learning displacement data.
- the training displacement data is data (aging displacement map) that represents the time-series displacement of the ground surface of the region obtained by analyzing the observation data obtained by observing the same region many times at multiple times. Also written). The transition of displacement in time series is referred to as aged displacement.
- the first receiving unit 111 sends the received learning displacement data to the first extracting unit 112.
- the first extraction unit 112 receives the learning displacement data from the first receiving unit 111. For example, the first extraction unit 112 extracts the point information of the point where the transition of the height represented by the displacement data included in the learning displacement data is observed from the learning displacement data. The first extraction unit 112 extracts the value of the geospatial information at the point whose position is represented by the extracted point information from the geospatial information stored in the geospatial information storage unit 131 described later.
- the geospatial information is, for example, information representing at least one of the state of the ground surface and the underground state of the ground surface.
- the geospatial information may be at least one of the information obtained from a so-called geographic information system (Geographic Information System).
- Geographic Information System Geographic Information System
- the geospatial information may be data obtained by observation from an artificial satellite, an aircraft, or the like.
- the geospatial information may be data obtained by a field survey.
- the geospatial information may be information representing the result of analysis based on the data obtained by measurement or survey.
- the geospatial information may be artificially determined information based on the data obtained by measurement or survey.
- the geospatial information may be referred to as GIS (Geographic Information System) data.
- the geospatial information may be acquired from the geospatial information system in advance and stored in the geospatial information storage unit 131.
- a plurality of types of geospatial information may be stored in the geospatial information storage unit 131.
- the geospatial information may be represented in a format in which the value of the geospatial information of the point specified by the point information (for example, latitude and longitude) can be specified. Specific examples of geospatial information will be described in detail later.
- the first extraction unit 112 may extract the value of the predetermined type of geospatial information at the point specified by the point information.
- the first extraction unit 112 may extract the values of all types of geospatial information stored in the geospatial information storage unit 131 at the points specified by the point information.
- the first extraction unit 112 does not have to extract the value of the geospatial information.
- the first extraction unit 112 sets the value of the geospatial information to a value indicating that the value does not exist (for example, 0, etc.). It may be set.
- the first extraction unit 112 extracts the learning displacement data (in other words, the aged displacement map) and the area where the learning displacement data represents the aged displacement (specifically, a plurality of points in the area).
- the value of the geospatial information is sent to the learning unit 113.
- Geospatial Information Storage 131 stores geospatial information.
- the geospatial information is stored in the geospatial information storage unit 131 in a form capable of identifying the state of the ground surface at a designated point.
- the geospatial information may be represented by, for example, a value representing a state for each mesh in which the ground surface is divided.
- the first extraction unit 112 extracts the value of the geospatial information representing the state in the mesh including the position specified by the point information as the value of the geospatial information of the point specified by the point information. do.
- the size and shape of the mesh may be determined for each type of geospatial information.
- the geospatial information may be represented in other formats.
- the geospatial information may be represented, for example, by a boundary line between regions having different states and a value representing a state within the region separated by the boundary line.
- the first extraction unit 1120 extracts a value representing a state in the area including the position specified by the point information as a value of the geospatial information of the point specified by the point information.
- the format of the geospatial information may be defined for each type of geospatial information.
- Specific geospatial information includes, for example, type of embankment site, average slope angle, average precipitation (for example, average annual precipitation), surface geology, steep slope designation, sediment disaster warning area designation, liquefaction risk, etc. It may be whether or not a rainwater infiltration basin is possible, easiness of shaking during an earthquake, lowland where drainage is difficult, land use in urban areas, natural terrain classification, artificial terrain classification, surface geology, riverbed, facility information (presence or absence of construction, etc.).
- the type of embankment construction site may represent the method of embankment, which is determined by the shape of the ground surface on which the embankment was made.
- the types of embankment construction sites are, for example, “valley-filled embankment”, which is an embankment in which valleys and swamps are filled with embankment, and "belly-type embankment”, which is an embankment made on slopes.
- the type of embankment site may further represent the scale of the embankment.
- the type of embankment satisfying the standard for example, embankment having an area of 3000 square meters or more
- the type of embankment that does not meet the criteria may be "valley-filled embankment”.
- the angle of the embankment that meets the standard is 20 degrees or more with respect to the horizontal plane, and the height of the embankment is 5.
- the type of embankment that is greater than or equal to a meter may be a large-scale embankment.
- the type of embankment that does not meet the criteria may be "belly-type embankment".
- the value of the type of embankment construction site may be any one of different numerical values, which are appropriately assigned in advance to, for example, "valley-filled embankment” and "belly-filled embankment”.
- FIG. 9 is a diagram showing an example of the type of embankment land. The example shown in FIG. 9 shows the distribution of the embankment land on the ground surface for each type of embankment land.
- the average inclination angle may be, for example, data of the average inclination angle of the ground surface calculated in mesh units.
- the value of the average inclination angle may be the calculated average inclination angle of the ground surface.
- the average precipitation may be, for example, data of the average precipitation on the ground surface calculated in mesh units.
- the average precipitation value may be the calculated average precipitation on the ground surface.
- the surface geology may be data representing the surface geology (in other words, the type of geology) on the surface of the earth.
- the type of geology may be predetermined. Different numerical values may be assigned to each type of geology in advance.
- the value of the surface geology may be any one of the numerical values appropriately assigned to the geology in advance.
- the steep slope designation may be data indicating whether or not it is designated as a steep slope by, for example, a local government.
- the value of the steep slope designation may be, for example, a numerical value indicating that the slope is designated as a steep slope, or a numerical value indicating that the slope is not designated as a steep slope. As these numerical values, numerical values different from each other may be appropriately determined in advance.
- the sediment-related disaster warning area designation may indicate whether or not it has been designated as a sediment-related disaster warning area by, for example, a local government.
- the value of the sediment-related disaster warning area designation may be, for example, a numerical value indicating that it is designated as a sediment-related disaster warning area, or a numerical value indicating that it is not designated as a sediment-related disaster warning area. As these numerical values, numerical values different from each other may be predetermined.
- the liquefaction risk may be, for example, data indicating the degree of risk of land liquefaction.
- the liquefaction risk value may be a numerical value indicating the degree of liquefaction risk of the land.
- the value of the liquefaction risk may be any one of a plurality of different numerical values representing different degrees. A numerical value indicating the degree of risk may be appropriately determined in advance.
- Whether or not a rainwater infiltration basin can be installed is information indicating whether or not an infiltration facility can be installed based on the "infiltration facility installation judgment map" that indicates the result of determining whether or not an infiltration facility can be installed based on, for example, topography, soil quality, and groundwater level. It's okay.
- the value of whether or not the rainwater infiltration basin can be installed may be a numerical value indicating that the installation is possible or a numerical value indicating that the installation is not possible. As these numerical values, numerical values different from each other may be appropriately determined in advance.
- the easiness of shaking at the time of an earthquake may be, for example, data indicating the degree of easiness of shaking of the ground surface when an earthquake occurs.
- the value of easiness of shaking at the time of an earthquake may be a numerical value indicating the degree of easiness of shaking of the ground surface in the event of an earthquake.
- the value of easiness of shaking at the time of an earthquake may be any one of a plurality of numerical values indicating the degree of easiness of shaking of the ground surface in the event of an earthquake.
- a numerical value indicating the degree of easiness of shaking of the ground surface may be appropriately determined in advance.
- the difficult-to-drain lowland may represent, for example, whether or not the land is a difficult-to-drain lowland estimated from the altitude of the land or the difference in altitude from the surroundings.
- the value of the lowland where drainage is difficult may be a numerical value indicating that the lowland is difficult to drain, or a numerical value indicating that the lowland is not difficult to drain. These numerical values may be appropriately determined in advance.
- Urban land use may be a type of land use in the area designated as a city.
- the type of land use in urban land use may be read, for example, from satellite images.
- a land use type selected from a plurality of predetermined types may be set for an area included in an urban area. Different numerical values may be appropriately assigned to each of the plurality of predetermined types.
- the value of the land use type set for the area may be the numerical value assigned to the type.
- the natural terrain classification may be, for example, a type of terrain in a place that is not a building built by humans.
- a plurality of terrain types that can be set as natural terrain classification may be appropriately determined in advance.
- the terrain type selected from a plurality of terrain types predetermined as the types that can be set as the natural terrain classification may be set.
- Different numerical values may be assigned to each of the plurality of types.
- the terrain value in the natural terrain classification set in the area may be a numerical value assigned to the terrain type set in the area.
- the artificial terrain classification may be, for example, a type of terrain in a place where a human has modified the terrain or a place where a building is built by a human.
- a plurality of terrain types that can be set as artificial terrain classification may be appropriately determined in advance.
- the terrain type selected from a plurality of terrain types predetermined as the types that can be set as the artificial terrain classification may be set.
- Different numerical values may be assigned to each of the plurality of types.
- the terrain value in the artificial terrain classification set in the area may be a numerical value assigned to the type of terrain set in the area.
- the surface geology may represent, for example, the type of soil on the surface of the earth.
- a plurality of soil types may be appropriately determined in advance.
- different numerical values which are appropriately determined in advance, may be assigned to each of the plurality of soil types.
- the type of soil based on the results of the survey may be set for the area.
- the surface geological value of the area may be a numerical value assigned to the type of soil set in the area.
- FIG. 10 is a diagram showing an example of surface geology.
- FIG. 10 shows the distribution of geology in the surface layer including the ground surface.
- the riverbed may be information indicating whether or not the area is a riverbed.
- a numerical value indicating that it is a riverbed in other words, a numerical value indicating a riverbed
- another numerical value indicating that it is not a riverbed in other words, a numerical value indicating a non-riverbed
- a numerical value representing the riverbed may be set in the area that is the riverbed.
- Numerical values representing non-river beds may be set in areas that are not riverbeds.
- the value of the riverbed of the area may be a numerical value indicating the riverbed or a numerical value representing a non-riverbed set in the area.
- FIG. 11 is a diagram showing a riverbed. In FIG. 11, the area determined to be the riverbed and the other areas are drawn.
- Facility information represents information about the facility.
- the facility information may represent any of various information about the facility, which is predetermined.
- the facility information indicates whether or not construction is underway.
- a numerical value indicating that construction is underway and another numerical value indicating that construction is not underway may be appropriately set in advance.
- Facility information indicating that construction is underway may be set for the area under construction.
- Facility information indicating that construction is not underway may be set for an area that is not under construction.
- a numerical value indicating that the construction is underway may be set.
- a numerical value indicating that it is not under construction may be set.
- the learning unit 113 receives the learning displacement data and the extracted geospatial information value of the region where the learning displacement data represents the secular displacement from the first extraction unit 112.
- the learning unit 113 performs learning using the received displacement data for learning and the value of the geospatial information. In this learning, the learning unit 113 determines a combination of geospatial information that contributes to the displacement of the height at the target point based on at least a part of the values of the geospatial information of the target point. , Learn the judgment model.
- the determination model of the present embodiment represents, for example, a parameter of a program that receives a value of geospatial information and outputs a combination of geospatial information that contributes to height displacement according to the value of the received geospatial information. It's okay.
- the determination model is, for example, when the value of the received geospatial information satisfies the condition for at least a part of the values of the geospatial information, the geography that contributes to the displacement of the height according to the condition. It may represent a parameter of a program that outputs a combination of spatial information.
- the judgment model is a combination of a condition that the received geospatial information value is for at least a part of the geospatial information value and a geospatial information that contributes to the height displacement when the condition is satisfied. Represented by. It should be noted that a plurality of conditions may exist.
- Each of the plurality of conditions may be a condition for at least a part of geospatial information that is not necessarily the same.
- a processor and a computer including such a processor) that executes the above-mentioned program using the above-mentioned parameters is also referred to as a determiner below.
- the learning unit 113 uses heterogeneous mixed learning as the learning algorithm.
- the learning algorithm is another algorithm that can learn a judgment model that receives the value of the geospatial information and outputs the combination of the geospatial information that contributes to the displacement of the height according to the condition for the value of the geospatial information.
- multivariate analysis may also be used for multiple regression analysis and the like.
- Heterogeneous mixture learning is described, for example, in the following references.
- Heterogeneous mixture learning refers to learning of a heterogeneous mixture prediction model that makes predictions by combining prediction models based on a combination of different explanatory variables.
- the heterogeneous mixture prediction model is represented by, for example, a plurality of sets of a combination of conditional expressions and a prediction formula when all the conditional expressions included in the combination are satisfied.
- Each conditional expression is, for example, a conditional expression for the value of any one explanatory variable.
- the combination of conditional expressions includes one or more conditional expressions.
- Each prediction formula is represented by a linear form of explanatory variables that are not necessarily the same.
- the learning unit 113 performs heterogeneous mixed learning so as to predict future height displacement, for example, height displacement after a predetermined period, using, for example, height displacement as an objective variable and geospatial information as an explanatory variable. conduct.
- the predetermined period may be appropriately set in advance.
- the learning unit 113 can obtain a plurality of sets of a combination of conditional expressions and a prediction expression when all the conditional expressions included in the combination are satisfied.
- Each of the conditional expressions represents a condition for a value of one geospatial information that is not necessarily the same.
- the combination of conditional expressions includes one or more conditional expressions as described above. The combination of these conditional expressions is referred to as a case classification condition.
- the prediction formula is a formula for predicting the displacement of the height.
- Each of the prediction formulas is represented by the linear form (linear sum) of one or more explanatory variables.
- Each of the explanatory variables represents any one geospatial information. It can be said that the geospatial information represented by the explanatory variables included in the prediction formula is the geospatial information that contributes to the displacement of the height.
- the case classification condition is satisfied means that all the conditional expressions included in the case classification condition are satisfied.
- the prediction formula for the case classification condition represents the prediction formula when the case classification condition is satisfied.
- the learning unit 113 determines the geospatial information represented by the explanatory variables included in the prediction formula for the case classification condition as the geospatial information that contributes to the displacement of the height. To generate.
- the determination model outputs the information of the geospatial information determined as the geospatial information that contributes to the displacement of the height.
- "generating a judgment model” refers to learning the judgment model and generating data representing the judgment model obtained by the learning.
- the learning unit 113 stores the obtained determination model (in other words, data representing the obtained determination model) in the model storage unit 125.
- Second receiving unit 121 receives information (for example, latitude and longitude information) that identifies the position of a point on the ground surface.
- information for example, latitude and longitude information
- the user may use the terminal device described above to input information for identifying the position of a point on the ground surface into the second receiving unit 121.
- the second receiving unit 121 receives information for identifying the position of a point on the ground surface from the terminal device.
- the point information received by the second receiving unit 121 is referred to as target point information.
- a point whose position is specified by the target point information is referred to as a target point.
- the target point information may represent the position of one target point. In that case, the target point information may include, for example, one combination of information representing latitude and information representing longitude.
- the target point information may represent the positions of a plurality of target points. In that case, the target point information may include, for example, a plurality of combinations of information representing latitude and information representing longitude.
- the target point information may represent, for example, the positions of a plurality of points (also referred to as grid points) that are regularly arranged in the area.
- the target point information may include information for specifying the area and information for specifying the target point in the area.
- the information for specifying the region is, for example, when the shape of the region is rectangular, for example, the latitude and longitude of one vertex and two vectors (first vector) representing the two sides of the rectangle starting from that vertex. It may include a vector and a second vector).
- the information for identifying the target point in the region may be, for example, an interval in which the target point exists in the direction of the first vector and an interval in which the target point exists in the direction of the second vector.
- the target point information is not limited to these examples.
- each part described below may repeat the operation for one target point for the plurality of target points.
- the second receiving unit 121 sends the received target point information to the second extracting unit 122.
- Second extraction unit 122 receives the target point information from the second receiving unit 121.
- the second extraction unit 122 extracts the value of the geospatial information at the target point specified by the received target point information from the geospatial information stored in the geospatial information storage unit 131.
- the second extraction unit 122 may extract a predetermined value of the geospatial information from all the geospatial information stored in the geospatial information storage unit 131. In this case, for example, the geospatial information that has no relation to the condition for the value of the geospatial information and is confirmed in advance not to contribute to the displacement of the height may be excluded from the target of the value extraction. If there is geospatial information for which a value has not been set at the target point, the second extraction unit 122 may set the value of the geospatial information to a numerical value (for example, 0) indicating that the value does not exist. good.
- a numerical value for example, 0
- the second extraction unit 122 sends the received target point information and the extracted value of the geospatial information at the target point to the determination unit 123.
- the determination unit 123 receives the value of the geospatial information at the target point from the second extraction unit 122.
- the determination unit 123 may receive the target point information from the second extraction unit 122.
- the determination unit 123 determines the combination of geospatial information that contributes to the height displacement at the target point according to the determination model stored in the model storage unit 125. Specifically, the determination unit 123 specifies, for example, the condition satisfied by the received value of the geospatial information at the target point among the plurality of conditions included in the determination model. The determination unit 123 determines that the combination of geospatial information that contributes to the height displacement when the specified condition is satisfied is the combination of the geospatial information that contributes to the height displacement at the target point. The combination of geospatial information that contributes to the displacement of the height at the target point can be regarded as the cause of the fluctuation in height. The number of types of geospatial information included in the combination of geospatial information may be one. The number of types of geospatial information included in the combination of geospatial information may be two or more.
- the determination unit 123 sends the determined information of the combination of geospatial information that contributes to the displacement of the height at the target point to the output unit 124.
- Output unit 124 receives from the determination unit 123 information on a combination of geospatial information that contributes to the displacement of the height at the target point.
- the output unit 124 outputs the received information of the combination of geospatial information that contributes to the displacement of the height at the target point.
- the output unit 124 may display, for example, a combination of geospatial information that contributes to the displacement of the height at the target point on a display or the like.
- the output unit 124 may send a combination of geospatial information that contributes to the displacement of the height at the target point to another information processing device, the terminal device described above, or the like.
- FIG. 2 is a flowchart showing an example of the operation of the analyzer 10 of the present embodiment.
- the first receiving unit 111 receives the displacement and the position of the height at a plurality of points on the ground surface (step S101). Specifically, the first receiving unit 111 receives information on displacement of height at a plurality of points on the ground surface and point information representing the positions of the plurality of points as learning displacement data. As mentioned above, the height displacement is, for example, the height displacement obtained by observation by SAR. The first receiving unit 111 sends the received height displacements and positions at a plurality of points on the ground surface to the first extracting unit 112.
- the first extraction unit 112 extracts the value of the geospatial information at a plurality of points (step S102). That is, the first extraction unit 112 stores the value of the geospatial information at the position indicated by the point information of each of the plurality of points received by the first receiving unit 111 in the geospatial information storage unit 131. Extract from spatial information. The first extraction unit 112 obtains the data received as the displacement data for learning (that is, the displacement of the height at a plurality of points on the ground surface and the point information of the plurality of points) and the value of the extracted geospatial information. , Is sent to the learning unit 113.
- the displacement data for learning that is, the displacement of the height at a plurality of points on the ground surface and the point information of the plurality of points
- the learning unit 113 learns the determination model (step S103). Specifically, the learning unit 113 receives the displacement of the height at a plurality of points on the ground surface, the point information of the plurality of points, and the value of the extracted geospatial information from the first extraction unit 112. .. The learning unit 113 learns the above-mentioned determination model by using the displacement of the height and the value of the geospatial information at each of the plurality of points. The learning unit 113 stores the determination model obtained as a result of learning in the model storage unit 125.
- the analyzer 10 may perform the above operations from step S101 to step S103 in advance. It is not necessary to perform the operation of step S104 following step S103.
- step S104 the second receiving unit 121 receives the position of the target point (that is, the target point information).
- the second receiving unit 121 sends the received target point information to the second extracting unit 122.
- the second extraction unit 122 extracts the value of the geospatial information at the target point (step S105).
- the second extraction unit 122 may extract the value of the geospatial information at the position specified by the received target point information.
- the determination unit 123 determines the combination of geospatial information that contributes to the displacement of the height at the target point by the determination model (step S106).
- the output unit 124 outputs the obtained combination of geospatial information (step S107).
- the analyzer 10 may perform the operations from step S104 to step S107 for each of the plurality of target points, for example.
- step S104 the analyzer 104 may collectively receive the displacements and positions of the heights of the plurality of target points. Then, the analyzer 104 may perform the operations of step S105 and step S106 for each of the plurality of target points.
- step S107 the analyzer 104 may collectively output a combination of geospatial information of a plurality of target points.
- This embodiment has the effect of being able to determine the factors that cause fluctuations in the height of the ground surface.
- the reason is that the learning unit 113 contributes to the displacement of the height at the position of the target point as a factor of the fluctuation of the height based on at least a part of the values of the geodata information at the position of the target point. This is because the determination model for determining the combination of is learned.
- the determination model generated by the learning unit 113 of this modification outputs a value indicating the magnitude of the contribution of the geospatial information in addition to the information of the geospatial information that contributes to the displacement of the height.
- the learning unit 113 uses, for example, the displacement of the height as the objective variable and the geospatial information as the explanatory variable to perform heterogeneous mixed learning, thereby performing the case classification condition and the prediction formula for the case classification condition. You can get multiple pairs of. Prior to learning, the learning unit 113 of the present embodiment converts the value of each geospatial information so that the range of the geospatial information is the same for each geospatial information (for example, 0 or more and 1 or less). ..
- the case classification condition is a combination of conditional expressions.
- the prediction formula for the case classification condition is a prediction formula when all the conditional expressions included in the case classification condition are satisfied.
- the prediction formula is represented by the linear form of the explanatory variables.
- Explanatory variables represent geospatial information.
- the learning unit 113 considers the geospatial information represented by the explanatory variables included in the prediction formula as the geospatial information that contributes to the displacement of the height. Then, the learning unit 113 considers the coefficient of the explanatory variable representing the geospatial information as the magnitude of the contribution of the geospatial information in the prediction formula.
- the learning unit 113 generates the following determination model.
- the judgment model determines the geospatial information represented by the explanatory variables included in the prediction formula for the case classification condition as the geospatial information that contributes to the displacement of the height.
- the judgment model also uses the coefficients of the explanatory variables included in the prediction formula for the case classification conditions when the case classification conditions are satisfied, and the magnitude of the contribution to the displacement of the height of the geospatial information represented by the explanatory variables. Judge as a sword.
- the determination model outputs the information of the geospatial information determined as the geospatial information contributing to the displacement of the height and the information indicating the magnitude of the contribution of the geospatial information to the displacement of the height.
- the determination unit 123 refers to a combination of geospatial information that contributes to the height displacement at the target point and the geospatial information included in the combination with respect to the height displacement according to the determination model stored in the model storage unit 125. Determine the magnitude of the contribution. Specifically, the determination unit 123 specifies, for example, a condition satisfied by the value of the received geospatial information among a plurality of conditions included in the determination model. The determination unit 123 determines that the combination of geospatial information that contributes to the height displacement when the specified condition is satisfied is the combination of the geospatial information that contributes to the height displacement at the target point.
- the determination unit 123 determines that the magnitude of the contribution of the geospatial information to the height displacement, which contributes to the height displacement when the specified condition is satisfied, is the height of the geospatial information at the target point. It is determined that it is the magnitude of the contribution to the displacement.
- the determination unit 123 sends to the output unit 124 information on the combination of geospatial information that contributes to the displacement of the height at the target point and information indicating the magnitude of the contribution of the geospatial information included in the combination.
- Output unit 124 receives from the determination unit 123 information on the combination of geospatial information that contributes to the displacement of the height at the target point and information indicating the magnitude of the contribution of the geospatial information included in the combination.
- the output unit 124 outputs the received information on the combination of geospatial information that contributes to the displacement of the height at the target point and the information indicating the magnitude of the contribution of the geospatial information included in the combination.
- the output unit 124 may display, for example, a combination of geospatial information that contributes to the displacement of the height at the target point and the magnitude of the contribution on a display or the like.
- the output unit 124 may send out the combination of geospatial information that contributes to the displacement of the height at the target point and the magnitude of the contribution to another information processing device, the terminal device described above, or the like.
- FIG. 3 is a block diagram showing the configuration of the analysis system 1 of the modified example of the first embodiment.
- the analysis system 1 includes a learning device 11, an analysis device 21, a geospatial information storage device 31, and a terminal device 51.
- the learning device 11, the analyzer 21, the geospatial information storage device 31, and the terminal device 51 are communicably connected to each other by the network 40, which is a communication network.
- the analysis system 1 realizes the functions of the analysis device 10 of the first embodiment by the learning device 11, the analysis device 21, and the geospatial information storage device 31.
- the terminal device 51 is the terminal device described above.
- FIG. 4 is a block diagram showing an example of a detailed configuration of the learning device 11, the analysis device 21, and the geospatial information storage device 31 included in the analysis system 1 of this modified example.
- the transfer of data between the components of the learning device 11, the analyzer 21, and the geospatial information storage device 31 realized by the network 40 of FIG. 3 is drawn by a line connecting the components. ing.
- the learning device 11 includes a first receiving unit 111, a first extracting unit 112, a learning unit 113, a first reading unit 114, and a transmitting unit 115.
- the first receiving unit 111, the first extraction unit 112, and the learning unit 113 are the same as the units of the first embodiment having the same name and having the same reference numerals.
- the first reading unit 114 reads the geospatial information from the geospatial information storage unit 131 of the geospatial information storage device 31 via the input / output unit 132. Specifically, the first reading unit 114 transmits a request for geospatial information to the input / output unit 132 of the geospatial information storage device 31, and is read from the geospatial information storage unit 131 by the input / output unit 132. , The requested geospatial information may be received from the input / output unit 132.
- the request for geospatial information may include point information that identifies the point (eg, latitude and longitude information).
- the requested geospatial information refers to the value of the geospatial information at the point specified by the point information.
- the transmission unit 115 transmits the determination model (in other words, the parameter of the determination device) learned by the learning unit 113 to the analyzer 21.
- the analyzer 21 includes a second receiving unit 121, a second extracting unit 122, a determination unit 123, an output unit 124, a model storage unit 125, a second reading unit 126, and a receiving unit 127.
- the second receiving unit 121, the second extraction unit 122, the determination unit 123, the output unit 124, and the model storage unit 125 are the same as the units of the first embodiment having the same name and having the same reference numerals. be.
- the second reading unit 126 reads the geospatial information from the geospatial information storage unit 131 of the geospatial information storage device 31 via the input / output unit 132. Specifically, the second reading unit 126 transmits a request for geospatial information to the input / output unit 132 of the geospatial information storage device 31, and is read from the geospatial information storage unit 131 by the input / output unit 132. , The requested geospatial information may be received from the input / output unit 132.
- the request for geospatial information may include point information that identifies the point (eg, latitude and longitude information).
- the request for geospatial information generated and transmitted by the second reading unit 126 may include type information that specifies the type of geospatial information.
- a plurality of types of type information may be specified.
- the input / output unit 132 sets the values of all types of geospatial information specified by the type information at the points specified by the point information, as described later. , Is sent to the second reading unit 126.
- the receiving unit 127 receives the determination model from the transmitting unit 115 of the learning device 11.
- the receiving unit 127 stores the received determination model in the model storage unit 125.
- the geospatial information storage device 31 includes a geospatial information storage unit 131 and an input / output unit 132.
- the geospatial information storage unit 131 is the same as the geospatial information storage unit 131 of the first embodiment.
- the input / output unit 132 receives a request for geospatial information.
- the source of the request for geospatial information is the first reading unit 114 or the second reading unit 126.
- the request for geospatial information may include information that identifies the location.
- the input / output unit 132 extracts the value of the geospatial information of the point specified by the information for specifying the point included in the request for the geospatial information from the geospatial information stored in the geospatial information storage unit 131. ..
- the input / output unit 132 may extract the values of all types of geospatial information at the specified points.
- the request for geospatial information may include information that identifies the type of geospatial information.
- the input / output unit 132 may extract the values of all types of geospatial information included in the request for geospatial information, which are specified by the information that specifies the type of geospatial information.
- the input / output unit 132 transmits the extracted geospatial information value to the source of the geospatial information request.
- the operation of the analysis system 1 of this modification is the same as the operation of the analysis device 10 of the first embodiment shown in FIG. 2, except for the following differences.
- the above-mentioned difference is that, for example, the reading of the geospatial information is performed via the first reading unit 114 and the input / output unit 132, or via the second reading unit 126 and the input / output unit 132, and the determination model. Is a point where the delivery is performed via the transmission unit 115 and the reception unit 127.
- FIG. 1 is a diagram showing a configuration of an analyzer 10 according to a second embodiment of the present disclosure.
- the configuration of the analyzer 10 of the present embodiment is the same as the configuration of the analyzer 10 of the first embodiment.
- the components of the analyzer 10 of the present embodiment are the same as the components of the analyzer 10 of the first embodiment, which are given the same name and reference numeral, except for the differences described below.
- the learning unit 113 of the present embodiment learns a determination model different from the determination model learned by the learning unit 113 of the first embodiment.
- the learning unit 113 of the present embodiment is the same as the learning unit 113 of the first embodiment.
- the learning unit 113 of the present embodiment receives the learning displacement data and the region in which the learning displacement data represents the secular displacement from the first extraction unit 112. Receives the extracted geospatial information values.
- the learning unit 113 of the present embodiment stores the determination model obtained by learning in the model storage unit 125.
- the learning unit 113 of the present embodiment also performs heterogeneous mixed learning so as to predict, for example, the height displacement after a predetermined period, for example, using the height displacement as the objective variable and the geospatial information as the explanatory variable.
- the learning unit 113 can obtain a plurality of sets of a combination of conditional expressions and a prediction expression when all the conditional expressions included in the combination are satisfied.
- each of the conditional expressions represents a condition for one geospatial information value that is not necessarily the same.
- the combination of conditional expressions includes one or more conditional expressions as described above. The combination of these conditional expressions is referred to as a case classification condition.
- the prediction formula is a formula for predicting the displacement of the height.
- Each of the prediction formulas is represented by the linear form of one or more explanatory variables.
- Each of the explanatory variables represents any one geospatial information. It can be said that the geospatial information represented by the explanatory variables included in the prediction formula is the geospatial information that contributes to the displacement of the height.
- the case classification condition is satisfied means that all the conditional expressions included in the case classification condition are satisfied.
- the prediction formula for the case classification condition represents the prediction formula when the case classification condition is satisfied.
- the learning unit 113 of the present embodiment When the case classification condition is satisfied, the learning unit 113 of the present embodiment generates a determination model that predicts the height displacement by the prediction formula for the case classification condition.
- the determination model outputs information on the displacement at the predicted height.
- the determination unit 123 of the present embodiment receives the value of the geospatial information at the target point from the second extraction unit 122, similarly to the determination unit 123 of the first embodiment.
- the determination unit 123 may receive the target point information from the second extraction unit 122.
- the determination unit 123 of the present embodiment predicts the displacement of the height at the target point according to the determination model stored in the model storage unit 125. Specifically, the determination unit 123 specifies, for example, the condition satisfied by the received value of the geospatial information at the target point among the plurality of conditions included in the determination model. The determination unit 123 predicts the displacement of the height by using the prediction formula when the specified condition is satisfied.
- the determination unit 123 sends information indicating the displacement of the predicted height to the output unit 124.
- Output unit 124 receives information representing the predicted height displacement from the determination unit 123.
- the output unit 124 outputs the received information representing the displacement of the height.
- the output destination of the output unit 124 is the same as the output unit of the output unit 124 of the first embodiment.
- FIG. 5 is a flowchart showing an example of the operation of the analyzer 10 of the present embodiment.
- steps S101 and S102 shown in FIG. 5 are the same as the operations of steps S101 and S102 of the analyzer 10 of the first embodiment shown in FIG.
- step S203 the learning unit 113 of the present embodiment generates the above-mentioned determination model for predicting the displacement of the height.
- the analyzer 10 of the present embodiment does not need to perform the operations after step S104 after step S101, step S102, and step S203.
- steps S104 and S105 are the same as the operations of steps S104 and S105 of the analyzer 10 of the first embodiment shown in FIG.
- step S206 the determination unit 123 predicts the height displacement at the target point by the determination model.
- step S207 the output unit 124 outputs the displacement at the predicted height.
- This embodiment has the effect of being able to predict fluctuations in the height of the ground surface.
- the reason is that the learning unit 113 learns a determination model that predicts the height displacement at the position of the target point based on at least a part of the values of the geospatial information at the position of the target point.
- the learning unit 113 may generate a plurality of determination models that individually predict the displacement of the height after each of the plurality of periods by performing heterogeneous mixed learning for each of the plurality of periods.
- the length of each of the plurality of periods may be, for example, a multiple of a predetermined length of a predetermined period.
- the length of each of the plurality of periods may be determined according to appropriately determined rules.
- the length of each of the plurality of periods may be specified by the user, for example.
- the learning unit 113 stores the generated plurality of determination models in the model storage unit 125.
- Each determination model may be configured to predict height displacement and output information representing the predicted height displacement and information representing the period.
- the determination unit 123 reads out a plurality of determination models stored in the model storage unit 125.
- the determination unit 123 predicts the displacement of the height after different periods have elapsed, respectively, according to the plurality of determined determination models read out.
- the determination unit 123 sends out the prediction of the displacement of the height after the lapse of different periods and each period to the output unit 124.
- the output unit 124 outputs a prediction of height displacement after each of a plurality of periods has elapsed.
- the output unit 124 may output information representing a plurality of periods and a prediction of height displacement after each period has elapsed.
- the learning unit 113 of this modification predicts the height displacement after a predetermined period elapses by the prediction formula for the case classification condition, and further contributes to the height displacement.
- the judgment model analyzes the geospatial information that contributes to the height displacement, and as described above, the geospatial information represented by the explanatory variables included in the prediction formula that predicts the height displacement, the height displacement. It is done by analyzing it as geospatial information that contributes to.
- the determination model outputs the information of the geospatial information analyzed as the geospatial information that contributes to the height displacement.
- the determination unit 123 predicts the height displacement and analyzes the geospatial information that contributes to the height displacement by the determination model.
- the determination unit 123 sends information representing the predicted height displacement and information representing the geospatial information contributing to the height displacement to the output unit 124.
- the output unit 124 outputs information representing the predicted height displacement and information representing the geospatial information that contributes to the height displacement.
- the second modification of the second embodiment can be configured like the first modification of the second embodiment.
- the learning unit 113 of the present modification may generate the same determination model as the determination model of the second modification of the second embodiment for each of the plurality of different periods. Specifically, the learning unit 113 generates a plurality of determination models for predicting the displacement of the height after the lapse of different periods and determining the cause of the displacement. As mentioned above, the factor is either geospatial information.
- the determination unit 123 predicts the displacement of the height after a plurality of different periods have elapsed at the target point and determines the factors contributing to the displacement of the height by the generated plurality of determination models.
- the determination unit 123 provides the output unit 124 with information indicating a predicted height displacement and information indicating a factor contributing to the height displacement after a plurality of different periods have elapsed at the target point. Send out.
- the output unit 124 outputs information representing the predicted displacement of the height and information representing the factors contributing to the displacement of the height.
- This modification is an example in which the second modification of the second embodiment is modified like the first modification of the first embodiment.
- the learning unit 113 of the present modification may generate a determination model for determining the magnitude of the contribution of the factor in addition to predicting the displacement of the height and determining the factor.
- the determination unit 123 of this modification determines the cause of the height displacement and the magnitude of the contribution of the factor.
- the determination unit 123 of this modification may send information indicating the displacement of the predicted height and information indicating the determined factor and the magnitude of the contribution of the factor to the output unit 124.
- the output unit 124 may output information representing the displacement of the predicted height and information representing the determined factor and the magnitude of the contribution of the factor.
- This modification is an example in which the third modification of the second embodiment is modified like the first modification of the first embodiment.
- the learning unit 113 of this modified example predicts the displacement of the height after a different period elapses and determines the factor of the displacement of the height, and in addition, the contribution of the factor is large. Generate multiple judgment models to judge the height.
- the determination unit 123 of the present modification uses a plurality of determination models to predict the displacement of the height after different periods have elapsed, and determine the factors of the displacement of the height and the magnitude of the contribution of the factors.
- the determination unit 123 may send information indicating the displacement of the predicted height and information indicating the determined factor and the magnitude of the contribution of the factor to the output unit 124.
- the output unit 124 outputs information representing the displacement of the predicted height and information representing the determined factor and the magnitude of the contribution of the factor.
- a sixth modification of the second embodiment is a combination of a plurality of devices as in the second modified example of the first embodiment. Can be achieved by.
- FIG. 6 is a diagram showing an example of the configuration of the analyzer 12 of the present embodiment.
- the analyzer 12 includes a first extraction unit 112 and a learning unit 113.
- the first extraction unit 112 extracts the geospatial information at each of the plurality of points on the ground surface from a plurality of types of geospatial information representing at least one of the state of the ground surface and the underground state of the ground surface. Extract the value of.
- the learning unit 113 determines the combination of the geospatial information that the determination model contributes to the displacement of the height based on at least a part of the values of the geospatial information. The determination model is learned based on the displacement and the extracted value of the geospatial information.
- the first extraction unit 112 and the learning unit 113 of the present embodiment function in the same manner as the first extraction unit 112 and the learning unit 113 of the first embodiment, respectively.
- FIG. 7 is a flowchart showing an example of the operation of the analyzer 12 of the present embodiment.
- the first extraction unit 112 extracts the value of the geospatial information at these plurality of points (step S102). Then, the learning unit 103 learns the determination model (step S103). The learning unit 103 of the present embodiment learns a determination model similar to the determination model of the first embodiment, similarly to the learning unit 103 of the first embodiment.
- FIG. 6 is a diagram showing an example of the configuration of the analyzer 12 of the present embodiment.
- the analyzer 12 includes a first extraction unit 112 and a learning unit 113.
- the first extraction unit 112 extracts the geospatial information at each of the plurality of points on the ground surface from a plurality of types of geospatial information representing at least one of the state of the ground surface and the underground state of the ground surface. Extract the value of.
- the learning unit 113 detects the height displacement at the plurality of points and the extracted value of the geospatial information so that the determination model predicts the height displacement based on at least a part of the values of the geospatial information. The determination model is learned based on the above.
- the first extraction unit 112 and the learning unit 113 of the present embodiment function in the same manner as the first extraction unit 112 and the learning unit 113 of the second embodiment, respectively.
- FIG. 7 is a flowchart showing an example of the operation of the analyzer 12 of the present embodiment.
- the first extraction unit 112 extracts the value of the geospatial information at these plurality of points (step S102). Then, the learning unit 103 learns the determination model (step S103). The learning unit 103 of the present embodiment learns a determination model similar to the determination model of the second embodiment, similarly to the learning unit 103 of the second embodiment.
- Each of the analyzer 10, the learning device 11, the analyzer 12, and the analyzer 21 according to the embodiment of the present disclosure includes a memory in which a program read from a storage medium is loaded, and a processor that executes the program. It can be realized by a computer.
- Each of the analyzer 10, the learning device 11, the analyzer 12, and the analyzer 21 according to the embodiment of the present disclosure can also be realized by dedicated hardware.
- Each of the analyzer 10, the learning device 11, the analyzer 12, and the analyzer 21 according to the embodiment of the present disclosure can also be realized by a combination of the above-mentioned computer and dedicated hardware.
- FIG. 8 is a diagram showing an example of a hardware configuration of a computer 1000 capable of realizing each of the analyzer 10, the learning device 11, the analyzer 12, and the analyzer 21 according to the embodiment of the present disclosure.
- the computer 1000 includes a processor 1001, a memory 1002, a storage device 1003, and an I / O (Input / Output) interface 1004.
- the computer 1000 can access the storage medium 1005.
- the memory 1002 and the storage device 1003 are storage devices such as a RAM (Random Access Memory) and a hard disk, for example.
- the storage medium 1005 is, for example, a storage device such as a RAM or a hard disk, a ROM (Read Only Memory), or a portable storage medium.
- the storage device 1003 may be a storage medium 1005.
- the processor 1001 can read and write data and programs to the memory 1002 and the storage device 1003.
- Processor 1001 can access other devices, for example, via the I / O interface 1004.
- Processor 1001 can access the storage medium 1005.
- the storage medium 1005 stores a program that causes the computer 1000 to operate as any one of the analyzer 10, the learning device 11, the analyzer 12, and the analyzer 21 according to the embodiment of the present disclosure.
- the processor 1001 is a program stored in the storage medium 1005 that causes the computer 1000 to operate as any one of the analyzer 10, the learning device 11, the analyzer 12, and the analyzer 21 according to the embodiment of the present disclosure. Load to. Then, when the processor 1001 executes the program loaded in the memory 1002, the computer 1000 uses the computer 1000 as the analyzer 10, the learning device 11, the analyzer 12, or the analyzer according to the embodiment of the present disclosure. Operates as 21.
- the first receiving unit 111, the first extracting unit 112, the learning unit 113, the first reading unit 114, and the transmitting unit 115 are, for example, by a processor 1001 that executes a program loaded into the memory 1002 from the storage medium 1005 that stores the program. It can be realized.
- the second receiving unit 121, the second extracting unit 122, the determining unit 123, the output unit 124, the second reading unit 126, and the receiving unit 127 execute, for example, a program loaded into the memory 1002 from the storage medium 1005 that stores the program. It can be realized by the processor 1001.
- the input / output unit 132 can be realized by, for example, a processor 1001 that executes a program loaded into the memory 1002 from the storage medium 1005 that stores the program.
- the model storage unit 125 and the geospatial information storage unit 131 can be realized by the memory 1002 included in the computer 1000 and the storage device 1003 such as a hard disk device.
- a part or all of the first receiving unit 111, the first extraction unit 112, the learning unit 113, the first reading unit 114, and the transmitting unit 115 can be realized by a dedicated circuit that realizes the functions of each unit.
- a part or all of the geospatial information storage unit 131 and the input / output unit 132 can also be realized by a dedicated circuit that realizes the functions of each unit.
- Appendix 1 A first method of extracting the value of the geospatial information at each of the plurality of points on the ground surface from a plurality of types of geospatial information representing at least one of the state of the ground surface and the underground state of the ground surface. Extraction means and The height displacement at a plurality of points is extracted so that the determination model determines the combination of the geospatial information that contributes to the height displacement based on at least a part of the values of the geospatial information. A learning means for learning the determination model based on the value of the geospatial information, and An analyzer equipped with.
- Appendix 2 A second extraction means for extracting the value of the geospatial information at the target point, and A determination means for determining a combination of geospatial information that contributes to the target displacement, which is a height displacement at the target point, based on the value of the geospatial information at the target point by the determination model.
- An output means that outputs a combination of the geospatial information that contributes to the target displacement, and The analyzer according to Appendix 1.
- the learning means learns the determination model so as to further determine the magnitude of each contribution of the combination of geospatial information that the determination model contributes to the height displacement.
- the determination means determines the magnitude of the contribution of geospatial information that contributes to the target displacement by the determination model.
- the analyzer according to Appendix 2, wherein the output means further outputs the determined magnitude of the contribution.
- the learning means extracts the height displacements at the plurality of points so that the determination model further predicts future height displacements based on at least some values of the geospatial information.
- the determination model is learned based on the value of the geospatial information, and the determination model is learned.
- the determination means further predicts the target displacement based on the value of the geospatial information at the target point by the determination model.
- the analyzer according to Appendix 2 or 3, wherein the output means outputs a prediction of the target displacement.
- the learning means derives a condition for each value of at least a part of the geospatial information and a prediction formula for predicting the displacement of the future height when all of the conditions are satisfied.
- the prediction formula is represented by a linear sum of variables representing the geospatial information.
- the determination model determines that the combination of the geospatial information represented by the variable in the prediction formula when all the conditions are satisfied is the combination of the geospatial information that contributes to the displacement of the height.
- the analyzer according to any one of the above items.
- Extraction means and The height displacement at the plurality of points and the extracted geospatial information value are used so that the determination model predicts the height displacement based on at least a part of the geodata information values.
- Learning means to learn the judgment model and An analyzer equipped with.
- Appendix 7 A second extraction means for extracting the value of the geospatial information at the target point, and A determination means for predicting a target displacement, which is a height displacement at the target point, based on the value of the geospatial information at the target point by the determination model. An output means for outputting the predicted target displacement and The analyzer according to Appendix 6.
- the value of the geospatial information at each of the plurality of points on the ground surface is extracted from a plurality of types of geospatial information representing at least one of the state of the ground surface and the underground state of the ground surface.
- the height displacement at a plurality of points is extracted so that the determination model determines the combination of the geospatial information that contributes to the height displacement based on at least a part of the values of the geospatial information.
- the determination model is learned based on the value of the geospatial information. Analytical method.
- the determination model is trained so that the determination model further determines the magnitude of each contribution of the combination of geospatial information that contributes to the height displacement. Using the determination model, the magnitude of the contribution of geospatial information that contributes to the target displacement is determined. Further, the analysis method according to Appendix 9, which outputs the determined magnitude of the contribution.
- a condition for each value of at least a part of the geospatial information and a prediction formula for predicting the displacement of the future height when all the conditions are satisfied are derived.
- the prediction formula is represented by a linear sum of variables representing the geospatial information.
- the determination model determines that the combination of the geospatial information represented by the variable in the prediction formula when all the conditions are satisfied is the combination of the geospatial information that contributes to the displacement of the height.
- the value of the geospatial information at each of the plurality of points on the ground surface is extracted from a plurality of types of geospatial information representing at least one of the state of the ground surface and the underground state of the ground surface.
- the height displacement at the plurality of points and the extracted geospatial information value are used so that the determination model predicts the height displacement based on at least a part of the geodata information values. Learn the decision model, Analytical method.
- Appendix 14 Extract the value of the geospatial information at the target point and The determination model predicts the target displacement, which is the displacement of the height at the target point, based on the value of the geospatial information at the target point. Output the predicted target displacement, The analysis method according to Appendix 13.
- Appendix 16 The program The second extraction process that extracts the value of the geospatial information at the target point, and A determination process for determining a combination of geospatial information that contributes to the target displacement, which is a height displacement at the target point, based on the value of the geospatial information at the target point by the determination model.
- Output processing that outputs a combination of the geospatial information that contributes to the target displacement, and The storage medium according to Appendix 15, which further causes a computer to execute the above.
- the learning process trains the determination model so as to further determine the magnitude of each contribution of the combination of geospatial information that the determination model contributes to the height displacement.
- the magnitude of the contribution of the geospatial information that contributes to the target displacement is determined by the determination model.
- the storage medium according to Appendix 16, wherein the output process further outputs the determined magnitude of the contribution.
- the learning process extracts the height displacements at the plurality of points so that the determination model further predicts future height displacements based on at least some values of the geospatial information.
- the determination model is learned based on the value of the geospatial information, and the determination model is learned.
- the determination process further predicts the target displacement based on the value of the geospatial information at the target point by the determination model.
- the learning process derives a condition for each value of at least a part of the geospatial information and a prediction formula for predicting the displacement of the future height when all of the conditions are satisfied.
- the prediction formula is represented by a linear sum of variables representing the geospatial information.
- the determination model determines that the combination of the geospatial information represented by the variable in the prediction formula when all the conditions are satisfied is the combination of the geospatial information that contributes to the displacement of the height.
- the storage medium according to any one of the above items.
- Appendix 20 A first method of extracting the value of the geospatial information at each of the plurality of points on the ground surface from a plurality of types of geospatial information representing at least one of the state of the ground surface and the underground state of the ground surface. Extraction process and The height displacement at the plurality of points and the extracted geospatial information value are used so that the determination model predicts the height displacement based on at least a part of the geodata information values. Learning process to learn the judgment model and A storage medium that stores a program that causes a computer to execute a program.
- Appendix 21 The program The second extraction process that extracts the value of the geospatial information at the target point, and The determination process for predicting the target displacement, which is the displacement of the height at the target point, based on the value of the geospatial information at the target point by the determination model. Output processing that outputs the predicted target displacement and 20.
- Analysis system 10 Analysis device 11 Learning device 12 Analysis device 21 Analysis device 31 Geospatial information storage device 40 Network 51 Terminal device 111 1st receiving unit 112 1st extraction unit 113 Learning unit 114 1st reading unit 115 Transmitting unit 121 2nd Receiving unit 122 Second extraction unit 123 Judging unit 124 Output unit 125 Model storage unit 126 Second reading unit 127 Receiving unit 131 Geospatial information storage unit 132 Input / output unit 1000 Computer 1001 Processor 1002 Memory 1003 Storage device 1004 I / O interface 1005 Storage medium
Landscapes
- Engineering & Computer Science (AREA)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
- Physics & Mathematics (AREA)
- Electromagnetism (AREA)
- Computer Networks & Wireless Communication (AREA)
- General Physics & Mathematics (AREA)
- Radar Systems Or Details Thereof (AREA)
- Testing Or Calibration Of Command Recording Devices (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
Description
<構成>
図1は、本開示の第1の実施形態に係る分析装置10の構成の例を表すブロック図である。図1に示す例では、分析装置10は、第1受取部111と、第1抽出部112と、学習部113と、第2受取部121と、第2抽出部122と、判定部123と、出力部124と、モデル記憶部125と、地理空間情報記憶部131とを含む。なお、分析装置10は、互いに通信可能に接続されている2つ以上の装置の組み合わせとして実現されていてもよい。また、ユーザがデータの分析装置10への入力などを行う端末装置が、例えば通信ネットワークを介して分析装置10に通信可能に接続されていてもよい。分析装置10が、互いに通信可能に接続されている3つの装置の組み合わせとして実現されている例は、変形例として後述される。
第1受取部111は、地表面の高さの変位を表すデータを、学習用のデータとして受け取る。例えば、ユーザが、上述の端末装置を使用して、地表面の高さの変位を表すデータを、第1受取部111に入力してもよい。この場合、第1受取部111は、地表面の高さの変位を表すデータを、その端末装置から受け取る。
第1抽出部112は、第1受取部111から学習用変位データを受け取る。第1抽出部112は、例えば、学習用変位データから、その学習用変位データに含まれる変位データによって表される高さの推移が観測された地点の地点情報を抽出する。第1抽出部112は、抽出した地点情報によって位置が表される地点における、地理空間情報の値を、後述の地理空間情報記憶部131に格納されている地理空間情報から抽出する。
地理空間情報記憶部131は、地理空間情報を記憶する。地理空間情報は、指定された地点における地表面の状態を特定できる形で、地理空間情報記憶部131に格納されている。
学習部113は、第1抽出部112から、学習用変位データと、その学習用変位データが経年変位を表す領域の、抽出された地理空間情報の値と、を受け取る。
(参考文献)” Fully-Automatic Bayesian Piecewise Sparse Linear Models”, Riki Eto, Ryohei Fujimaki, Satoshi Morinaga, Hiroshi Tamano, Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statistics, PMLR 33, pp. 238-246, 2014.
第2受取部121は、地表面上の地点の位置を特定する情報(例えば、緯度及び経度の情報)を受け取る。例えば、ユーザが、上述の端末装置を使用して、地表面上の地点の位置を特定する情報を、第2受取部121に入力してもよい。この場合、第2受取部121は、地表面上の地点の位置を特定する情報を、その端末装置から受け取る。
第2抽出部122は、第2受取部121から、対象地点情報を受け取る。第2抽出部122は、受け取った対象地点情報によって特定される対象地点における、地理空間情報の値を、地理空間情報記憶部131に格納されている地理空間情報から抽出する。第2抽出部122は、地理空間情報記憶部131に格納されている全ての地理空間情報のうち、あらかじめ定められている地理空間情報の値を抽出してもよい。この場合、例えば、地理空間情報の値に対する条件に関係が無く、高さの変位に寄与しないことがあらかじめ確かめられている地理空間情報を、値の抽出の対象から除外されていてよい。対象地点における値が設定されていない地理空間情報が存在する場合、第2抽出部122は、その地理空間情報の値を、値が存在しないことを表す数値(例えば0など)に設定してもよい。
判定部123は、対象地点における地理空間情報の値を、第2抽出部122から受け取る。判定部123は、対象地点情報を第2抽出部122から受け取ってもよい。
出力部124は、判定部123から、対象地点における高さの変位に寄与する地理空間情報の組み合わせの情報を受け取る。出力部124は、受け取った、対象地点における高さの変位に寄与する地理空間情報の組み合わせの情報を出力する。出力部124は、例えば、ディスプレイなどに、対象地点における高さの変位に寄与する地理空間情報の組み合わせを表示してもよい。出力部124は、対象地点における高さの変位に寄与する地理空間情報の組み合わせを、他の情報処理装置や上述の端末装置などに送出してもよい。
次に、第1の実施形態の分析装置10の動作について、図面を参照して詳細に説明する。
本実施形態には、地表面の高さの変動の要因を判定することができるという効果がある。その理由は、学習部113が、対象地点の位置における地理空間情報の少なくとも一部の値に基づいて、高さの変動の要因として、対象地点の位置における高さの変位に寄与する地理空間情報の組み合わせを判定する判定モデルを学習するからである。
次に、第1の実施形態の第1の変形例について説明する。本変形例の分析装置10の構成は、第1の実施形態の分析装置10の構成と同じである。本変形例の分析装置10の機能及び動作は、以下の相違点を除いて、第1の実施形態の分析装置10の機能及び動作と同じである。
本変形例の学習部113が生成する判定モデルは、高さの変位に寄与する地理空間情報の情報に加えて、地理空間情報の寄与の大きさを表す値を出力する。
判定部123は、モデル記憶部125に格納されている判定モデルに従って、対象地点における高さの変位に寄与する地理空間情報の組み合わせと、その組み合わせに含まれる地理空間情報の、高さの変位に対する寄与の大きさとを判定する。具体的には、判定部123は、例えば、判定モデルに含まれる複数の条件のうち、受け取った地理空間情報の値によって満たされる条件を特定する。判定部123は、特定した条件が満たされる場合の、高さの変位に寄与する地理空間情報の組み合わせが、対象地点における高さの変位に寄与する地理空間情報の組み合わせであると判定する。さらに、判定部123は、特定した条件が満たされる場合の高さの変位に寄与する地理空間情報の高さの変位に対する寄与の大きさが、それらの地理空間情報の、対象地点における高さの変位に対する寄与の大きさであると判定する。
出力部124は、判定部123から、対象地点における高さの変位に寄与する地理空間情報の組み合わせの情報と、組み合わせに含まれる地理空間情報の寄与の大きさを表す情報とを受け取る。出力部124は、受け取った、対象地点における高さの変位に寄与する地理空間情報の組み合わせの情報と、組み合わせに含まれる地理空間情報の寄与の大きさを表す情報とを出力する。出力部124は、例えば、ディスプレイなどに、対象地点における高さの変位に寄与する地理空間情報の組み合わせと寄与の大きさとを表示してもよい。出力部124は、対象地点における高さの変位に寄与する地理空間情報の組み合わせと寄与の大きさとを、他の情報処理装置や上述の端末装置などに送出してもよい。
図3は、第1の実施形態の変形例の分析システム1の構成を表すブロック図である。図3に示す例では、分析システム1は、学習装置11と、分析装置21と、地理空間情報記憶装置31、端末装置51とを含む。学習装置11、分析装置21、地理空間情報記憶装置31、及び、端末装置51は、通信ネットワークであるネットワーク40によって、通信可能に互いに接続されている。分析システム1は、第1の実施形態の分析装置10の機能を、学習装置11と、分析装置21と、地理空間情報記憶装置31とによって実現する。端末装置51は、上述の端末装置である。
図1は、本開示の第2の実施形態の分析装置10の構成を表す図である。本実施形態の分析装置10の構成は、第1の実施形態の分析装置10の構成と同じである。本実施形態の分析装置10の構成要素は、以下で説明する相違点を除いて、同一の名称及び符号が付与されている、第1の実施形態の分析装置10の構成要素と同じである。
本実施形態の学習部113は、第1の実施形態の学習部113が学習する判定モデルと異なる判定モデルを学習する。その他の点において、本実施形態の学習部113は、第1の実施形態の学習部113と同じである。例えば、第1の実施形態の学習部113と同様に、本実施形態の学習部113は、第1抽出部112から、学習用変位データと、その学習用変位データが経年変位を表す領域の、抽出された地理空間情報の値と、を受け取る。第1の実施形態の学習部113と同様に、本実施形態の学習部113は、学習によって得られた判定モデルを、モデル記憶部125に格納する。
本実施形態の判定部123は、第1の実施形態の判定部123と同様に、対象地点における地理空間情報の値を、第2抽出部122から受け取る。判定部123は、対象地点情報を第2抽出部122から受け取ってもよい。
出力部124は、予測された高さの変位を表す情報を判定部123から受け取る。出力部124は、受け取った、高さの変位を表す情報を出力する。出力部124の出力先は、第1の実施形態の出力部124の出力部と同様である。
次に、本実施形態の分析装置10の動作について説明する。
本実施形態には、地表面の高さの変動を予測することができるという効果がある。その理由は、学習部113が、対象地点の位置における地理空間情報の少なくとも一部の値に基づいて、対象地点の位置における高さの変位の予測を行う判定モデルを学習するからである。
次に、第2の実施形態の第1の変形例について説明する。本変形例の分析装置10の構成は、図1に示す、第2の実施形態の分析装置10の構成と同じである。
次に、第2の実施形態の第2の変形例について説明する。本変形例の分析装置10の構成は、図1に示す、第2の実施形態の分析装置10の構成と同じである。
次に、第2の実施形態の第3の変形例について説明する。本変形例の分析装置10の構成は、図1に示す、第2の実施形態の分析装置10の構成と同じである。
次に、第2の実施形態の第4の変形例について説明する。本変形例の分析装置10の構成は、図1に示す、第2の実施形態の分析装置10の構成と同じである。
次に、第2の実施形態の第5の変形例について説明する。本変形例の分析装置10の構成は、図1に示す、第2の実施形態の分析装置10の構成と同じである。
第2の実施形態、及び、第2の実施形態の第1から第5の変形例の分析装置10の機能は、第1の実施形態の第2の変形例のような、複数の装置の組み合わせによって実現できる。
次に、本開示の第3の実施形態について、図面を参照して詳細に説明する。
図6は、本実施形態の分析装置12の構成の例を表す図である。
図7は、本実施形態の分析装置12の動作の例を表すフローチャートである。
本実施形態には、第1の実施形態と同じ効果がある。その理由は、第1の実施形態の効果が生じる理由と同じである。
次に、本開示の第4の実施形態について、図面を参照して詳細に説明する。
図6は、本実施形態の分析装置12の構成の例を表す図である。
図7は、本実施形態の分析装置12の動作の例を表すフローチャートである。
本実施形態には、第2の実施形態と同じ効果がある。その理由は、第1の実施形態の効果が生じる理由と同じである。
本開示の実施形態に係る分析装置10、学習装置11、分析装置12、分析装置21の各々は、記憶媒体から読み出されたプログラムがロードされたメモリと、そのプログラムを実行するプロセッサとを含むコンピュータによって実現することができる。本開示の実施形態に係る分析装置10、学習装置11、分析装置12、分析装置21の各々は、専用のハードウェアによって実現することもできる。本開示の実施形態に係る分析装置10、学習装置11、分析装置12、分析装置21の各々は、前述のコンピュータと専用のハードウェアとの組み合わせによって実現することもできる。
地表面の状態及び当該地表面の地下の状態の少なくともいずれかをそれぞれ表す、複数の種類の地理空間情報から、前記地表面の複数の地点の各々における前記地理空間情報の値を抽出する第1抽出手段と、
判定モデルが高さの変位に対して寄与する前記地理空間情報の組み合わせを前記地理空間情報の少なくとも一部の値に基づいて判定するように、前記複数の地点における高さの変位と抽出した前記前記地理空間情報の値とに基づいて前記判定モデルを学習する学習手段と、
を備える分析装置。
対象地点における前記地理空間情報の値を抽出する第2抽出手段と、
前記判定モデルによって、前記対象地点における前記地理空間情報の値に基づく、前記対象地点における高さの変位である対象変位に寄与する地理空間情報の組み合わせを判定する判定手段と、
前記対象変位に寄与する前記地理空間情報の組み合わせを出力する出力手段と、
を備える付記1に記載の分析装置。
前記学習手段は、前記判定モデルが高さの変位に対して寄与する前記地理空間情報の組み合わせの各々の寄与の大きさをさらに判定するように、前記判定モデルを学習し、
前記判定手段は、前記判定モデルによって、前記対象変位に寄与する地理空間情報の寄与の大きさを判定し、
前記出力手段は、さらに、判定された前記寄与の大きさを出力する
付記2に記載の分析装置。
前記学習手段は、前記判定モデルが、前記地理空間情報の少なくとも一部の値に基づいて将来の高さの変位の予測をさらに行うように、前記複数の地点における前記高さの変位と抽出した前記前記地理空間情報の値とに基づいて前記判定モデルを学習し、
前記判定手段は、前記判定モデルによって、前記対象地点における前記前記地理空間情報の値に基づく、前記対象変位の予測をさらに行い、
前記出力手段は、前記対象変位の予測を出力する
付記2または3に記載の分析装置。
前記学習手段は、前記地理空間情報の少なくとも一部の各々の値に対する条件と、当該条件の全てが満たされる場合において前記将来の高さの変位の予測を行う予測式とを導出し、
前記予測式は、前記地理空間情報をそれぞれ表す変数の線形和によって表され、
前記判定モデルは、前記条件がすべて満たされる場合における前記予測式の、前記変数が表す前記地理空間情報の組み合わせを、高さの変位に寄与する前記地理空間情報の組み合わせと判定する
付記1乃至4のいずれか1項に記載の分析装置。
地表面の状態及び当該地表面の地下の状態の少なくともいずれかをそれぞれ表す、複数の種類の地理空間情報から、前記地表面の複数の地点の各々における前記地理空間情報の値を抽出する第1抽出手段と、
判定モデルが高さの変位を前記地理空間情報の少なくとも一部の値に基づいて予測するように、前記複数の地点における高さの変位と抽出した前記前記地理空間情報の値とに基づいて前記判定モデルを学習する学習手段と、
を備える分析装置。
対象地点における前記地理空間情報の値を抽出する第2抽出手段と、
前記判定モデルによって、前記対象地点における前記地理空間情報の値に基づく、前記対象地点における高さの変位である対象変位を予測する判定手段と、
予測された前記対象変位を出力する出力手段と、
を備える付記6に記載の分析装置。
地表面の状態及び当該地表面の地下の状態の少なくともいずれかをそれぞれ表す、複数の種類の地理空間情報から、前記地表面の複数の地点の各々における前記地理空間情報の値を抽出し、
判定モデルが高さの変位に対して寄与する前記地理空間情報の組み合わせを前記地理空間情報の少なくとも一部の値に基づいて判定するように、前記複数の地点における高さの変位と抽出した前記前記地理空間情報の値とに基づいて前記判定モデルを学習する、
分析方法。
対象地点における前記地理空間情報の値を抽出し、
前記判定モデルによって、前記対象地点における前記地理空間情報の値に基づく、前記対象地点における高さの変位である対象変位に寄与する地理空間情報の組み合わせを判定し、
前記対象変位に寄与する前記地理空間情報の組み合わせを出力する、
付記8に記載の分析方法。
前記判定モデルが高さの変位に対して寄与する前記地理空間情報の組み合わせの各々の寄与の大きさをさらに判定するように、前記判定モデルを学習し、
前記判定モデルによって、前記対象変位に寄与する地理空間情報の寄与の大きさを判定し、
さらに、判定された前記寄与の大きさを出力する
付記9に記載の分析方法。
前記判定モデルが、前記地理空間情報の少なくとも一部の値に基づいて将来の高さの変位の予測をさらに行うように、前記複数の地点における前記高さの変位と抽出した前記前記地理空間情報の値とに基づいて前記判定モデルを学習し、
前記判定モデルによって、前記対象地点における前記前記地理空間情報の値に基づく、前記対象変位の予測をさらに行い、
前記対象変位の予測を出力する
付記9または10に記載の分析方法。
前記地理空間情報の少なくとも一部の各々の値に対する条件と、当該条件の全てが満たされる場合において前記将来の高さの変位の予測を行う予測式とを導出し、
前記予測式は、前記地理空間情報をそれぞれ表す変数の線形和によって表され、
前記判定モデルは、前記条件がすべて満たされる場合における前記予測式の、前記変数が表す前記地理空間情報の組み合わせを、高さの変位に寄与する前記地理空間情報の組み合わせと判定する
付記8乃至11のいずれか1項に記載の分析方法。
地表面の状態及び当該地表面の地下の状態の少なくともいずれかをそれぞれ表す、複数の種類の地理空間情報から、前記地表面の複数の地点の各々における前記地理空間情報の値を抽出し、
判定モデルが高さの変位を前記地理空間情報の少なくとも一部の値に基づいて予測するように、前記複数の地点における高さの変位と抽出した前記前記地理空間情報の値とに基づいて前記判定モデルを学習する、
分析方法。
対象地点における前記地理空間情報の値を抽出し、
前記判定モデルによって、前記対象地点における前記地理空間情報の値に基づく、前記対象地点における高さの変位である対象変位を予測し、
予測された前記対象変位を出力する、
付記13に記載の分析方法。
地表面の状態及び当該地表面の地下の状態の少なくともいずれかをそれぞれ表す、複数の種類の地理空間情報から、前記地表面の複数の地点の各々における前記地理空間情報の値を抽出する第1抽出処理と、
判定モデルが高さの変位に対して寄与する前記地理空間情報の組み合わせを前記地理空間情報の少なくとも一部の値に基づいて判定するように、前記複数の地点における高さの変位と抽出した前記前記地理空間情報の値とに基づいて前記判定モデルを学習する学習処理と、
をコンピュータに実行させるプログラムを記憶する記憶媒体。
前記プログラムは、
対象地点における前記地理空間情報の値を抽出する第2抽出処理と、
前記判定モデルによって、前記対象地点における前記地理空間情報の値に基づく、前記対象地点における高さの変位である対象変位に寄与する地理空間情報の組み合わせを判定する判定処理と、
前記対象変位に寄与する前記地理空間情報の組み合わせを出力する出力処理と、
をさらにコンピュータに実行させる付記15に記載の記憶媒体。
前記学習処理は、前記判定モデルが高さの変位に対して寄与する前記地理空間情報の組み合わせの各々の寄与の大きさをさらに判定するように、前記判定モデルを学習し、
前記判定処理は、前記判定モデルによって、前記対象変位に寄与する地理空間情報の寄与の大きさを判定し、
前記出力処理は、さらに、判定された前記寄与の大きさを出力する
付記16に記載の記憶媒体。
前記学習処理は、前記判定モデルが、前記地理空間情報の少なくとも一部の値に基づいて将来の高さの変位の予測をさらに行うように、前記複数の地点における前記高さの変位と抽出した前記前記地理空間情報の値とに基づいて前記判定モデルを学習し、
前記判定処理は、前記判定モデルによって、前記対象地点における前記前記地理空間情報の値に基づく、前記対象変位の予測をさらに行い、
前記出力処理は、前記対象変位の予測を出力する
付記16または17に記載の記憶媒体。
前記学習処理は、前記地理空間情報の少なくとも一部の各々の値に対する条件と、当該条件の全てが満たされる場合において前記将来の高さの変位の予測を行う予測式とを導出し、
前記予測式は、前記地理空間情報をそれぞれ表す変数の線形和によって表され、
前記判定モデルは、前記条件がすべて満たされる場合における前記予測式の、前記変数が表す前記地理空間情報の組み合わせを、高さの変位に寄与する前記地理空間情報の組み合わせと判定する
付記15乃至18のいずれか1項に記載の記憶媒体。
地表面の状態及び当該地表面の地下の状態の少なくともいずれかをそれぞれ表す、複数の種類の地理空間情報から、前記地表面の複数の地点の各々における前記地理空間情報の値を抽出する第1抽出処理と、
判定モデルが高さの変位を前記地理空間情報の少なくとも一部の値に基づいて予測するように、前記複数の地点における高さの変位と抽出した前記前記地理空間情報の値とに基づいて前記判定モデルを学習する学習処理と、
をコンピュータに実行させるプログラムを記憶する記憶媒体。
前記プログラムは、
対象地点における前記地理空間情報の値を抽出する第2抽出処理と、
前記判定モデルによって、前記対象地点における前記地理空間情報の値に基づく、前記対象地点における高さの変位である対象変位を予測する判定処理と、
予測された前記対象変位を出力する出力処理と、
をコンピュータに実行させる付記20に記載の記憶媒体。
10 分析装置
11 学習装置
12 分析装置
21 分析装置
31 地理空間情報記憶装置
40 ネットワーク
51 端末装置
111 第1受取部
112 第1抽出部
113 学習部
114 第1読出部
115 送信部
121 第2受取部
122 第2抽出部
123 判定部
124 出力部
125 モデル記憶部
126 第2読出部
127 受信部
131 地理空間情報記憶部
132 入出力部
1000 コンピュータ
1001 プロセッサ
1002 メモリ
1003 記憶装置
1004 I/Oインタフェース
1005 記憶媒体
Claims (21)
- 地表面の状態及び当該地表面の地下の状態の少なくともいずれかをそれぞれ表す、複数の種類の地理空間情報から、前記地表面の複数の地点の各々における前記地理空間情報の値を抽出する第1抽出手段と、
判定モデルが高さの変位に対して寄与する前記地理空間情報の組み合わせを前記地理空間情報の少なくとも一部の値に基づいて判定するように、前記複数の地点における高さの変位と抽出した前記前記地理空間情報の値とに基づいて前記判定モデルを学習する学習手段と、
を備える分析装置。 - 対象地点における前記地理空間情報の値を抽出する第2抽出手段と、
前記判定モデルによって、前記対象地点における前記地理空間情報の値に基づく、前記対象地点における高さの変位である対象変位に寄与する地理空間情報の組み合わせを判定する判定手段と、
前記対象変位に寄与する前記地理空間情報の組み合わせを出力する出力手段と、
を備える請求項1に記載の分析装置。 - 前記学習手段は、前記判定モデルが高さの変位に対して寄与する前記地理空間情報の組み合わせの各々の寄与の大きさをさらに判定するように、前記判定モデルを学習し、
前記判定手段は、前記判定モデルによって、前記対象変位に寄与する地理空間情報の寄与の大きさを判定し、
前記出力手段は、さらに、判定された前記寄与の大きさを出力する
請求項2に記載の分析装置。 - 前記学習手段は、前記判定モデルが、前記地理空間情報の少なくとも一部の値に基づいて将来の高さの変位の予測をさらに行うように、前記複数の地点における前記高さの変位と抽出した前記前記地理空間情報の値とに基づいて前記判定モデルを学習し、
前記判定手段は、前記判定モデルによって、前記対象地点における前記前記地理空間情報の値に基づく、前記対象変位の予測をさらに行い、
前記出力手段は、前記対象変位の予測を出力する
請求項2または3に記載の分析装置。 - 前記学習手段は、前記地理空間情報の少なくとも一部の各々の値に対する条件と、当該条件の全てが満たされる場合において前記将来の高さの変位の予測を行う予測式とを導出し、
前記予測式は、前記地理空間情報をそれぞれ表す変数の線形和によって表され、
前記判定モデルは、前記条件がすべて満たされる場合における前記予測式の、前記変数が表す前記地理空間情報の組み合わせを、高さの変位に寄与する前記地理空間情報の組み合わせと判定する
請求項1乃至4のいずれか1項に記載の分析装置。 - 地表面の状態及び当該地表面の地下の状態の少なくともいずれかをそれぞれ表す、複数の種類の地理空間情報から、前記地表面の複数の地点の各々における前記地理空間情報の値を抽出する第1抽出手段と、
判定モデルが高さの変位を前記地理空間情報の少なくとも一部の値に基づいて予測するように、前記複数の地点における高さの変位と抽出した前記前記地理空間情報の値とに基づいて前記判定モデルを学習する学習手段と、
を備える分析装置。 - 対象地点における前記地理空間情報の値を抽出する第2抽出手段と、
前記判定モデルによって、前記対象地点における前記地理空間情報の値に基づく、前記対象地点における高さの変位である対象変位を予測する判定手段と、
予測された前記対象変位を出力する出力手段と、
を備える請求項6に記載の分析装置。 - 地表面の状態及び当該地表面の地下の状態の少なくともいずれかをそれぞれ表す、複数の種類の地理空間情報から、前記地表面の複数の地点の各々における前記地理空間情報の値を抽出し、
判定モデルが高さの変位に対して寄与する前記地理空間情報の組み合わせを前記地理空間情報の少なくとも一部の値に基づいて判定するように、前記複数の地点における高さの変位と抽出した前記前記地理空間情報の値とに基づいて前記判定モデルを学習する、
分析方法。 - 対象地点における前記地理空間情報の値を抽出し、
前記判定モデルによって、前記対象地点における前記地理空間情報の値に基づく、前記対象地点における高さの変位である対象変位に寄与する地理空間情報の組み合わせを判定し、
前記対象変位に寄与する前記地理空間情報の組み合わせを出力する、
請求項8に記載の分析方法。 - 前記判定モデルが高さの変位に対して寄与する前記地理空間情報の組み合わせの各々の寄与の大きさをさらに判定するように、前記判定モデルを学習し、
前記判定モデルによって、前記対象変位に寄与する地理空間情報の寄与の大きさを判定し、
さらに、判定された前記寄与の大きさを出力する
請求項9に記載の分析方法。 - 前記判定モデルが、前記地理空間情報の少なくとも一部の値に基づいて将来の高さの変位の予測をさらに行うように、前記複数の地点における前記高さの変位と抽出した前記前記地理空間情報の値とに基づいて前記判定モデルを学習し、
前記判定モデルによって、前記対象地点における前記前記地理空間情報の値に基づく、前記対象変位の予測をさらに行い、
前記対象変位の予測を出力する
請求項9または10に記載の分析方法。 - 前記地理空間情報の少なくとも一部の各々の値に対する条件と、当該条件の全てが満たされる場合において前記将来の高さの変位の予測を行う予測式とを導出し、
前記予測式は、前記地理空間情報をそれぞれ表す変数の線形和によって表され、
前記判定モデルは、前記条件がすべて満たされる場合における前記予測式の、前記変数が表す前記地理空間情報の組み合わせを、高さの変位に寄与する前記地理空間情報の組み合わせと判定する
請求項8乃至11のいずれか1項に記載の分析方法。 - 地表面の状態及び当該地表面の地下の状態の少なくともいずれかをそれぞれ表す、複数の種類の地理空間情報から、前記地表面の複数の地点の各々における前記地理空間情報の値を抽出し、
判定モデルが高さの変位を前記地理空間情報の少なくとも一部の値に基づいて予測するように、前記複数の地点における高さの変位と抽出した前記前記地理空間情報の値とに基づいて前記判定モデルを学習する、
分析方法。 - 対象地点における前記地理空間情報の値を抽出し、
前記判定モデルによって、前記対象地点における前記地理空間情報の値に基づく、前記対象地点における高さの変位である対象変位を予測し、
予測された前記対象変位を出力する、
請求項13に記載の分析方法。 - 地表面の状態及び当該地表面の地下の状態の少なくともいずれかをそれぞれ表す、複数の種類の地理空間情報から、前記地表面の複数の地点の各々における前記地理空間情報の値を抽出する第1抽出処理と、
判定モデルが高さの変位に対して寄与する前記地理空間情報の組み合わせを前記地理空間情報の少なくとも一部の値に基づいて判定するように、前記複数の地点における高さの変位と抽出した前記前記地理空間情報の値とに基づいて前記判定モデルを学習する学習処理と、
をコンピュータに実行させるプログラムを記憶する記憶媒体。 - 前記プログラムは、
対象地点における前記地理空間情報の値を抽出する第2抽出処理と、
前記判定モデルによって、前記対象地点における前記地理空間情報の値に基づく、前記対象地点における高さの変位である対象変位に寄与する地理空間情報の組み合わせを判定する判定処理と、
前記対象変位に寄与する前記地理空間情報の組み合わせを出力する出力処理と、
をさらにコンピュータに実行させる請求項15に記載の記憶媒体。 - 前記学習処理は、前記判定モデルが高さの変位に対して寄与する前記地理空間情報の組み合わせの各々の寄与の大きさをさらに判定するように、前記判定モデルを学習し、
前記判定処理は、前記判定モデルによって、前記対象変位に寄与する地理空間情報の寄与の大きさを判定し、
前記出力処理は、さらに、判定された前記寄与の大きさを出力する
請求項16に記載の記憶媒体。 - 前記学習処理は、前記判定モデルが、前記地理空間情報の少なくとも一部の値に基づいて将来の高さの変位の予測をさらに行うように、前記複数の地点における前記高さの変位と抽出した前記前記地理空間情報の値とに基づいて前記判定モデルを学習し、
前記判定処理は、前記判定モデルによって、前記対象地点における前記前記地理空間情報の値に基づく、前記対象変位の予測をさらに行い、
前記出力処理は、前記対象変位の予測を出力する
請求項16または17に記載の記憶媒体。 - 前記学習処理は、前記地理空間情報の少なくとも一部の各々の値に対する条件と、当該条件の全てが満たされる場合において前記将来の高さの変位の予測を行う予測式とを導出し、
前記予測式は、前記地理空間情報をそれぞれ表す変数の線形和によって表され、
前記判定モデルは、前記条件がすべて満たされる場合における前記予測式の、前記変数が表す前記地理空間情報の組み合わせを、高さの変位に寄与する前記地理空間情報の組み合わせと判定する
請求項15乃至18のいずれか1項に記載の記憶媒体。 - 地表面の状態及び当該地表面の地下の状態の少なくともいずれかをそれぞれ表す、複数の種類の地理空間情報から、前記地表面の複数の地点の各々における前記地理空間情報の値を抽出する第1抽出処理と、
判定モデルが高さの変位を前記地理空間情報の少なくとも一部の値に基づいて予測するように、前記複数の地点における高さの変位と抽出した前記前記地理空間情報の値とに基づいて前記判定モデルを学習する学習処理と、
をコンピュータに実行させるプログラムを記憶する記憶媒体。 - 前記プログラムは、
対象地点における前記地理空間情報の値を抽出する第2抽出処理と、
前記判定モデルによって、前記対象地点における前記地理空間情報の値に基づく、前記対象地点における高さの変位である対象変位を予測する判定処理と、
予測された前記対象変位を出力する出力処理と、
をコンピュータに実行させる請求項20に記載の記憶媒体。
Priority Applications (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US17/788,007 US12360235B2 (en) | 2020-03-31 | 2020-03-31 | Analysis device, analysis method, and storage medium |
| PCT/JP2020/014752 WO2021199245A1 (ja) | 2020-03-31 | 2020-03-31 | 分析装置、分析方法及び記憶媒体 |
| JP2022512969A JP7375915B2 (ja) | 2020-03-31 | 2020-03-31 | 分析装置、分析方法及びプログラム |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/JP2020/014752 WO2021199245A1 (ja) | 2020-03-31 | 2020-03-31 | 分析装置、分析方法及び記憶媒体 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2021199245A1 true WO2021199245A1 (ja) | 2021-10-07 |
Family
ID=77928213
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/JP2020/014752 Ceased WO2021199245A1 (ja) | 2020-03-31 | 2020-03-31 | 分析装置、分析方法及び記憶媒体 |
Country Status (3)
| Country | Link |
|---|---|
| US (1) | US12360235B2 (ja) |
| JP (1) | JP7375915B2 (ja) |
| WO (1) | WO2021199245A1 (ja) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN119902167A (zh) * | 2025-02-11 | 2025-04-29 | 西安电子科技大学 | 基于vmd和稀疏贝叶斯盲分离的雷达抗间歇采样干扰方法 |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP7703201B2 (ja) * | 2020-07-06 | 2025-07-07 | 国立大学法人東北大学 | 液状化評価モデル生成装置、液状化評価装置、液状化評価モデル生成プログラム及び液状化評価プログラム |
| JP7616228B2 (ja) * | 2020-09-10 | 2025-01-17 | 日本電気株式会社 | 推定装置、推定方法及びプログラム |
| JP7639408B2 (ja) * | 2021-03-03 | 2025-03-05 | 富士通株式会社 | 説明情報出力プログラム,説明情報出力方法および説明情報出力装置 |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2009087235A (ja) * | 2007-10-02 | 2009-04-23 | Toshiba Corp | モデル構築装置並びにモデル構築のための方法およびプログラム |
| WO2015041295A1 (ja) * | 2013-09-18 | 2015-03-26 | 国立大学法人東京大学 | 地表種別分類方法および地表種別分類プログラム並びに地表種別分類装置 |
| US20160321818A1 (en) * | 2015-04-30 | 2016-11-03 | Harris Corporation | Geospatial imaging system providing segmentation and classification features and related methods |
| JP2018048898A (ja) * | 2016-09-21 | 2018-03-29 | 日本電気株式会社 | 画像処理装置、画像処理方法及びプログラム |
| JP2018194404A (ja) * | 2017-05-16 | 2018-12-06 | 株式会社パスコ | 機械学習方法、及び地表変動判定方法 |
| JP2019035598A (ja) * | 2017-08-10 | 2019-03-07 | 株式会社パスコ | 土地被覆学習データ生成装置、土地被覆学習データ、土地被覆分類予測装置及び土地被覆学習データ生成プログラム |
| CN109669184A (zh) * | 2019-02-25 | 2019-04-23 | 电子科技大学 | 一种基于全卷积网络的合成孔径雷达方位模糊消除方法 |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7298891B2 (en) * | 2004-07-15 | 2007-11-20 | Harris Corporation | Bare earth digital elevation model extraction for three-dimensional registration from topographical points |
| US7142984B2 (en) * | 2005-02-08 | 2006-11-28 | Harris Corporation | Method and apparatus for enhancing a digital elevation model (DEM) for topographical modeling |
| US8244026B2 (en) * | 2008-01-09 | 2012-08-14 | Tiltan Systems Engineering Ltd. | Apparatus and method for automatic airborne LiDAR data processing and mapping using data obtained thereby |
| GB2559566B (en) * | 2017-02-08 | 2022-01-12 | Ordnance Survey Ltd | Topographic data machine learning method and system |
| US20200166626A1 (en) * | 2017-05-17 | 2020-05-28 | Nec Corporation | Information processing device, information processing method, and storage medium having program stored thereon |
| JP7059466B2 (ja) | 2017-11-02 | 2022-04-26 | 国際航業株式会社 | 予測座標変動算出システム |
| JP7244794B2 (ja) | 2018-08-03 | 2023-03-23 | 日本電気株式会社 | Sar画像解析システム |
| WO2022147534A1 (en) * | 2021-01-04 | 2022-07-07 | Realmfive, Inc. | Preventative failure for agricultural applications |
| US12260464B2 (en) * | 2021-09-27 | 2025-03-25 | Hexcuity Limited | System and method for automated forest inventory mapping |
-
2020
- 2020-03-31 JP JP2022512969A patent/JP7375915B2/ja active Active
- 2020-03-31 US US17/788,007 patent/US12360235B2/en active Active
- 2020-03-31 WO PCT/JP2020/014752 patent/WO2021199245A1/ja not_active Ceased
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2009087235A (ja) * | 2007-10-02 | 2009-04-23 | Toshiba Corp | モデル構築装置並びにモデル構築のための方法およびプログラム |
| WO2015041295A1 (ja) * | 2013-09-18 | 2015-03-26 | 国立大学法人東京大学 | 地表種別分類方法および地表種別分類プログラム並びに地表種別分類装置 |
| US20160321818A1 (en) * | 2015-04-30 | 2016-11-03 | Harris Corporation | Geospatial imaging system providing segmentation and classification features and related methods |
| JP2018048898A (ja) * | 2016-09-21 | 2018-03-29 | 日本電気株式会社 | 画像処理装置、画像処理方法及びプログラム |
| JP2018194404A (ja) * | 2017-05-16 | 2018-12-06 | 株式会社パスコ | 機械学習方法、及び地表変動判定方法 |
| JP2019035598A (ja) * | 2017-08-10 | 2019-03-07 | 株式会社パスコ | 土地被覆学習データ生成装置、土地被覆学習データ、土地被覆分類予測装置及び土地被覆学習データ生成プログラム |
| CN109669184A (zh) * | 2019-02-25 | 2019-04-23 | 电子科技大学 | 一种基于全卷积网络的合成孔径雷达方位模糊消除方法 |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN119902167A (zh) * | 2025-02-11 | 2025-04-29 | 西安电子科技大学 | 基于vmd和稀疏贝叶斯盲分离的雷达抗间歇采样干扰方法 |
Also Published As
| Publication number | Publication date |
|---|---|
| JP7375915B2 (ja) | 2023-11-08 |
| US12360235B2 (en) | 2025-07-15 |
| JPWO2021199245A1 (ja) | 2021-10-07 |
| US20230042178A1 (en) | 2023-02-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP7375915B2 (ja) | 分析装置、分析方法及びプログラム | |
| Goldstein et al. | Approaches to simulating the “March of Bricks and Mortar” | |
| Thiery et al. | Application of expert rules in indirect approaches for landslide susceptibility assessment | |
| Marques et al. | Sea cliff instability susceptibility at regional scale: a statistically based assessment in the southern Algarve, Portugal | |
| US20240153264A1 (en) | Inundation and overflow prediction system using drone images and artificial intelligence | |
| Shastry et al. | Using steady‐state backwater analysis to predict inundated area from national water model streamflow simulations | |
| CN115147011B (zh) | 一种地表水与林地损害鉴定评估方法及装置、设备和介质 | |
| CN115907357A (zh) | 基于遥感和边坡失稳模型的输电杆塔选址优化系统及方法 | |
| Gyasi-Agyei et al. | Interpolation of daily rainfall networks using simulated radar fields for realistic hydrological modelling of spatial rain field ensembles | |
| JP7294529B2 (ja) | 分析装置、分析方法及びプログラム | |
| Dehingia et al. | Decadal transformation of land use-land cover and future spatial expansion in Bangalore metropolitan region, India: Open-source geospatial machine learning approach | |
| Bhardwaj et al. | Landslide susceptibility mapping using probabilistic frequency ratio and Shannon entropy for Chamoli, Uttarakhand Himalayas | |
| Bhattarai et al. | Rapid prediction of urban flooding at street-scale using physics-informed machine learning-based surrogate modeling | |
| JP7380847B2 (ja) | 分析装置、分析方法及びプログラム | |
| Marques et al. | Statistically based sea cliff instability hazard assessment of Burgau-Lagos coastal section (Algarve, Portugal) | |
| Shannon et al. | LiDAR-based sinkhole detection and mapping in Knox County, Tennessee | |
| Williams et al. | Geographic information system data considerations in the context of the enhanced bathtub model for coastal inundation | |
| CN119692532A (zh) | 一种基于城市内涝风险评估模型的绿地优化方法及相关装置 | |
| CN115600047B (zh) | 一种基于栅格分析的小流域面平均降雨量测算方法和系统 | |
| Lin et al. | Evaluation of a geospatial liquefaction model using land damage data from the 2016 Kaikoura earthquake | |
| Kim et al. | Urban flood inundation simulation based on high-precision 3D modeling | |
| JP7499137B2 (ja) | 地震時諸現象統合評価モデル生成方法、地震時諸現象統合評価モデル生成装置、地震時諸現象統合評価方法、及び、地震時諸現象統合評価装置 | |
| Romah | Advanced methods in sea level rise vulnerability assessment | |
| Cortez et al. | Modeling and spatio-temporal prediction rainfall data applied to mass movements in Belo Horizonte, Minas Gerais-Brazil | |
| JP7616228B2 (ja) | 推定装置、推定方法及びプログラム |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 20928280 Country of ref document: EP Kind code of ref document: A1 |
|
| ENP | Entry into the national phase |
Ref document number: 2022512969 Country of ref document: JP Kind code of ref document: A |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 20928280 Country of ref document: EP Kind code of ref document: A1 |
|
| WWG | Wipo information: grant in national office |
Ref document number: 17788007 Country of ref document: US |