Anomaly detection device, method, and program

Abstract

By analyzing crustal deformation data, we can clearly identify precursory anomalies such as pre-slips that precede major earthquakes, and use this data to predict earthquakes and estimate the optimal epicenter location from among potential epicenters. [Solution] The crustal deformation vector calculation unit calculates the difference vector of the change in position information at each observation station at a fixed interval such as 5 minutes, and calculates this as the crustal deformation vector for each observation station. The correlation value calculation unit performs processing based on the crustal deformation vectors for each observation station output by the crustal deformation vector calculation unit. The correlation value calculation unit selects one station as the central observation station and selects several observation stations in its vicinity as surrounding observation stations. In a certain situation, the correlation value calculation unit can select any number of surrounding observation stations from among observation stations whose distance from the central observation station is within a predetermined distance.

Description

Technical Field

[0001] The present invention relates to an abnormality detection device, and more particularly to a technique for detecting abnormalities in crustal movements.

Background Art

[0002] Before a major earthquake occurs, many studies have been conducted to capture precursors just before the earthquake. In particular, there are many studies attempting to capture the immediate precursor slip (press slip) among earthquake precursor phenomena, but studies using existing Global Navigation Satellite System (GNSS) are based on relatively recent papers (Non-Patent Document 1). Regarding the study of capturing abnormalities in the ionosphere just before an earthquake, there is the correlation analysis method proposed by the present inventor (Patent Document 1), which discloses a method of capturing changes in the electron density of the ionosphere at multiple points and improving the signal-to-noise ratio by taking correlations.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Non-Patent Documents

[0004]

Non-Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] According to the technology disclosed in Non-Patent Document 1, the dot product of the pre-slip with the amount of slip (vector) calculated after the earthquake occurs is taken to detect the immediate pre-slip. If the dot product is large, it is determined that the amount of slip is large, so it cannot be used for earthquake prediction. In other words, it is not a technology that can capture earthquake precursors using only data from before the earthquake occurs.

[0006] According to the technology disclosed in Patent Document 1, it has been successful to detect ionospheric anomalies about an hour before major earthquakes such as the Tohoku offshore earthquake. However, there is no conclusive evidence from the perspective of crustal deformation to determine whether these anomalies are truly related to the Tohoku offshore earthquake. Therefore, in order to make more accurate earthquake predictions, it is necessary to examine the causal relationship between earthquakes and ionospheric anomalies, and to distinguish between space weather anomalies and ionospheric anomalies as earthquake precursor phenomena.

[0007] The technology disclosed in Non-Patent Document 1 has been criticized for its unreliable results due to common mode errors specific to positioning satellites, and there has been ongoing debate as to whether the pre-slip obtained in Non-Patent Document 1 is truly pre-slip or merely noise.

[0008] This disclosure is made in light of the above-mentioned background and aims to provide a technology that solves all three of the aforementioned problems.

[0009] Specifically, the first problem is solved by calculating the amount of slip using only data from before the earthquake occurred, then solving the second problem (relationship with earthquakes) by obtaining a new correlation with localized crustal deformation fluctuations, and further solving the problem of common-mode errors and other noises by increasing the signal-to-noise ratio by correlating crustal deformation at multiple observation points. [Means for solving the problem]

[0010] An anomaly detection device is provided according to one embodiment. The anomaly detection device includes an input unit that acquires observation vectors from each of a plurality of observation stations for observing crustal deformation vectors, and a control unit that determines anomalies in the observation vectors and the anomaly region. The control unit selects a central observation station from the plurality of observation stations, selects a plurality of surrounding observation stations from the plurality of observation stations based on the distance from the central observation station, calculates a correlation value by the dot product of the observation vectors of each of the plurality of surrounding observation stations and the observation vector of the central observation station, determines whether the observation vector of the central observation station is anomaly based on the correlation value, and determines a predetermined region including the central observation station that has been determined to be anomaly as an anomaly region.

[0011] In a given situation, the anomaly detection device determines whether the observation vector of the central observation station is anomaly based on correlation values, which includes determining whether the observation vector of the central observation station is anomaly based on the average value of correlation values ​​calculated from the dot product of the observation vector of the central observation station and the observation vectors of multiple observation stations surrounding the central observation station.

[0012] In a given scenario, the system further includes a memory unit that stores a virtual epicenter location, information on a first distance, and information on a second distance. Selecting a central observation station from among multiple observation stations based on the virtual epicenter location and the distance from the central observation station includes selecting a central observation station from a region where the distance from the virtual epicenter location to the central observation station is greater than or equal to the first distance and less than or equal to the second distance.

[0013] In a given situation, determining whether the observation vector of the central observation station is anomalous or not based on the correlation value is: This includes determining whether the observation vector of the central observation station is abnormal, based on the average value of the correlation calculated by the dot product of the observation vector of the central observation station and the observation vectors of multiple observation stations surrounding the central observation station, and the dot product of a unit vector pointing from a hypothetical epicenter location towards the central observation station and a unit vector in the same direction as the observation vector of the central observation station.

[0014] In a given situation, determining whether an area is anomalous if its distance from a hypothetical epicenter is greater than or equal to a first distance and less than or equal to a second distance, based on correlation values, includes determining that an area is anomalous if its distance from a hypothetical epicenter is greater than or equal to a first distance and less than or equal to a second distance, based on the increase per unit time of the value obtained by dividing the sum of the average values ​​of correlation values ​​calculated from the dot product of the observation vector of each central observation station and the observation vectors of multiple observation stations surrounding that central observation station, by the total number of central observation stations in that area.

[0015] In a given situation, determining whether an area is anomalous if its distance from a hypothetical epicenter is greater than or equal to a first distance and less than or equal to a second distance is based on correlation values, includes determining whether an area is anomalous if its distance from a hypothetical epicenter is greater than or equal to a first distance and less than or equal to a second distance is based on the increase per unit time of the sum of the correlation values ​​calculated by the dot product of the observation vector of each central observation station and the observation vectors of multiple observation stations surrounding that central observation station, divided by the total number of central observation stations in that area.

[0016] In a given scenario, the system includes a memory unit that stores multiple virtual epicenter locations, information on a first distance, and information on a second distance. For each virtual epicenter location, the system determines whether an area where the distance from the virtual epicenter location is greater than or equal to the first distance and less than or equal to the second distance is abnormal. This determination is made based on the increase per unit time of the value obtained by dividing the sum of the average values ​​of the correlation values ​​calculated by the dot product of the observation vector of each central observation station and the observation vectors of multiple observation stations surrounding that central observation station, by the total number of central observation stations in that area. The system further includes an epicenter selection unit that selects one virtual epicenter location from among multiple virtual epicenter locations as a high-probability epicenter location based on the increase per unit time of the value obtained by dividing the sum of the average values ​​of the correlation values ​​calculated by the dot product of the observation vector of each central observation station and the observation vectors of multiple observation stations surrounding that central observation station, by the total number of central observation stations in that area.

[0017] In a given scenario, the system includes a memory unit that stores multiple virtual epicenter locations, information on a first distance, and information on a second distance, and for each virtual epicenter location, To determine whether an area is abnormal if its distance from a hypothetical epicenter is greater than or equal to a first distance and less than or equal to a second distance, the following is determined: For all central observation stations within the area, the average of the correlation values ​​calculated by the dot product of the observation vector of each central observation station and the observation vectors of multiple observation stations surrounding that central observation station, and the sum of the dot products of the unit vector pointing from the hypothetical epicenter to the central observation station and the unit vector in the same direction as the observation vector of the central observation station, divided by the total number of central observation stations in the area, is used to determine if an area is abnormal. Furthermore, the following is determined based on the increase per unit time of the sum of the correlation values ​​calculated by the dot product of the observation vector of each central observation station and the observation vectors of multiple observation stations surrounding that central observation station, and the sum of the dot products of the unit vector pointing from the hypothetical epicenter to the central observation station and the unit vector in the same direction as the observation vector of the central observation station, divided by the total number of central observation stations in the area. It includes an epicenter selection unit that selects one virtual epicenter location from multiple virtual epicenter locations as a high-probability epicenter location.

[0018] Another embodiment provides a method for detecting localized anomalies in crustal deformation. The method includes the steps of: acquiring observation vectors from multiple observation stations for observing crustal deformation vectors; determining anomalies in the observation vectors and the anomaly region; selecting a central observation station from among the multiple observation stations; selecting multiple surrounding observation stations from among the multiple observation stations based on the distance from the central observation station; calculating a correlation value by the dot product of the observation vectors of each of the multiple surrounding observation stations and the observation vector of the central observation station; and determining whether the observation vector of the central observation station is anomaly based on the correlation value; and determining a predetermined region including the central observation station that has been determined to be anomaly as an anomaly region.

[0019] In a certain situation, the step of determining whether the observation vector of the central observation station is abnormal based on the correlation value includes determining that the observation vector of the central observation station is abnormal based on the average value of the correlation values calculated by the inner product of the observation vector of the central observation station and the observation vectors of a plurality of observation stations around the central observation station.

[0020] In a certain situation, further comprising the step of storing a virtual epicenter position, information on a first distance, and information on a second distance, the step of selecting the central observation station from among a plurality of observation stations based on the distance from the virtual epicenter position to the central observation station includes selecting the central observation station from a region where the distance from the virtual epicenter position to the central observation station is greater than or equal to the first distance and less than or equal to the second distance.

[0021] In a certain situation, the step of determining whether the observation vector of the central observation station is abnormal based on the correlation value includes determining that the observation vector of the central observation station is abnormal based on the average value of the correlation values calculated by the inner product of the observation vector of the central observation station and the observation vectors of a plurality of observation stations around the central observation station, and the inner product of the unit vector pointing from the virtual epicenter position to the central observation station and the unit vector in the same direction as the observation vector of the central observation station.

[0022] In a certain situation, the step of determining whether a region where the distance from the virtual epicenter position is greater than or equal to the first distance and less than or equal to the second distance is abnormal based on the correlation value includes, for all central observation stations within the region, calculating the sum of the average values of the correlation values calculated by the inner product of the observation vector of each central observation station and the observation vectors of a plurality of observation stations around the central observation station, and determining that the region where the distance from the virtual epicenter position is greater than or equal to the first distance and less than or equal to the second distance is an abnormal region based on the increase amount per unit time of the value obtained by dividing the sum by the total number of all central observation stations within the region.

[0023] In a certain situation, the step of determining whether a region where the distance from a virtual epicenter position is greater than or equal to a first distance and less than or equal to a second distance is abnormal based on a correlation value includes, for all central observation stations within the region, the average value of the correlation values calculated by the inner product of the observation vector of each central observation station and the observation vectors of a plurality of observation stations around the central observation station, and the sum of the inner products of the unit vector pointing from the virtual epicenter position to the central observation station and the unit vector in the same direction as the observation vector of the central observation station. Based on the increase amount per unit time of the value obtained by dividing by the total number of all central observation stations within the region, the step of determining that the region where the distance from the virtual epicenter position is greater than or equal to the first distance and less than or equal to the second distance is an abnormal region is included.

[0024] In a certain situation, it includes the step of storing a plurality of virtual epicenter positions, information on a first distance, and information on a second distance. For each virtual epicenter position, The step of determining whether a region where the distance from the virtual epicenter position is greater than or equal to the first distance and less than or equal to the second distance is abnormal includes, for all central observation stations within the region, the sum of the average values of the correlation values calculated by the inner product of the observation vector of each central observation station and the observation vectors of a plurality of observation stations around the central observation station. Based on the increase amount per unit time of the value obtained by dividing by the total number of all central observation stations within the region, it is determined that the region where the distance from the virtual epicenter position is greater than or equal to the first distance and less than or equal to the second distance is an abnormal region. Furthermore, based on the increase amount per unit time of the value obtained by dividing the sum of the average values of the correlation values calculated by the inner product of the observation vector of the central observation station and the observation vectors of a plurality of observation stations around the central observation station by the total number of all central observation stations within the region, it includes an epicenter selection step of selecting one virtual epicenter position as the high-probability epicenter position from a plurality of virtual epicenter positions. Based on the increase amount per unit time of the value obtained by dividing by the total number of all central observation stations within the region, it includes an epicenter selection step of selecting one virtual epicenter position as the high-probability epicenter position from a plurality of virtual epicenter positions.

[0025] In a given phase, the system includes the step of storing multiple virtual epicenter locations, information on a first distance, and information on a second distance. For each virtual epicenter location, the system determines whether an area where the distance from the virtual epicenter location is greater than or equal to the first distance and less than or equal to the second distance is abnormal, based on the increase per unit time of the value obtained by dividing the sum of the dot products of the average correlation value calculated from the dot product of the observation vector of each central observation station and the observation vectors of multiple observation stations surrounding that central observation station, and the dot product of the unit vector pointing from the virtual epicenter location to the central observation station and the unit vector in the same direction as the observation vector of the central observation station, by the total number of central observation stations in that area. The process includes an epicenter step in which an area where the distance from a hypothetical epicenter is greater than or equal to a first distance and less than or equal to a second distance is determined to be an anomaly region, and further, based on the increase per unit time of the value obtained by dividing the average of the correlation values ​​calculated from the dot product of the observation vector of the central observation station and the observation vectors of multiple observation stations surrounding the central observation station, and the sum of the dot products of the unit vector pointing from the hypothetical epicenter to the central observation station and the unit vector in the same direction as the observation vector of the central observation station, by the total number of central observation stations in the region, one hypothetical epicenter is selected as a high-probability epicenter from among multiple hypothetical epicenter locations.

[0026] According to another embodiment, a program is provided for causing a computer to perform the above method. [Effects of the Invention]

[0027] According to one embodiment, in determining anomalies in crustal deformation, it is possible to improve the signal-to-noise ratio related to common mode errors associated with positioning satellites, and to detect pre-slip, which is a precursory slip, from data before a major earthquake occurs. Furthermore, when there are multiple candidate epicenters, it is possible to select the most likely epicenter, that is, to estimate the epicenter.

[0028] The above and other purposes, features, aspects and advantages of this disclosure will become apparent from the following detailed description of this disclosure, as understood in conjunction with the attached drawings. [Brief explanation of the drawing]

[0029] [Figure 1] This figure shows an example of an overall structure of a crustal deformation observation system according to a certain embodiment. [Figure 2] An example of the functional configuration of an anomaly detection device 120 according to a certain embodiment is shown. [Figure 3] This figure shows an example of the configuration of an anomaly detection device 120 according to a certain embodiment. [Figure 4] This is the result of a correlation analysis performed by an anomaly detection device 120 when there are eight surrounding observation stations around a central observation station. The lower right shows the trend of the correlation value. It can be seen that the correlation value increased linearly in the data from 1-2 hours immediately before the 2011 Tohoku earthquake. The upper right shows the crustal deformation in the east-west direction (data at 5-minute intervals), and the upper right shows the crustal deformation in the north-south direction (data at 5-minute intervals). The lower right shows the dot product of the seismic solution and the crustal deformation vector for the 2011 Tohoku earthquake. The unit of the vertical axis (correlation value) in the lower right is m × m = m². [Figure 5] This graph shows the average behavior of correlation values ​​during the 2011 Tohoku earthquake, as detected by the anomaly detection device 120, categorized by distance from the epicenter. The horizontal axis represents time, indicating how many hours before the main shock it was. [Figure 6] Enlarged view of [Figure 5]. This shows the average behavior of correlation values ​​during the 2011 Tohoku earthquake, as detected by the anomaly detection device 120, categorized by distance from the epicenter. The horizontal axis represents time, indicating how many hours before the main shock it was. [Modes for carrying out the invention]

[0030] The embodiments of the technical concept relating to this disclosure will be described below with reference to the drawings. In the following description, identical parts are denoted by the same reference numerals. Their names and functions are the same. Therefore, detailed descriptions of them will not be repeated. In the following description, when referring to multiple identical configurations, they may be expressed as configuration 101A, 101B, and so on. Also, when referring to them collectively, they will be expressed as configuration 101.

[0031] Figure 1 shows an example of the overall structure of the anomaly detection system 100 according to this embodiment. The anomaly detection system 100 acquires and analyzes crustal deformation vectors over a certain period (e.g., every 5 minutes) from data obtained by each receiving station by the GNSS satellite 102, and can detect or predict anomalies in the crustal deformation vectors at each value.

[0032] The anomaly detection system 100 consists of multiple observation stations 105, The system includes an observation data database (DB) 110, an anomaly detection device 120, a service provider 130, and a user terminal 140. Multiple observation stations 105, the observation data database 110, the anomaly detection device 120, and the service provider 130 are configured to communicate with each other via a network 150.

[0033] In some cases, the anomaly detection system 100 may include only the anomaly detection device 120. In that case, the anomaly detection system 100 may cooperate with external devices or services, such as multiple observation stations 105, an observation data DB 110, a service provider 130, and a user terminal 140.

[0034] Observation station 105 observes crustal deformation vectors at each observation station 105 by communicating with the GNSS satellite 102. Observation station 105 transmits the acquired observation vectors to the observation data DB 110. Multiple observation stations 105 can be deployed in areas where crustal deformation can be observed.

[0035] The observation data DB 110 acquires and stores observation vectors from multiple observation stations 105. The observation data DB 110 can be located on any device, such as a server. In some cases, the observation vectors may be crustal deformation vectors at regular intervals. In that case, a server equipped with the observation DB 110 stores the observation vectors as crustal deformation vectors in the observation data DB 110. In other aspects, the observation vector may include delay information of communication between the GNSS satellite 102 and the observation station 105, position information of the GNSS satellite 102, and position information of the observation station 105. In that case, the server equipped with the observation data DB 110 can calculate a crustal deformation vector from the information included in the observation vector and store the crustal deformation vector in the observation data DB 110.

[0036] Furthermore, in other contexts, the observation data DB110 may be represented as a relational database table, or as any other data format such as JSON (JavaScript® Object Notation). Also, in other contexts, the observation station 105, observation data DB110, and GNSS satellite 102 may constitute a crustal deformation observation system.

[0037] The observation data DB 110 is used to obtain crustal deformation data from observation station 105. In some cases, the anomaly detection device 120 may calculate a crustal deformation vector based on the observation data from observation station 105. The anomaly detection device 120 may also provide an alert to service provider 130. In other cases, the anomaly detection device 120 may be implemented as a PC (Personal Computer), a server, or a cloud service.

[0038] The service provider 130 may provide various forecasts or value-added services utilizing forecasts to the user terminal 140 based on information acquired from the anomaly detection device 120. In some cases, the service provider 130 may be implemented as a server or a cloud service. In other cases, the functions of the service provider 130 may be included in the anomaly detection device 120.

[0039] The user terminal 140 may receive alerts from the anomaly detection device 120 or the service provider 130 and display the content of the alerts on its display. In some cases, the user terminal 140 may be a PC, smartphone, tablet, wearable device, television, radio, car driving monitor, or other device.

[0040] Figure 2 shows an example of the functional configuration of the anomaly detection device 120. In some cases, the configuration shown in Figure 2 may be implemented as software, in which case these configurations can be executed by the hardware shown in Figure 3.

[0041] The anomaly detection device 120 comprises a virtual epicenter location DB 210, an observation station location data DB 220, a virtual epicenter location observation station distance calculation unit 230, a crustal deformation vector calculation unit 240, a correlation value calculation unit 250, an epicenter selection unit 260, and an anomaly determination unit 270. In a given scenario, the virtual epicenter location observation DB 210 and the observation station location data DB 220 may be represented as tables in a relational database, or they may be represented in any other data format such as JSON.

[0042] The virtual epicenter location DB210 stores location information for virtual epicenter locations that are candidates for earthquakes. In some cases, the location information may be information indicating latitude and longitude. In other cases, the location information may be information indicating each region when the observed region (such as the Japanese archipelago) is divided into arbitrary regions such as a grid.

[0043] The observation station location DB220 stores the location information of observation station 105. In some cases, the location information may be information indicating latitude and longitude. In other cases, the location information may be information indicating each region when the area to be observed (such as the Japanese archipelago) is divided into arbitrary regions such as a grid.

[0044] The virtual epicenter location observation station distance calculation unit 230 calculates the distance between each virtual epicenter location and each observation station 105. In some cases, the virtual epicenter location observation station distance calculation unit 230 may calculate the distance between virtual epicenter location observation stations based on the difference in latitude and longitude between the virtual epicenter location and the observation station 105. In other cases, the virtual epicenter location observation station distance calculation unit 230 may calculate the distance between the virtual epicenter location and the observation station 105 based on how many squares apart the virtual epicenter location and the observation station 105 are on a grid. In that case, the virtual epicenter location observation station distance calculation unit 230 can calculate the distance between the virtual epicenter location and the observation station 105 by multiplying the number of squares between the virtual epicenter location and the observation station 105 by the distance per square.

[0045] The crustal deformation vector calculation unit 240 calculates the difference vector of the change in position information at each observation station 105 at regular intervals such as 5 minutes, and calculates this as the crustal deformation vector for each observation station 105.

[0046] The correlation value calculation unit 250 performs processing based on the crustal deformation vectors of each observation station 105 output by the crustal deformation vector calculation unit 240. The correlation value calculation unit 250 selects one station as the central observation station and selects several observation stations in its vicinity as surrounding observation stations. In a given scenario, the correlation value calculation unit 250 can select any number of surrounding observation stations from among the observation stations 105 whose distance from the central observation station is within a predetermined distance.

[0047] Next, the correlation value calculation unit 250 calculates the correlation value between the crustal deformation vector at the central observation station at a certain time t and the crustal deformation vectors of each surrounding observation station at a certain time t. The crustal deformation vector at the central observation station and the crustal deformation vectors of each surrounding observation station are calculated based on the crustal deformation vector at the same time t. The correlation value calculation unit 250 selects each of the observation stations 105 in turn as the central observation station and repeatedly performs the above correlation value calculation process.

[0048] Furthermore, the correlation value calculation unit 250 calculates a relative value based on the correlation value at each observation station 105 at a certain time t. The correlation value calculation unit 250 calculates the median and standard deviation of the correlation values ​​at each observation station 105 at a certain time t. Then, based on the calculated median and standard deviation of the correlation values, the correlation value calculation unit 250 calculates a relative value indicating how much each observation station 105 differs from the median of the correlation values. For each correlation value of all observation stations 105, the correlation value calculation unit 250 calculates a relative value. The above relative value calculation process is repeatedly executed. The correlation value calculation unit 250 may calculate the relative values ​​of each observation station 105 using known techniques. For example, the correlation value calculation unit 250 may perform the relative value calculation shown in Patent Document 1.

[0049] The epicenter selection unit 260 classifies each observation station 105 for each virtual epicenter location according to the distance calculated by the virtual epicenter location observation station distance calculation unit 230. For all observation stations 105 whose virtual epicenter location observation station distance falls within the same distance range, for example, if the virtual epicenter location observation station distance falls between 120 km and 140 km, the unit calculates the average of the correlation values ​​calculated by the correlation value calculation unit 250 and selects the virtual epicenter location with the highest average increase rate of the correlation value as the high-probability epicenter location. In other situations, if the average increase rate of the correlation value within the distance range becomes larger as the distance between the virtual epicenter location observation stations decreases, the virtual epicenter location may be selected as the high-probability epicenter location.

[0050] The location of the high-probability epicenter should only be selected when the average rate of increase of the correlation value exceeds a certain threshold, and should not be selected otherwise.

[0051] The anomaly detection unit 270 determines whether the relative values ​​of all observation stations 105 at a given time are above a predetermined threshold. For example, the anomaly detection unit 270 may determine that the observed value (crustal deformation vector) of observation station 105A is anomaly based on the fact that the relative value of observation station 105A is above a predetermined threshold. In a given situation, the anomaly detection unit 270 may determine whether the observed value (crustal deformation vector) of observation station 105A is anomaly by determining whether the correlation value is above a predetermined threshold, without calculating the relative value. In that case, the anomaly detection unit 270 may determine that the observed value (crustal deformation vector) of observation station 105A is anomaly based on the fact that the correlation value is above a predetermined threshold.

[0052] Figure 3 shows an example of the configuration of the anomaly detection device 120. The anomaly detection device 120 includes a CPU (Central Processing Unit) 301, a primary storage device 302, a secondary storage device 303, an external device interface 304, an input interface 305, an output interface 306, and a communication interface 307.

[0053] The CPU 301 can execute programs to implement various functions of the anomaly detection device 120. The CPU 301 is composed of, for example, at least one integrated circuit. The integrated circuit may consist of at least one CPU, at least one FPGA (Field Programmable Gate Array), or ASIC (Application Specific Integrated Circuit), or a combination thereof.

[0054] The secondary storage device 303 is a non-volatile memory and may store programs executed by the CPU 301 and data referenced by the CPU 301. In this case, the CPU 301 references the data read from the secondary storage device 303 to the primary storage device 302. In some cases, the secondary storage device 303 may be implemented by an HDD (Hard Disk Drive), SSD (Solid State Drive), EPROM (Erasable Programmable Read Only Memory), EEPROM (Electrically Erasable Programmable Read Only Memory), or flash memory.

[0055] The external device interface 304 can be connected to any external device such as a printer, scanner, and external HDD. In some cases, the external device interface may be implemented by a USB (Universal Serial Bus) terminal or the like.

[0056] The input interface 305 can be connected to any input device such as a keyboard, mouse, touchpad, or gamepad. In some cases, the input interface 305 may be implemented by a USB terminal, PS / 2 terminal, and Bluetooth® module, etc.

[0057] The output interface 306 can be connected to any output device such as a cathode ray tube display, a liquid crystal display, or an organic EL (Electro-Luminescence) display. In some cases, the output interface 306 may be connected to a USB terminal, a D-Sub terminal, a DVI (Digital Visual Interface) terminal, and an HDMI (registered trademark) terminal. This may be implemented by a High-Definition Multimedia Interface (H-HD) terminal or the like.

[0058] The communication interface 307 is connected to wired or wireless network equipment. In some cases, the communication interface 307 may be implemented by a wired LAN (Local Area) port and a Wi-Fi® (Wireless Fidelity) module, etc. In other cases, the communication interface 307 is Data may be sent and received using communication protocols such as TCP / IP (Transmission Control Protocol / Internet Protocol), UDP (User Datagram Protocol), and MQTT (Message Queuing Telemetry Transport).

[0059] The left portion of Figure 4 shows an example of the selection of observation station 105, and the arrows indicate the crustal deformation vectors observed at each observation station 105. Observation station 105, marked with a filled-in triangle, is the central observation station, and the other eight triangles represent surrounding observation stations.

[0060] The right portion of Figure 5 shows the crustal deformation from 48 hours before the magnitude 9.0 Tohoku-oki earthquake on March 11, 2011, to just before it occurred, based on the observation station selected on the left. The most important figure is the correlation analysis result in the lower right, which shows that the correlation value increased from about 2 hours before the earthquake. It is noteworthy that this is probably the first data in the world to capture the pre-slip of the 2011 Tohoku-oki earthquake using continuous crustal deformation data. This increasing trend in the correlation value was captured solely from data before the earthquake occurred, and this data forms the basis of the present invention. In the other figures on the right, from top to bottom, are deformation in the east-west direction, deformation in the north-south direction, the dot product with the seismic solution (crustal deformation solution calculated after the earthquake), and the results of the correlation analysis, all with a time scale from -48 hours to 0 hours.

[0061] Figure 6 plots the increase in the correlation value of crustal deformation according to the present invention for the 2011 Tohoku earthquake, using average values ​​at different distances from the epicenter. A characteristic behavior is that the closer the distance from the epicenter, the higher the rate of increase in the correlation value, starting two hours before the earthquake. This suggests that the amount of crustal deformation correlation value captured by the present invention has captured (perhaps for the first time in the world) the "slip" that indicates an earthquake precursor. Previously, this could not be captured because the vectors of each observation station were being looked at individually, but it is clear that the present invention has captured the "pre-slip" phenomenon clearly.

[0062] Figure 7 is an enlarged view of Figure 6. It can be seen more clearly that the rate of increase of the correlation value changes for each distance from the epicenter, as shown in

[0061] . This time from two hours before the earthquake to the earthquake (lead time of two hours) is consistent with the anomaly one hour before the earthquake disclosed in Patent Document 1.

[0063] Conversely, as shown in Figures 6 and 7, the distribution of correlation values ​​has a clear relationship with the distance from the epicenter. Therefore, as disclosed in this invention, when there are multiple virtual epicenter locations, it is possible to solve the optimization problem of determining which is the most likely (most probable) epicenter. Multiple virtual epicenter locations are determined by dividing the Japanese archipelago into multiple grids and estimating the most probable epicenter location from among them. Furthermore, by similarly dividing that grid into multiple grids, it is possible to estimate the epicenter location with high accuracy. This invention not only detects pre-slip crustal deformation vectors based on correlation values, but also enables the estimation of the epicenter's location.

[0064] As described above, the anomaly detection device 120 according to this embodiment calculates a correlation value based on multiple virtual epicenter locations and crustal deformation vectors of multiple surrounding observation stations selected from the central observation station. It notifies of an anomaly based on the correlation value or the rate of increase of the correlation value, and further enables the estimation of the epicenter location where an earthquake will occur. In particular, the ability to observe crustal deformation at ground observation stations to detect pre-slips of a large earthquake approximately two hours in advance is considered useful because it can prevent human casualties and accidents at nuclear power plants and railway infrastructure without incurring significant social costs.

[0065] The embodiments disclosed herein should be considered in all respects to be illustrative and not restrictive. The scope of this disclosure is indicated by the claims rather than the foregoing description, and all modifications are intended to be included in the sense and scope equivalent to the claims. Furthermore, the disclosures described in the embodiments and each variation are intended to be implemented, as far as possible, individually or in combination. [Explanation of Symbols]

[0066] 100 Anomaly Detection System 102GNSS satellite 105 observation stations 110 Observation Data DB (Data Base) 120 Anomaly Detection Device 130 service providers 140 user terminals 150 networks 210 Virtual Epicenter Location Database 220 Observation Station Location Data Database 230 Virtual Epicenter Location Observation Station Distance Calculation Unit 240 Crustal deformation vector calculation unit 250 Correlation Value Calculation Unit 260 Epicenter Selection Section 270 Abnormality Judgment Unit 301CPU 302 Primary storage 303 Secondary storage device 304 External Device Interface 305 Input Interface 306 output interface 307 Communication Interface

Claims

1. The system includes an input unit that acquires observation vectors from multiple observation stations for observing crustal deformation vectors, and a control unit that determines anomalies in the observation vectors and the areas of those anomalies. The control unit, Select a central observation station from among the aforementioned multiple observation stations. Based on the distance from the central observation station, select multiple surrounding observation stations from among the multiple observation stations. The observation vectors of each of the aforementioned multiple surrounding observation stations and the observation vector of the central observation station The correlation value is calculated using the dot product of the aforementioned observation vectors. An anomaly detection device that determines whether the observation vector of the central observation station is abnormal based on the correlation value, and determines a predetermined region including the central observation station that has been determined to be abnormal as an abnormal region.

2. Based on the correlation value, determining whether the observation vector of the central observation station is abnormal is: An anomaly detection device according to claim 1, which includes determining that the observation vector of the central observation station is abnormal based on the average value of the correlation calculated by the dot product of the observation vector of the central observation station and the observation vectors of a plurality of observation stations located around the central observation station.

3. The system further includes a memory unit that stores a virtual epicenter location, information on a first distance, and information on a second distance. Selecting the central observation station from among the plurality of observation stations based on the distance between the virtual epicenter location and the central observation station is performed if the distance between the virtual epicenter location and the central observation station is greater than or equal to the first distance and less than or equal to the second distance. An anomaly detection device according to any one of claims 1 to 2, comprising selecting a central observation station from the area below.

4. Based on the correlation value, determining whether the observation vector of the central observation station is abnormal is: An anomaly detection device according to claim 3, which includes determining that the observation vector of the central observation station is abnormal based on the average value of the correlation calculated by the dot product of the observation vector of the central observation station and the observation vectors of a plurality of observation stations located around the central observation station, and the dot product of a unit vector pointing from the virtual epicenter location to the central observation station and a unit vector in the same direction as the observation vector of the central observation station.

5. Based on the correlation value, determining whether an area where the distance from the hypothetical epicenter is greater than or equal to the first distance and less than or equal to the second distance is abnormal is performed for all central observation stations within that area. Based on the increase per unit time of the value obtained by dividing the sum of the average values ​​of the correlation values ​​calculated by the dot product of the observation vector of each central observation station and the observation vectors of a plurality of observation stations located around the central observation station by the total number of central observation stations in that region, An anomaly detection device according to claim 3, comprising determining that an area where the distance from the virtual epicenter location is greater than or equal to the first distance and less than or equal to the second distance is an anomaly area.

6. Based on the correlation value, determining whether an area where the distance from the hypothetical epicenter is greater than or equal to the first distance and less than or equal to the second distance is abnormal is performed for all central observation stations within that area. Based on the increase per unit time of the value obtained by dividing the average of the correlation values ​​calculated by the dot product of the observation vector of each central observation station and the observation vectors of a plurality of observation stations located around the central observation station, and the sum of the dot products of the unit vector pointing from the hypothetical epicenter location to the central observation station and the unit vector in the same direction as the observation vector of the central observation station, by the total number of central observation stations in that region, An anomaly detection device according to claim 4, comprising determining that an area where the distance from the virtual epicenter location is greater than or equal to the first distance and less than or equal to the second distance is an anomaly area.

7. It has a memory unit that stores multiple virtual epicenter locations, information on a first distance, and information on a second distance. For each of the aforementioned hypothetical epicenter locations, Determining whether an area where the distance from the hypothetical epicenter is greater than or equal to the first distance and less than or equal to the second distance is abnormal is performed for all central observation stations within that area. Based on the increase per unit time of the value obtained by dividing the sum of the average values ​​of the correlation values ​​calculated by the dot product of the observation vector of each central observation station and the observation vectors of a plurality of observation stations located around the central observation station by the total number of central observation stations in that region, The region where the distance from the aforementioned hypothetical epicenter is greater than or equal to the first distance and less than or equal to the second distance is determined to be an abnormal region, and further Based on the increase per unit time of the value obtained by dividing the sum of the average values ​​of the correlation values ​​calculated by the dot product of the observation vector of the central observation station and the observation vectors of a plurality of observation stations located around the central observation station by the total number of central observation stations in that region, An anomaly detection device according to claim 5, further comprising an epicenter selection unit that selects one virtual epicenter location as a high-probability epicenter location from the plurality of virtual epicenter locations.

8. It includes a memory unit that stores multiple virtual epicenter locations, information on a first distance, and information on a second distance. For each of the aforementioned hypothetical epicenter locations, Determining whether an area where the distance from the hypothetical epicenter is greater than or equal to the first distance and less than or equal to the second distance is abnormal is performed for all central observation stations within that area. Based on the increase per unit time of the value obtained by dividing the average of the correlation values ​​calculated by the dot product of the observation vector of each central observation station and the observation vectors of multiple observation stations located around the central observation station, and the sum of the dot products of the unit vector pointing from the hypothetical epicenter location to the central observation station and the unit vector in the same direction as the observation vector of the central observation station, by the total number of central observation stations in that region, The region where the distance from the aforementioned hypothetical epicenter is greater than or equal to the first distance and less than or equal to the second distance is determined to be an abnormal region, and further Anomaly detection device according to claim 6, which includes an epicenter selection unit that selects one virtual epicenter location as a high-probability epicenter location from a plurality of virtual epicenter locations based on the increase per unit time of the sum of the average of the correlation values ​​calculated by the dot product of the observation vector of the central observation station and the observation vectors of a plurality of observation stations located around the central observation station, and the sum of the dot products of the unit vector pointing from the virtual epicenter location towards the central observation station and the unit vector in the same direction as the observation vector of the central observation station, divided by the total number of central observation stations in the region.

9. A method for detecting anomalies in crustal deformation, The method comprises the steps of acquiring observation vectors from multiple observation stations for observing crustal deformation vectors, and determining anomalies in the observation vectors and the areas of such anomalies. A central observation station is selected from the aforementioned multiple observation stations, and based on the distance from the central observation station, multiple surrounding observation stations are selected from the aforementioned multiple observation stations. The steps include: calculating a correlation value by the dot product of the observation vectors of each of the plurality of surrounding observation stations and the observation vector of the central observation station; A method comprising the steps of determining whether the observation vector of the central observation station is abnormal based on the correlation value, and determining a predetermined region including the central observation station that was determined to be abnormal as an abnormal region.

10. The step of determining whether the observation vector of the central observation station is abnormal based on the correlation value is: The method according to claim 9, further comprising determining that the observation vector of the central observation station is abnormal based on the average value of the correlation calculated by the dot product of the observation vector of the central observation station and the observation vectors of a plurality of observation stations located around the central observation station.

11. The method according to claim 10, further comprising the step of storing a virtual epicenter location, information of a first distance, and information of a second distance, wherein the step of selecting the central observation station from among the plurality of observation stations based on the distance from the virtual epicenter location and the central observation station includes selecting the central observation station from a region in which the distance from the virtual epicenter location and the central observation station is greater than or equal to the first distance and less than or equal to the second distance.

12. The step of determining whether the observation vector of the central observation station is abnormal based on the correlation value is: The method of claim 11, comprising the step of determining that the observation vector of the central observation station is abnormal, based on the average value of the correlation calculated by the dot product of the observation vector of the central observation station and the observation vectors of a plurality of observation stations located around the central observation station, and the dot product of a unit vector pointing from the hypothetical epicenter location to the central observation station and a unit vector in the same direction as the observation vector of the central observation station.

13. Based on the correlation value, the step of determining whether an area where the distance from the hypothetical epicenter is greater than or equal to the first distance and less than or equal to the second distance is abnormal is performed for all of the central observation stations within that area. Based on the increase per unit time of the value obtained by dividing the sum of the average values ​​of the correlation values ​​calculated by the dot product of the observation vector of each central observation station and the observation vectors of a plurality of observation stations located around the central observation station by the total number of central observation stations in that region, The method according to claim 11, further comprising the step of determining that an area whose distance from the virtual epicenter is greater than or equal to the first distance and less than or equal to the second distance is an abnormal area.

14. Based on the correlation value, the step of determining whether an area where the distance from the hypothetical epicenter is greater than or equal to the first distance and less than or equal to the second distance is abnormal is performed for all of the central observation stations within that area. The method according to claim 12, further comprising the step of determining that an area where the distance from the virtual epicenter is greater than or equal to the first distance and less than or equal to the second distance is an abnormal area, based on the increase per unit time of the sum of the average of the correlation values ​​calculated by the dot product of the observation vector of each of the central observation stations and the observation vectors of a plurality of observation stations located around the central observation station, and the sum of the dot products of the unit vector pointing from the virtual epicenter to the central observation station and the unit vector in the same direction as the observation vector of the central observation station, divided by the total number of central observation stations in that area.

15. The process includes the step of storing multiple virtual epicenter locations, first distance information, and second distance information. For each of the aforementioned hypothetical epicenter locations, The step of determining whether an area where the distance from the hypothetical epicenter is greater than or equal to the first distance and less than or equal to the second distance is abnormal is performed for all of the central observation stations within that area. Based on the increase per unit time of the value obtained by dividing the sum of the average values ​​of the correlation values ​​calculated by the dot product of the observation vector of each central observation station and the observation vectors of multiple observation stations located around the central observation station by the total number of central observation stations in that region, The region where the distance from the aforementioned hypothetical epicenter is greater than or equal to the first distance and less than or equal to the second distance is determined to be an abnormal region, and further Based on the increase per unit time of the value obtained by dividing the sum of the average values ​​of the correlation values ​​calculated by the dot product of the observation vector of the central observation station and the observation vectors of a plurality of observation stations located around the central observation station by the total number of central observation stations in that region, The method according to claim 13, further comprising an epicenter selection step of selecting one virtual epicenter location from the plurality of virtual epicenter locations as a high-probability epicenter location.

16. The process includes the step of storing multiple virtual epicenter locations, first distance information, and second distance information. For each of the aforementioned hypothetical epicenter locations, Determining whether an area where the distance from the hypothetical epicenter is greater than or equal to the first distance and less than or equal to the second distance is abnormal is performed for all central observation stations within that area. Based on the increase per unit time of the value obtained by dividing the average of the correlation values ​​calculated by the dot product of the observation vector of each central observation station and the observation vectors of multiple observation stations located around the central observation station, and the sum of the dot products of the unit vector pointing from the hypothetical epicenter location to the central observation station and the unit vector in the same direction as the observation vector of the central observation station, by the total number of central observation stations in that region, The region where the distance from the aforementioned hypothetical epicenter is greater than or equal to the first distance and less than or equal to the second distance is determined to be an abnormal region, and further The average value of the correlation calculated by the dot product of the observation vector of the central observation station and the observation vectors of a plurality of observation stations located around the central observation station, Based on the increase per unit time of the value obtained by dividing the sum of the dot products of the unit vector pointing from the hypothetical epicenter location to the central observation station and the unit vector of the central observation station in the same direction as the observation vector by the total number of central observation stations in that region, The method according to claim 14, further comprising the epicenter step of selecting one virtual epicenter location from the plurality of virtual epicenter locations as a high-probability epicenter location.

17. A program for causing a computer to perform the method described in any one of claims 1 to 17.

Images