JP6677300B2 - Epicenter distance estimation device, epicenter distance estimation method, and program - Google Patents

Description

The present invention, for estimating the epicenter distance when the earthquake occurred, relates epicentral distance estimation apparatus and epicentral distance estimation method further relates to a program for realizing these.

  When an earthquake occurs, it is necessary to quickly determine the epicenter distance in order to estimate the seismic intensity and the arrival time of the main motion in each place. Normally, the depth of the epicenter distance is specified based on seismic intensity detected by seismic intensity meters at multiple locations.

  However, if the location of the epicenter is on the sea floor or in an area where seismic intensity is low, it will take too long to acquire seismic intensity measured by multiple seismic intensity meters. Is delayed. For this reason, in recent years, a technique for specifying the epicenter distance using only the seismic intensity measured by a single seismic intensity meter has been developed.

As such a technology, "the rise of the strength of the seismic wave form data at the time of the earthquake of reach, becomes stronger as there is an earthquake close to the epicenter, epicenter becomes gentle as a distant earthquake," taking advantage of the fact A technique for estimating the epicenter distance is known (for example, see Patent Document 1).

  Specifically, in the technology disclosed in Patent Document 1, the absolute value of time-series data obtained from a seismograph is y (t), the time is t, and the time when the seismometer detects an earthquake is t = 0. The waveform shape of the initial part of the earthquake acquired by the seismometer is fitted by the function shown by the following equation (1). In the following equation 1, A is a parameter relating to the maximum amplitude of the initial motion portion, and B is a parameter relating to a time change of the initial motion amplitude of the earthquake waveform. Actually, in the fitting, an operation is performed in which complicated individual point attributes are reflected on parameters A and B in which what influences are unknown, depending on human experience and intuition. .

  In the technique disclosed in Patent Document 1, the parameters A and B are obtained by the least squares method. Among these, there is a correlation between the parameter B and the epicenter distance, but it has been found that this correlation is not affected by the magnitude. Therefore, if the correlation between the parameter B and the epicenter distance is formulated in advance, the epicenter distance is specified by calculating the parameter B using Equation 1 from the waveform shape of the initial part of the earthquake. According to the technology disclosed in Patent Literature 1, the epicenter distance can be quickly specified from the waveform shape of the initial part of the earthquake.

JP-A-2002-277557

However, were the technique disclosed in Patent Document 1, depending on the situation, there is a possibility that a situation that can not be calculated coefficients A and B are generated, there is a problem that reliability is not sufficient. In addition, the technique disclosed in Patent Document 1 has a problem that it is difficult to reduce the time required for calculating the epicenter distance.

An example of the object of the present invention is to solve the above problem, to stably calculate the epicenter distance, and to shorten the calculation time, an epicenter distance estimation device, an epicenter distance estimation method, and a program Is to provide.

To achieve the above object, an epicenter distance estimation device according to one aspect of the present invention includes:
An earthquake information acquisition unit for acquiring waveform data of the generated earthquake;
An estimation processing unit that estimates the epicenter distance by applying the acquired waveform data to a learning model obtained by learning the relationship between the waveform data of the earthquake and the epicenter distance,
It is characterized by having.

To achieve the above object, the epicenter distance estimation method according to one aspect of the present invention includes:
(A) acquiring waveform data of the earthquake that occurred;
(B) estimating the epicenter distance by applying the acquired waveform data to a learning model obtained by learning the relationship between the waveform data of the earthquake and the epicenter distance;
Characterized by having

Furthermore, in order to achieve the above object, a program according to one aspect of the present invention includes:
On the computer,
(A) acquiring waveform data of the earthquake that occurred;
(B) estimating the epicenter distance by applying the acquired waveform data to a learning model obtained by learning the relationship between the waveform data of the earthquake and the epicenter distance;
Ru is running, characterized in that.

  As described above, according to the present invention, the calculation of the epicenter distance can be performed stably, and the calculation time can be reduced.

FIG. 1 is a block diagram illustrating a schematic configuration of the epicenter distance estimation device according to the first embodiment of the present invention. FIG. 2 is a block diagram specifically illustrating a configuration of the epicenter distance estimation device according to the first embodiment of the present invention. FIG. 3 is a diagram illustrating an example of input data and correct data used for learning in the first embodiment. FIG. 4 is a flowchart showing an operation of the epicenter distance estimating apparatus according to Embodiment 1 of the present invention when performing a learning process. FIG. 5 is a flowchart showing an operation of the epicenter distance estimating apparatus according to Embodiment 1 of the present invention when performing estimation processing. FIG. 6 is a block diagram specifically showing a configuration of the epicenter distance estimation device according to Embodiment 2 of the present invention. FIG. 7 is a flowchart showing an operation of the epicenter distance estimating apparatus according to Embodiment 2 of the present invention when performing a learning process. FIG. 8 is a flowchart showing an operation of the epicenter distance estimating apparatus according to Embodiment 2 of the present invention when performing estimation processing. FIG. 9 is a block diagram illustrating an example of a computer that realizes the epicenter distance estimation device according to the first and second embodiments of the present invention.

(Embodiment 1)
Hereinafter, an epicenter distance estimation device, an epicenter distance estimation method, and a program according to Embodiment 1 of the present invention will be described with reference to FIGS. 1 to 5.

[Device configuration]
First, a schematic configuration of the epicenter distance estimation device according to the first embodiment will be described with reference to FIG. FIG. 1 is a block diagram illustrating a schematic configuration of the epicenter distance estimation device according to the first embodiment of the present invention.

  The epicenter distance estimating apparatus 10 according to the first embodiment shown in FIG. 1 is an apparatus for estimating the epicenter distance from waveform data measured when an earthquake occurs. As shown in FIG. 1, the epicenter distance estimation device 10 includes an earthquake information acquisition unit 11 and an estimation processing unit 12.

  The earthquake information acquisition unit 11 acquires waveform data of the generated earthquake. The estimation processing unit 12 estimates the epicenter distance by applying the waveform data acquired by the earthquake information acquisition unit 11 to the learning model. The learning model is obtained in advance by learning the relationship between the waveform data of the earthquake and the epicenter distance.

  As described above, in the first embodiment, unlike the related art, since the epicenter distance is estimated without fitting the waveform data to the function, the epicenter distance can be stably calculated. Further, in the first embodiment, since the calculation processing by the least squares method is not necessary, the calculation time can be shortened.

  Next, the configuration of the epicenter distance estimation device according to the first embodiment will be described more specifically with reference to FIG. FIG. 2 is a block diagram specifically illustrating a configuration of the epicenter distance estimation device according to the first embodiment of the present invention.

  As shown in FIG. 2, in the first embodiment, the epicenter distance estimation device 10 is connected to an earthquake detection device 20 and a comprehensive monitoring system 30 such as seismic activity via a network. Among them, the earthquake detection device 20 includes a seismometer, and when a seismic wave is detected by the seismometer, transmits the detected seismic waveform data to the epicenter distance estimation device 10. In the first embodiment, the earthquake detection device 20 is the acquisition destination of the waveform data of the earthquake information acquisition unit 11.

  Further, in the example of FIG. 2, only a single earthquake detection device 20 is illustrated, but the number of earthquake detection devices 20 to which the epicenter distance estimation device 10 is connected is not particularly limited. However, the earthquake detection device 20 from which the earthquake information obtaining unit 11 obtains may be any one of them.

  The integrated monitoring system 30 for seismic activity etc. is a system owned by the Japan Meteorological Agency in Japan. When an earthquake occurs, the magnitude of the Meteorological Agency is calculated, and the tsunami height is predicted based on the calculated Meteorological Agency magnitude. Furthermore, the comprehensive monitoring system 30 for seismic activity and the like distributes the calculated Meteorological Agency magnitude and the predicted tsunami height to various media as an emergency earthquake bulletin.

  In the first embodiment, the epicenter distance estimating apparatus 10 inputs the estimated epicenter distance to the integrated monitoring system 30 for seismic activity and the like. For this reason, the comprehensive monitoring system 30 for seismic activity and the like uses the epicenter distance estimated by the epicenter distance estimation device 10 to execute the calculation of the JMA magnitude and the prediction of the tsunami height.

  As shown in FIG. 2, in the first embodiment, the epicenter distance estimation device 10 includes a learning information acquisition unit 13, a learning unit 14 in addition to the above-described earthquake information acquisition unit 11 and estimation processing unit 12. , A storage unit 15. FIG. 2 shows an example of the epicenter distance estimation device 10, and the learning information acquisition unit 13, the learning unit 14, and the storage unit 15 may be provided in a device other than the epicenter distance estimation device 10.

  The learning information acquisition unit 13 acquires waveform data as input data in learning by the learning unit 14 to be described later and an epicenter distance that is also correct data in learning, and inputs these to the learning unit 14. The source of the input data and the correct answer data is not particularly limited.

  The learning unit 14 uses the waveform data of the earthquake as input data, uses the epicenter distance of the earthquake as correct data, learns the relationship between the waveform data and the epicenter distance, and creates a learning model 16 showing the learning result. The learning unit 14 stores the generated learning model 16 in the storage unit 15.

  FIG. 3 is a diagram showing an example of input data and correct answer data used for learning in the first embodiment. FIG. 3 shows a plurality of waveform data having different epicenter distances. Each waveform data shown in FIG. 3 is waveform data of an earthquake observed in the past. The epicenter distance corresponding to each waveform data is the correct answer data. The learning unit 14 performs learning using the waveform data shown in FIG. 3 as input data and the epicenter distance as correct data.

Further, in the first embodiment, data published by the Japan Meteorological Agency can be used as the correct answer data. The data released by the Japan Meteorological Agency include the epicenter distance and the epicenter element of each observation point, and these are the measurement values (http: //www.data.jma.go .jp / svd / eqev / data / bulletin / deck.html). Further, in the first embodiment, the input data and correct answer data used for learning, and even better in which seismic intensity is calculated from the seismic is the set value (e.g., seismic intensity 4) or more.

  In the first embodiment, the learning unit 14 can construct a neural network by machine learning, for example, and use the neural network as the learning model 16. Specifically, in the hierarchical neural network including the input layer, the intermediate layer, and the output layer, the learning unit 14 uses the input data and the correct data to calculate the connection weight between the nodes adjacent to the layer. A learning model is generated by adjusting the value.

In the first embodiment, “learning” by the learning unit 14 means so-called “machine learning”. Further, by the learning section 14 "learning" is not limited to the deep learning using a neural network described above, learning using logistic regression, learning using support vector machines, learning using a decision tree, heterogeneous learning And so on.

  In the first embodiment, the earthquake information acquisition unit 11 receives the waveform data of the generated earthquake from a single earthquake detection device 20. Further, the earthquake information acquisition unit 11 sends the received waveform data to the estimation processing unit 12.

  In the first embodiment, the estimation processing unit 12 accesses the storage unit 15, acquires the learning model 16, and applies the waveform data transmitted from the learning information acquiring unit 13 to the acquired learning model 16. By doing so, the epicenter distance is estimated.

  Further, in the first embodiment, the learning unit 14 further learns the relationship between the waveform data, the epicenter distance and the depth of the epicenter using the epicenter depth of the earthquake as the correct answer data, and obtains a learning model. 16 can also be generated. In this case, the estimation processing unit 12 can also estimate the depth of the epicenter in addition to the epicenter distance.

  Further, in the first embodiment, the learning unit 14 can use, in addition to the waveform data, point data of a point where the waveform data is obtained as input data. In this case, the learning unit 14 generates a learning model 16 by learning the relationship between the waveform data and the point data, and the epicenter distance (or epicenter distance and epicenter depth).

  Here, the point where the waveform data is obtained is a point where the seismic wave that is the basis of the waveform data is observed. Examples of the spot data include surface ground amplification factor, data indicating the state of the plate, data indicating a volcano existing near the spot where the seismic wave was observed, crustal thickness, lithospheric thickness, and the like. As described above, if the learning model 16 is generated by using the waveform data and the point data as the input data, the accuracy of the estimation process can be improved.

  When the point data is used as the learning input data, the earthquake information acquisition unit 11 adds, in addition to the waveform data of the generated earthquake, the point data of the point where the waveform data is obtained, that is, the earthquake detection device 20. The location data of the location where is installed is also acquired.

  In addition, the point data may be stored in advance in the storage unit 15 for each of the earthquake detection devices 20, and in this embodiment, the earthquake information acquisition unit 11 reads from the storage unit 15 every time waveform data is acquired. Get the corresponding point data. Further, the point data may be transmitted together with the waveform data from the earthquake detecting device 20, and in this embodiment, the earthquake information acquisition unit 11 acquires the point data together with the waveform data.

  Further, when the point data is used as input data for learning, the estimation processing unit 12 applies the acquired waveform data and the point data to the learning model 16 generated by the learning unit 14 to calculate the epicenter distance. (Or epicenter distance and epicenter depth).

  In the first embodiment, the input data and the correct answer data are not limited to the above example. As input data, data other than waveform data and point data may be used. Further, data other than the epicenter distance and the depth of the epicenter may be used as the correct answer data.

[Device operation]
Next, the operation of the epicenter distance estimation device 10 according to the first embodiment will be described with reference to FIGS. In the following description, FIGS. 1 to 3 are appropriately referred to. In the first embodiment, the epicenter distance estimation method is performed by operating the epicenter distance estimation device 10. Therefore, the description of the epicenter distance estimation method according to the first embodiment will be replaced with the following description of the operation of the epicenter distance estimation device 10.

  In the first embodiment, the epicenter distance estimation device 10 mainly executes a learning process and an estimation process. First, the learning process will be described. FIG. 4 is a flowchart showing an operation of the epicenter distance estimating apparatus according to Embodiment 1 of the present invention when performing a learning process.

  As shown in FIG. 4, first, the learning information acquisition unit 13 acquires input data and correct answer data (step A1). Specifically, in step A1, the learning information acquisition unit 13 acquires the point data in addition to the waveform data as input data, and also acquires the depth of the epicenter as the correct answer data in addition to the epicenter distance.

  Next, the learning unit 14 determines whether the learning model 16 already exists (Step A2). Specifically, the learning unit 14 determines whether the learning model 16 is stored in the storage unit 15.

  As a result of the determination in step A2, if the learning model 16 does not yet exist, the learning unit 14 learns the relationship between the waveform data and the point data, the epicenter distance and the depth of the epicenter, and indicates the learning model indicating the learning result. 16 is newly generated (step A3).

  Specifically, in step A3, the learning unit 14 constructs a neural network by learning, and sets this as a learning model 16. The learning unit 14 stores the created learning model 16 in the storage unit 15.

  On the other hand, as a result of the determination in step A2, when the learning model 16 already exists, the learning unit 14 updates the existing learning model 16 using the input data and the correct data acquired in step A1. (Step A4). Specifically, the learning unit 14 updates the value of the connection weight between the nodes using the input data and the correct data acquired in step A1.

  By executing steps A1 to A4, a learning model is created or updated. After that, an estimation process is performed using the created or updated learning model. FIG. 5 is a flowchart showing an operation of the epicenter distance estimating apparatus according to Embodiment 1 of the present invention when performing estimation processing.

  As shown in FIG. 5, first, when the waveform data of the generated earthquake is transmitted from the earthquake detection device 20, the earthquake information acquisition unit 11 receives the transmitted waveform data (step B1).

  Next, the earthquake information acquiring unit 11 acquires, from the storage unit 15, point data of a point where the earthquake detecting device 20 that has transmitted the waveform data is installed (step B2). When the point data is transmitted together with the waveform data, the earthquake information acquisition unit 11 receives the transmitted point data.

  Next, the estimation processing unit 12 applies the waveform data received in step B1 and the point data obtained in step B2 to the learning model 16 created or updated by the learning processing illustrated in FIG. , The epicenter distance and the epicenter depth are estimated (step B3).

  By performing steps B1 to B3, the epicenter distance and the depth of the epicenter are estimated based on the waveform data acquired from the single earthquake detection device 20.

[Effects of First Embodiment]
As described above, according to the first embodiment, the epicenter distance and the epicenter depth are estimated from a single waveform data without fitting the waveform data to the function. In addition, since the estimation process is performed by the learning model 16, the epicenter distance and the epicenter depth are calculated stably and in a short time.

  In other words, in the first embodiment, unlike the conventional method disclosed in Patent Document 1, it is not necessary to manually reflect complicated individual point attributes to parameters whose effects are unknown. become. According to the first embodiment, it is possible to obtain information of an accuracy that can be used only by objective waveform data and machine learning. The epicenter distance estimation device 10 according to the first embodiment can be introduced to many points and many regions.

  Further, as described above, in the present embodiment, the depth of the epicenter can be estimated by a single waveform data, but the technique disclosed in Patent Document 1 cannot estimate the depth of the epicenter. When the technique disclosed in Patent Document 1 is used, measurement results of a plurality of seismometers are required for measuring the depth of the epicenter.

[Modification 1]
Next, a modification of the first embodiment will be described. First, in the first modification, the learning unit 14 generates the learning model 16 for each set waveform amount. Specifically, the waveform amount is represented by an elapsed time from the occurrence of the earthquake. Therefore, the learning unit 14 cuts out the waveform data corresponding to the elapsed time from the waveform data acquired by the learning information acquisition unit 13 for each set elapsed time, and performs learning using the extracted waveform data as input data. A learning model 16 is generated. Thereby, the learning model 16 is generated for each waveform amount.

  In the first modification, the estimation processing unit 12 calculates the waveform amount of the waveform data of the generated earthquake, and based on the calculated waveform amount, selects a learning model to be used from among a plurality of generated learning models 16. Select a model. Then, the estimation processing unit 12 applies the waveform data of the generated earthquake to the selected learning model 16 to estimate the epicenter distance (or the epicenter distance and the depth of the epicenter).

  In general, the estimation of the epicenter distance and the epicenter depth for an emergency earthquake bulletin is obtained even when the waveform amount of the waveform data is small. For this reason, the waveform data acquired by the earthquake information acquisition unit 11 is not necessarily constant, and the waveform amount of the waveform data used for generating the learning model and the waveform amount of the waveform data in the generated earthquake coincide. And the estimation accuracy may be reduced. However, according to the first modification, the learning model 16 is selected according to the waveform amount of the waveform data of the generated earthquake, so that the above-described decrease in estimation accuracy is avoided.

[Modification 2]
In the second modification, the learning unit 14 generates a learning model 16 for each observation point of waveform data serving as input data. Specifically, the learning unit 14 generates the learning model 16 for each earthquake detection device (for each seismometer) using only the waveform data acquired there.

  In the second modification, the estimation processing unit 12 specifies the observation point of the waveform data of the generated earthquake (that is, the earthquake detection device 20 that is the transmission source of the waveform data), and based on the specified observation point. The learning model to be used is selected from the plurality of generated learning models 16. Then, the estimation processing unit 12 applies the waveform data of the generated earthquake to the selected learning model 16 to estimate the epicenter distance (or the epicenter distance and the depth of the epicenter).

  According to the second modification, the estimation process can be performed in accordance with the characteristics of each observation point without performing learning using the characteristics of each observation point, that is, using the point data. In addition, at an observation point at which input data cannot be sufficiently secured, sufficient learning becomes difficult, so that it is difficult to generate a learning model for such an observation point.

[Modification 3]
In the third modification, the learning unit 14 generates the learning model 16 for each ground characteristic at the observation point of the waveform data serving as input data. Specifically, for example, it is assumed that observation points (earthquake detection devices 20) are grouped according to ground characteristics (point data values) such as a ground amplification factor. In this case, the learning unit 14 generates the learning model 16 for each group using only the waveform data obtained in the group.

  In the third modification, the estimation processing unit 12 specifies the ground characteristics of the observation point of the waveform data of the generated earthquake, and, based on the specified ground characteristics, from among the plurality of learning models 16 that have been generated, Select the learning model to use. Then, the estimation processing unit 12 applies the waveform data of the generated earthquake to the selected learning model 16 to estimate the epicenter distance (or the epicenter distance and the depth of the epicenter).

  According to the third modification, even if there is an observation point for which input data is not sufficiently secured, an estimation process that matches the characteristics of the observation point can be performed without performing learning using the point data. .

[program]
The program according to the first embodiment may be any program that causes a computer to execute steps A1 to A4 shown in FIG. 4 and steps B1 to B3 shown in FIG. By installing and executing this program on a computer, the epicenter distance estimation apparatus 10 and the epicenter distance estimation method according to the first embodiment can be realized. In this case, the CPU (Central Processing Unit) of the computer functions as the earthquake information acquisition unit 11, the estimation processing unit 12, the learning information acquisition unit 13, and the learning unit 14 and performs processing.

  Further, the program according to the first embodiment may be executed by a computer system configured by a plurality of computers. In this case, each computer may function as one of the earthquake information acquisition unit 11, the estimation processing unit 12, the learning information acquisition unit 13, and the learning unit 14, respectively. Further, storage unit 15 may be configured on a computer different from the computer that executes the program according to the present embodiment.

(Embodiment 2)
Next, an epicenter distance estimation device, an epicenter distance estimation method, and a program according to Embodiment 2 of the present invention will be described with reference to FIGS.

[Device configuration]
First, the configuration of the epicenter distance estimation device according to the second embodiment will be described with reference to FIG. FIG. 6 is a block diagram specifically showing a configuration of the epicenter distance estimation device according to Embodiment 2 of the present invention.

  As shown in FIG. 6, the epicenter distance estimating device 40 according to the second embodiment includes a waveform preprocessing unit 41. In this regard, the epicenter distance estimating device 40 according to the first embodiment shown in FIGS. It is different from the device 10. Hereinafter, the description will focus on the differences from the first embodiment.

  The waveform preprocessing unit 41 performs preprocessing on waveform data used as input data in the learning unit 14 and waveform data acquired in the earthquake information acquisition unit 11. Examples of the preprocessing include an image conversion process, an envelope conversion process, a bandpass conversion process, a differential conversion process, and a Fourier transform process.

  Specifically, the image conversion process is a process of converting waveform data into image data of an image for displaying the waveform data in a graph. According to the image conversion process, the learning unit 14 performs the learning based on the image data, so that the learning process is considered to be easy.

  The envelope conversion process is a process for making the waveform of the waveform data gentle. By executing the envelope processing, it becomes easy to specify the rising characteristics of the seismic wave, and thus the learning model 16 reflecting the rising characteristics of the seismic wave is generated.

  The bandpass conversion process is a process that makes a waveform of a specific cycle stand out. According to the band-pass conversion processing, the features of the seismic waves are emphasized, so that the learning model 16 reflecting the features of the seismic waves is generated.

  The differential conversion process is a process of differentiating the waveform data and converting the data into acceleration data. Also in the case of the differential transformation processing, the rise characteristic of the seismic wave can be easily specified, so that the learning model 16 reflecting the rise characteristic of the seismic wave is generated.

  Further, the Fourier transform process is a process for obtaining a frequency distribution of waveform data. According to the Fourier transform, the difference in the cycle of each waveform data is emphasized, so that the learning model 16 reflecting the cycle of the seismic wave is generated.

  The waveform preprocessing unit 41 can execute any one or more of image conversion processing, envelope conversion processing, bandpass conversion processing, differential conversion processing, and Fourier transform processing.

[Device operation]
Next, the operation of the epicenter distance estimation device 40 according to the second embodiment will be described with reference to FIGS. In the following description, FIGS. 1 to 6 are appropriately referred to. In the second embodiment, the epicenter distance estimation method is performed by operating the epicenter distance estimation device 40. Therefore, the description of the epicenter distance estimation method in the second embodiment will be replaced with the following description of the operation of the epicenter distance estimation device 40.

  First, the learning process will be described. FIG. 7 is a flowchart showing an operation of the epicenter distance estimating apparatus according to Embodiment 2 of the present invention when performing a learning process.

  As shown in FIG. 7, first, the learning information acquisition unit 13 acquires input data and correct answer data (step A11). Further, the learning information acquisition unit 13 inputs the acquired data to the waveform preprocessing unit 41.

  Next, the waveform pre-processing unit 41 performs pre-processing on the waveform data included in the input data acquired in step A11 (step A12). Then, the waveform preprocessing unit 41 inputs the preprocessed waveform data, other input data (point data), and the correct answer data to the learning unit 14.

  Next, the learning unit 14 determines whether the learning model 16 already exists (Step A13).

  As a result of the determination in step A13, if the learning model 16 does not yet exist, the learning unit 14 learns the relationship between the waveform data and the point data, the epicenter distance and the depth of the epicenter, and indicates the learning model indicating the learning result. 16 is newly generated (step A14).

  On the other hand, if the result of the determination in step A13 is that the learning model 16 already exists, the learning unit 14 updates the existing learning model 16 using the input data and the correct data (step A15).

  By executing steps A11 to A15, the learning model 16 is created or updated. After that, an estimation process is performed using the created or updated learning model 16. FIG. 8 is a flowchart showing an operation of the epicenter distance estimating apparatus according to Embodiment 2 of the present invention when performing estimation processing.

  As shown in FIG. 8, first, when the waveform data of the generated earthquake is transmitted from the earthquake detection device 20, the earthquake information acquisition unit 11 receives the transmitted waveform data (step B11).

  Next, the earthquake information acquisition unit 11 acquires, from the storage unit 15, point data of a point where the earthquake detection device 20 that has transmitted the waveform data is installed (step B12). When the point data is transmitted together with the waveform data, the earthquake information acquisition unit 11 receives the transmitted point data.

  Next, the waveform pre-processing unit 41 performs pre-processing on the waveform data received in step B11 (step B13). Then, the waveform preprocessing unit 41 inputs the preprocessed waveform data and the point data to the estimation processing unit 12.

  Next, the estimation processing unit 12 applies the waveform data after the pre-processing in step B13 and the point data acquired in step B12 to the learning model 16 created or updated by the learning processing illustrated in FIG. Then, the epicenter distance and the depth of the epicenter are estimated (step B14).

  By executing steps B11 to B14, also in the second embodiment, the epicenter distance and the depth of the epicenter are estimated based on the waveform data acquired from the single earthquake detection device 20, as in the first embodiment. Will be done.

[Effects of Second Embodiment]
As described above, in the second embodiment, the preprocessing performed by the waveform preprocessing unit 41 suppresses noise and makes features appear in waveform data used for learning. Therefore, according to the second embodiment, the accuracy of the learning model is improved, and as a result, the estimation accuracy is also improved.

(Physical configuration)
Here, a computer that realizes the epicenter distance estimating apparatus by executing the programs in the first and second embodiments will be described with reference to FIG. FIG. 9 is a block diagram illustrating an example of a computer that realizes the epicenter distance estimation device according to the first and second embodiments of the present invention.

  As shown in FIG. 9, the computer 110 includes a CPU 111, a main memory 112, a storage device 113, an input interface 114, a display controller 115, a data reader / writer 116, and a communication interface 117. These units are connected via a bus 121 so as to be able to perform data communication with each other.

  The CPU 111 expands the programs (codes) according to the present embodiment stored in the storage device 113 in the main memory 112 and executes them in a predetermined order to perform various operations. The main memory 112 is typically a volatile storage device such as a DRAM (Dynamic Random Access Memory). Further, the program according to the present embodiment is provided in a state stored in computer-readable recording medium 120. Note that the program in the present embodiment may be distributed on the Internet connected via the communication interface 117.

  Specific examples of the storage device 113 include a semiconductor storage device such as a flash memory in addition to a hard disk drive. The input interface 114 mediates data transmission between the CPU 111 and input devices 118 such as a keyboard and a mouse. The display controller 115 is connected to the display device 119 and controls display on the display device 119.

  The data reader / writer 116 mediates data transmission between the CPU 111 and the recording medium 120, executes reading of a program from the recording medium 120, and writing of a processing result in the computer 110 to the recording medium 120. The communication interface 117 mediates data transmission between the CPU 111 and another computer.

  Further, specific examples of the recording medium 120 include a general-purpose semiconductor recording device such as CF (Compact Flash (registered trademark)) and SD (Secure Digital), a magnetic recording medium such as a flexible disk (Flexible Disk), or a CD-ROM. An optical recording medium such as a ROM (Compact Disk Read Only Memory) can be used.

  Further, the epicenter distance estimation devices 10 and 40 in the first and second embodiments can be realized by using hardware corresponding to each unit instead of a computer in which a program is installed. Furthermore, the epicenter distance estimation devices 10 and 40 may be partially realized by a program and the remaining part may be realized by hardware.

Some or all of the above-described embodiments can be expressed by the following (Appendix 1) to (Appendix 30 ), but are not limited to the following description.

(Appendix 1)
An earthquake information acquisition unit for acquiring waveform data of the generated earthquake;
An estimation processing unit that estimates the epicenter distance by applying the acquired waveform data to a learning model obtained by learning the relationship between the waveform data of the earthquake and the epicenter distance,
An epicenter distance estimation device, comprising:

(Appendix 2)
Further comprising a learning unit for learning the relationship between the waveform data and the epicenter distance using the waveform data of the earthquake as input data and learning the relationship between the epicenter distance of the earthquake as the correct answer data and generating a learning model indicating the learning result;
The estimation processing unit, by applying the acquired waveform data to the learning model generated by the learning unit, to estimate the epicenter distance,
The epicentral distance estimation device according to supplementary note 1.

(Appendix 3)
The learning unit learns the relationship between the waveform data and the point data and the epicenter distance, using the point data of the point where the waveform data is obtained as input data in addition to the waveform data, and Produces
The earthquake information acquisition unit, in addition to the waveform data, also acquires the point data of the point where the waveform data of the generated earthquake was obtained,
The estimation processing unit, for the learning model, in addition to the waveform data, also applies the acquired point data, to estimate the epicenter distance,
The epicentral distance estimation device according to attachment 2.

(Appendix 4)
The learning processing unit further learns a relationship between the waveform data, the epicenter distance, and the depth of the epicenter, further using the epicenter depth of the earthquake as the correct answer data, and generates the learning model,
The estimation processing unit estimates the depth of the epicenter in addition to the epicenter distance,
The epicentral distance estimation device according to Supplementary Note 2 or 3.

(Appendix 5)
The learning processing unit constructs a neural network by learning, and uses the neural network as a learning model,
The epicentral distance estimation device according to any one of supplementary notes 2 to 4.

(Appendix 6)
The learning unit generates the learning model for each waveform amount of waveform data serving as input data,
The estimation processing unit calculates the waveform amount of the acquired waveform data, and based on the calculated waveform amount, from among the generated learning models, selects the learning model to be used, and selects the selected learning model. Applying the acquired waveform data to a learning model to estimate the epicenter distance,
The epicenter distance estimating apparatus according to any one of supplementary notes 2 to 5.

(Appendix 7)
The learning unit generates the learning model for each observation point of waveform data serving as input data,
The estimation processing unit specifies an observation point of the acquired waveform data, and based on the specified observation point, from among the generated learning models, selects the learning model to be used, and selects the selected learning model. Applying the acquired waveform data to a learning model to estimate the epicenter distance,
The epicenter distance estimating apparatus according to any one of supplementary notes 2 to 5.

(Appendix 8)
The learning unit generates the learning model for each ground characteristic at an observation point of waveform data serving as input data,
The estimation processing unit specifies the ground characteristics of the observation point of the acquired waveform data, and based on the specified ground characteristics, from among the generated learning models, selects the learning model to be used, To the selected learning model, applying the obtained waveform data, to estimate the epicenter distance,
The epicenter distance estimating apparatus according to any one of supplementary notes 2 to 5.

(Appendix 9)
The learning unit further includes a waveform preprocessing unit that performs preprocessing on the waveform data used as the input data and the waveform data acquired by the earthquake information acquisition unit.
The epicentral distance estimation device according to any one of supplementary notes 1 to 8.

(Appendix 10)
The waveform preprocessing unit performs at least one of image conversion processing, envelope conversion processing, bandpass conversion processing, differential conversion processing, and Fourier transform processing as the preprocessing.
The epicentral distance estimation device according to attachment 9.

(Appendix 11)
(A) acquiring waveform data of the earthquake that occurred;
(B) estimating the epicenter distance by applying the acquired waveform data to a learning model obtained by learning the relationship between the waveform data of the earthquake and the epicenter distance;
An epicenter distance estimation method, comprising:

(Appendix 12)
(C) using the waveform data of the earthquake as input data, learning the relationship between the waveform data and the epicenter distance using the epicenter distance of the earthquake as correct data, and generating a learning model indicating the learning result,
In the step (b), the acquired waveform data is applied to the learning model generated in the step (c) to estimate an epicenter distance.
An epicenter distance estimation method according to appendix 11.

(Appendix 13)
In the step (c), in addition to the waveform data, point data of a point where the waveform data is obtained is also used as input data to learn the relationship between the waveform data and the point data and the epicenter distance, Generating the learning model,
In the step (a), in addition to the waveform data, the location data of the location where the waveform data of the occurred earthquake is obtained,
(B) estimating the epicenter distance by applying the acquired point data to the learning model in addition to the waveform data;
An epicenter distance estimation method according to attachment 12.

(Appendix 14)
In the step (c), the learning model is generated by further learning the relationship between the waveform data and the epicenter distance and the depth of the epicenter using the epicenter depth of the earthquake as the correct answer data. And
In the step (b), the depth of the epicenter is estimated in addition to the epicenter distance.
An epicenter distance estimation method according to appendix 12 or 13.

(Appendix 15)
In the step (c), a neural network is constructed by learning, and the neural network is used as a learning model.
An epicenter distance estimation method according to any one of supplementary notes 12 to 14.

(Appendix 16)
In the step (c), the learning model is generated for each waveform amount of waveform data as input data,
In the step (b), a waveform amount of the acquired waveform data is calculated, and the learning model to be used is selected from each of the generated learning models based on the calculated waveform amount. Applying the acquired waveform data to the learning model, and estimating the epicenter distance,
An epicenter distance estimation method according to any one of supplementary notes 12 to 15.

(Appendix 17)
In the step (c), the learning model is generated for each observation point of the waveform data serving as input data,
In the step (b), an observation point of the acquired waveform data is specified, and the learning model to be used is selected from each of the generated learning models based on the specified observation point. Applying the acquired waveform data to the learning model, and estimating the epicenter distance,
An epicenter distance estimation method according to any one of supplementary notes 12 to 15.

(Appendix 18)
In the step (c), the learning model is generated for each ground characteristic at an observation point of waveform data serving as input data,
In the step (b), the ground characteristics of the observation point of the acquired waveform data are specified, and the learning model to be used is selected from the generated learning models based on the specified ground characteristics. And applying the acquired waveform data to the selected learning model to estimate the epicenter distance;
An epicenter distance estimation method according to any one of supplementary notes 12 to 15.

(Appendix 19)
(D) performing a pre-process on the waveform data used as the input data in the step (c) and the waveform data acquired in the step (a).
An epicenter distance estimation method according to any one of supplementary notes 11 to 18.

(Appendix 20)
In the step (d), at least one of an image conversion process, an envelope conversion process, a bandpass conversion process, a differential conversion process, and a Fourier transform process is executed as the preprocessing.
An epicenter distance estimation method according to attachment 19.

(Appendix 21)
On the computer,
(A) acquiring waveform data of the earthquake that occurred;
(B) estimating the epicenter distance by applying the acquired waveform data to a learning model obtained by learning the relationship between the waveform data of the earthquake and the epicenter distance;
Ru is the execution, program.

(Appendix 22)
On the computer,
(C) using the waveform data of the earthquake as input data, learning the relationship between the waveform data and the epicenter distance using the epicenter distance of the earthquake as correct data, and generating a learning model indicating the learning result, further executing
In the step (b), the acquired waveform data is applied to the learning model generated in the step (c) to estimate an epicenter distance.
The program according to supplementary note 21.

(Appendix 23)
In the step (c), in addition to the waveform data, point data of a point where the waveform data is obtained is also used as input data to learn the relationship between the waveform data and the point data and the epicenter distance, Generating the learning model,
In the step (a), in addition to the waveform data, the location data of the location where the waveform data of the occurred earthquake is obtained,
(B) estimating the epicenter distance by applying the acquired point data to the learning model in addition to the waveform data;
The program according to supplementary note 22.

(Appendix 24)
In the step (c), the learning model is generated by further learning the relationship between the waveform data and the epicenter distance and the depth of the epicenter using the epicenter depth of the earthquake as the correct answer data. And
In the step (b), the depth of the epicenter is estimated in addition to the epicenter distance.
The program according to Supplementary Note 22 or 23.

(Appendix 25)
In the step (c), a neural network is constructed by learning, and the neural network is used as a learning model.
The program according to any one of supplementary notes 22 to 24.

(Supplementary Note 26)
In the step (c), the learning model is generated for each waveform amount of waveform data as input data,
In the step (b), a waveform amount of the acquired waveform data is calculated, and the learning model to be used is selected from each of the generated learning models based on the calculated waveform amount. Applying the acquired waveform data to the learning model, and estimating the epicenter distance,
The program according to any one of supplementary notes 22 to 25.

(Appendix 27)
In the step (c), the learning model is generated for each observation point of the waveform data serving as input data,
In the step (b), an observation point of the acquired waveform data is specified, and the learning model to be used is selected from each of the generated learning models based on the specified observation point. Applying the acquired waveform data to the learning model, and estimating the epicenter distance,
The program according to any one of supplementary notes 22 to 25.

(Appendix 28)
In the step (c), the learning model is generated for each ground characteristic at an observation point of waveform data serving as input data,
In the step (b), the ground characteristics of the observation point of the acquired waveform data are specified, and the learning model to be used is selected from the generated learning models based on the specified ground characteristics. And applying the acquired waveform data to the selected learning model to estimate the epicenter distance;
The program according to any one of supplementary notes 22 to 25.

(Appendix 29)
On the computer,
(D) executing pre-processing on the waveform data used as the input data in the step (c) and the waveform data acquired in the step (a).
The program according to any one of supplementary notes 21 to 28.

(Appendix 30)
In the step (d), at least one of an image conversion process, an envelope conversion process, a bandpass conversion process, a differential conversion process, and a Fourier transform process is executed as the preprocessing.
The program according to supplementary note 29.

  Although the present invention has been described with reference to the exemplary embodiments, the present invention is not limited to the above exemplary embodiments. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.

  This application claims priority based on Japanese Patent Application No. 2016-136310 filed on Jul. 8, 2016, the disclosure of which is incorporated herein in its entirety.

  As described above, according to the present invention, the calculation of the epicenter distance can be performed stably, and the calculation time can be reduced. INDUSTRIAL APPLICABILITY The present invention is useful for a system that needs to distribute information about an earthquake as soon as possible when an earthquake occurs.

10 Epicenter distance estimation device (first embodiment)
DESCRIPTION OF SYMBOLS 11 Earthquake information acquisition part 12 Estimation processing part 13 Learning information acquisition part 14 Learning part 15 Storage part 16 Learning model 20 Earthquake detection device 30 Integrated monitoring system for seismic activity 40 Epicenter distance estimation device (Embodiment 2)
41 Waveform preprocessing unit 110 Computer 111 CPU
112 Main memory 113 Storage device 114 Input interface 115 Display controller 116 Data reader / writer 117 Communication interface 118 Input device 119 Display device 120 Recording medium 121 Bus

Claims (10)

A learning unit for inputting the waveform data of the earthquake as input data, learning the relationship between the waveform data and the epicenter distance as the correct answer data of the epicenter distance of the earthquake, and generating a learning model indicating a learning result;
An earthquake information acquisition unit for acquiring waveform data of the generated earthquake;
An estimation processing unit that estimates the epicenter distance by applying the acquired waveform data to a learning model obtained by learning the relationship between the waveform data of the earthquake and the epicenter distance by the learning unit,
Equipped with a,
The learning unit generates the learning model for each waveform amount of waveform data serving as input data,
The estimation processing unit calculates the waveform amount of the acquired waveform data, and based on the calculated waveform amount, from among the generated learning models, selects the learning model to be used, and selects the selected learning model. Applying the acquired waveform data to a learning model to estimate the epicenter distance,
An epicenter distance estimation device characterized by the following: The learning unit learns the relationship between the waveform data and the point data and the epicenter distance, using the point data of the point where the waveform data is obtained as input data in addition to the waveform data, and Produces
The earthquake information acquisition unit, in addition to the waveform data, also acquires the point data of the point where the waveform data of the generated earthquake was obtained,
The estimation processing unit, for the learning model, in addition to the waveform data, also applies the acquired point data, to estimate the epicenter distance,
The epicenter distance estimation device according to claim 1 . The learning processing unit further learns a relationship between the waveform data, the epicenter distance, and the depth of the epicenter, further using the epicenter depth of the earthquake as the correct answer data, and generates the learning model,
The estimation processing unit estimates the depth of the epicenter in addition to the epicenter distance,
The epicenter distance estimation device according to claim 1 . The learning processing unit constructs a neural network by learning, and uses the neural network as a learning model,
The epicenter distance estimation device according to claim 1. The learning unit generates the learning model for each observation point of waveform data serving as input data,
The estimation processing unit specifies an observation point of the acquired waveform data, and based on the specified observation point, from among the generated learning models, selects the learning model to be used, and selects the selected learning model. Applying the acquired waveform data to a learning model to estimate the epicenter distance,
Epicentral distance estimation apparatus according to claim 1. The learning unit generates the learning model for each ground characteristic at an observation point of waveform data serving as input data,
The estimation processing unit specifies the ground characteristics of the observation point of the acquired waveform data, and based on the specified ground characteristics, from among the generated learning models, selects the learning model to be used, To the selected learning model, applying the obtained waveform data, to estimate the epicenter distance,
Epicentral distance estimation apparatus according to claim 1. The learning unit further includes a waveform preprocessing unit that performs preprocessing on the waveform data used as the input data and the waveform data acquired by the earthquake information acquisition unit.
An epicenter distance estimation device according to any one of claims 1 to 6 . The waveform preprocessing unit performs at least one of image conversion processing, envelope conversion processing, bandpass conversion processing, differential conversion processing, and Fourier transform processing as the preprocessing.
An epicenter distance estimation device according to claim 7 . (A) using the waveform data of the earthquake as input data, learning the relationship between the waveform data and the epicenter distance using the epicenter distance of the earthquake as correct data, and generating a learning model indicating the learning result;
( B ) acquiring waveform data of the generated earthquake;
( C ) estimating the epicenter distance by applying the acquired waveform data to a learning model obtained by learning the relationship between the waveform data of the earthquake and the epicenter distance in the step (a). ,
Have a,
In the step (a), the learning model is generated for each waveform amount of waveform data serving as input data;
In the step (c), a waveform amount of the acquired waveform data is calculated, and the learning model to be used is selected from each of the generated learning models based on the calculated waveform amount. Applying the acquired waveform data to the learning model, and estimating the epicenter distance,
An epicenter distance estimation method characterized by the following: On the computer,
(A) using the waveform data of the earthquake as input data, learning the relationship between the waveform data and the epicenter distance using the epicenter distance of the earthquake as correct data, and generating a learning model indicating the learning result;
( B ) acquiring waveform data of the generated earthquake;
( C ) estimating the epicenter distance by applying the acquired waveform data to a learning model obtained by learning the relationship between the waveform data of the earthquake and the epicenter distance in the step (a). ,
Was executed,
In the step (a), the learning model is generated for each waveform amount of waveform data serving as input data;
In the step (c), a waveform amount of the acquired waveform data is calculated, and the learning model to be used is selected from each of the generated learning models based on the calculated waveform amount. Applying the acquired waveform data to the learning model, and estimating the epicenter distance,
program.

Images