Marine earthquake positioning method, device and equipment and readable storage medium

Abstract

The invention provides a marine earthquake positioning method, device and equipment and a readable storage medium, and relates to the technical field of earthquake prediction, and the method comprises the steps: obtaining a plurality of historical earthquake data sets; acquiring a real-time seismic data set; performing difference preprocessing on the plurality of historical seismic data sets and the real-time seismic data set to obtain a plurality of target seismic data sets and a prediction seismic data set; training a preset decision tree model based on the target seismic data set, and when a preset model index meets a set condition, stopping training to obtain a target positioning model; and inputting the predicted seismic data set into the target positioning model to obtain a target longitude and a target latitude in the target epicenter. According to the invention, the decision tree model is constructed to learn the relationship between the latitude and longitude difference between the seismic measurement table and the epicenter in the historical data and the time difference between the seismic measurement table and the epicenter trigger, so that the epicenter position at the current moment is predicted, and compared with the prior art, the marine seismic epicenter can be positioned more accurately and quickly.

Description

Technical Field

[0001] The present invention relates to the technical field of earthquake prediction, and more particularly, to a method, device, equipment and readable storage medium for locating marine earthquakes. Background Art

[0002] Epicenter location is an important part of an earthquake early warning system. The traditional epicenter location method achieves the location effect by constructing a Voronoi diagram. This method often requires seismic waves to pass through multiple seismic measurement stations to obtain a relatively accurate location. Moreover, if the epicenter is far from the seismic measurement stations, a large error will occur. Since the occurrence location of marine earthquakes is relatively far from the seismic measurement stations on land, it is difficult to obtain the accurate location of the epicenter through traditional location algorithms. Summary of the Invention

[0003] The purpose of the present invention is to provide a method, device, equipment and readable storage medium for locating marine earthquakes to improve the above problems. To achieve the above purpose, the technical solutions adopted by the present invention are as follows:

[0004] In a first aspect, the present application provides a method for locating marine earthquakes, including:

[0005] Obtain a plurality of historical earthquake data sets, where the historical earthquake data sets include information of multiple seismic measurement stations that trigger alarms and epicenter information. The information of the seismic measurement stations that trigger alarms includes the longitude, latitude and trigger time of the measurement station, and the epicenter information includes the epicenter longitude and epicenter latitude;

[0006] Obtain a real-time earthquake data set, where the real-time earthquake data set includes information of the seismic measurement station that triggers an alarm;

[0007] Perform difference preprocessing on the plurality of historical earthquake data sets and the real-time earthquake data set to obtain a plurality of target earthquake data sets and a predicted earthquake data set. The difference preprocessing is used to represent the calculation of differences on the information in the historical earthquake data sets;

[0008] Train a preset decision tree model based on the target earthquake data set. When the preset model index meets the set conditions, stop training to obtain a target location model;

[0009] Input the predicted earthquake data set into the target location model to obtain the target longitude and target latitude of the target epicenter.

[0010] In a second aspect, the present application also provides a device for locating marine earthquakes, including:

[0011] A first acquisition unit, configured to acquire a plurality of historical earthquake data sets, where the historical earthquake data sets include information of a plurality of seismic measurement stations that trigger alarms and epicenter information, the information of the seismic measurement stations that trigger alarms includes the longitude of the measurement station, the latitude of the measurement station, and the triggering time of the measurement station, and the epicenter information includes the longitude of the epicenter and the latitude of the epicenter;

[0012] A second acquisition unit, configured to acquire a real-time earthquake data set, where the real-time earthquake data set includes information of a seismic measurement station that triggers an alarm;

[0013] A preprocessing unit, configured to perform difference preprocessing on the plurality of historical earthquake data sets and the real-time earthquake data set to obtain a plurality of target earthquake data sets and a predicted earthquake data set, where the difference preprocessing is used to represent performing difference calculation on the information in the historical earthquake data sets;

[0014] A first training unit, configured to train a preset decision tree model based on the target earthquake data set, and stop training when the preset model metrics meet the set conditions to obtain a target positioning model;

[0015] An input unit, configured to input the predicted earthquake data set into the target positioning model to obtain the target longitude and target latitude of the target epicenter.

[0016] In a third aspect, the present application further provides a positioning device for marine earthquakes, including:

[0017] A memory, configured to store a computer program;

[0018] A processor, configured to implement the steps of the positioning method for marine earthquakes when executing the computer program.

[0019] In a fourth aspect, the present application further provides a readable storage medium, where a computer program is stored on the readable storage medium, and when the computer program is executed by a processor, the steps of the above-mentioned positioning method for marine earthquakes are implemented.

[0020] The beneficial effects of the present invention are as follows:

[0021] The present invention constructs a decision tree model to learn the relationship between the longitude and latitude differences between the seismic measurement stations and the epicenter in historical data, and the triggering time difference between the seismic measurement stations and the epicenter, so as to predict the epicenter position at the current moment. Compared with the prior art, it can more accurately and quickly locate the epicenter of marine earthquakes.

[0022] Other features and advantages of the present invention will be described in the subsequent description, and part of them will become obvious from the description, or will be understood by implementing the embodiments of the present invention. The objectives and other advantages of the present invention can be realized and obtained by the structures specifically pointed out in the written description, claims, and drawings. Brief Description of the Drawings

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0024] Figure 1 It is a schematic flow chart of the method for positioning marine earthquakes described in the embodiments of the present invention;

[0025] Figure 2 It is a schematic structural diagram of the device for positioning marine earthquakes described in the embodiments of the present invention;

[0026] Figure 3 It is a schematic diagram of earthquake triggering described in the embodiments of the present invention;

[0027] Figure 4 It is a schematic structural diagram of the equipment for positioning marine earthquakes described in the embodiments of the present invention.

[0028] Reference signs in the figures:

[0029] 10. First training unit; 20. Second training unit; 30. Preprocessing unit; 40. First training unit; 50. Input unit;

[0030] 800. Equipment for positioning marine earthquakes; 801. Processor; 802. Memory; 803. Multimedia component; 804. I / O interface; 805. Communication component. Detailed Embodiments

[0031] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Generally, the components of the embodiments of the present invention described and illustrated in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed present invention, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

[0032] It should be noted that similar reference numerals and letters denote similar items in the following figures. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. At the same time, in the description of the present invention, terms such as "first" and "second" are only used for distinguishing descriptions and cannot be construed as indicating or implying relative importance.

[0033] Embodiment 1:

[0034] This embodiment provides a method for locating marine earthquakes.

[0035] See Figure 1 , the figure shows that this method includes step S10, step S20, step S30, step S40, and step S50.

[0036] Step S10. Obtain multiple historical earthquake data sets. The historical earthquake data sets include information of multiple seismic measurement stations that trigger alarms and epicenter information. The information of the seismic measurement stations that trigger alarms includes the longitude of the measurement station, the latitude of the measurement station, and the trigger time of the measurement station. The epicenter information includes the longitude of the epicenter and the latitude of the epicenter;

[0037] Step S20. Obtain real-time earthquake data sets. The real-time earthquake data sets include information of seismic measurement stations that trigger alarms;

[0038] Specifically, all earthquake data within a historical time period can be obtained through official channels such as the China Earthquake Networks Center. Usually, earthquakes in the sea area around the area to be predicted and earthquakes in surrounding countries that may affect the area to be predicted are selected to ensure a certain correlation between historical earthquake data and real-time predicted earthquake data. Earthquake data sets with a magnitude less than or equal to zero are deleted from all historical earthquake data, and earthquake events with less than 5 seismic measurement station responses after the earthquake occurrence are also deleted to reduce the interference of collected error data on the historical earthquake data sets. Earthquake data with a trigger time difference greater than 30 seconds between any responding seismic measurement station and the first responding seismic measurement station also needs to be deleted. The remaining earthquake data is used as the historical earthquake data sets for the training and verification of the positioning model.

[0039] Step S30. Perform difference preprocessing on multiple historical earthquake data sets and real-time earthquake data sets to obtain multiple target earthquake data sets and predicted earthquake data sets. The difference preprocessing is used to represent the difference calculation of the information in the historical earthquake data sets;

[0040] Specifically, considering that the epicenter position of a marine earthquake is far from the seismic measurement stations set on land, the epicenter position information cannot be accurately obtained through the existing positioning of seismic measurement stations. The correlation between the longitude and latitude differences between the seismic measurement stations and the epicenter, as well as the time difference between the trigger time of the seismic measurement stations and the epicenter time, needs to be considered to accurately locate the epicenter position in the ocean.

[0041] Specifically, step S30 specifically includes steps S31, S32, S33, S34, S35, S36, and S37:

[0042] Step S30. Compare the trigger times of multiple measurement stations, and use the measurement station corresponding to the minimum measurement station trigger time as the target measurement station;

[0043] Step S30. Calculate the trigger time differences between the remaining measurement stations and the target measurement station to obtain multiple measurement station trigger time differences;

[0044] Step S30. Calculate the longitude differences between the remaining measurement stations and the target measurement station to obtain multiple measurement station longitude differences;

[0045] Step S30. Calculate the latitude differences between the remaining measurement stations and the target measurement station to obtain multiple measurement station latitude differences;

[0046] Step S30. Calculate the difference between the epicenter longitude and the longitude of the target measurement station to obtain the epicenter longitude difference;

[0047] Step S30. Calculate the difference between the epicenter latitude and the latitude of the target measurement station to obtain the epicenter latitude difference;

[0048] Step S30. Use the multiple measurement station trigger time differences, multiple measurement station longitude differences, multiple measurement station latitude differences, epicenter longitude difference, and epicenter latitude difference as the target earthquake dataset.

[0049] Specifically, after an earthquake occurs, record the response time of the first earthquake measurement station as T1, the response time of the second earthquake measurement station as T2, the response time of the third earthquake measurement station as T3, and so on. Using the first earthquake measurement station that triggers an alarm as the standard, calculate the time taken from the first earthquake measurement station to the second earthquake measurement station as t1 = T2 - T1, and the time differences from the subsequent earthquake measurement stations to the first earthquake measurement station can be obtained. As Figure 2 shown, it is a schematic diagram of earthquake triggering. Five earthquake measurement stations are marked in the figure. Among them, t1 is the difference between the trigger time of the second earthquake measurement station and the trigger time of the first earthquake measurement station, t2 is the difference between the trigger time of the third earthquake measurement station and the trigger time of the first earthquake measurement station, t3 is the difference between the trigger time of the fourth earthquake measurement station and the trigger time of the first earthquake measurement station, and t4 is the difference between the trigger time of the fifth earthquake measurement station and the trigger time of the first earthquake measurement station.

[0050] Step S40. Train a preset decision tree model based on the target earthquake dataset. When the preset model metrics meet the set conditions, stop training to obtain the target positioning model;

[0051] Specifically, this application uses a decision tree (LightGBM) model. The specific advantages of this model are as follows: The model uses the leaf splitting algorithm as the tree growth strategy. This strategy finds the leaf with the largest splitting gain from all current leaves each time, and then splits it. This process is repeated in a loop. The leaf splitting algorithm produces less error and higher positioning accuracy than the horizontal splitting algorithm. The model also uses the histogram algorithm. When constructing the tree, histogram information is used to calculate the splitting gain and quickly select the best splitting point, effectively reducing the amount of calculation. The model also uses the gradient-based unilateral sampling algorithm. The unilateral sampling algorithm excludes most samples with small gradients and only uses the remaining samples to calculate the information gain, which can reduce the data volume while ensuring the model positioning accuracy and effectively reducing the workload of calculation. The model also uses the mutually exclusive feature bundling algorithm, which bundles multiple features to form new features, effectively improving the model calculation speed.

[0052] Specifically, step S40 specifically includes steps S41, S42, S43, S44, S45, S46, S47, S48, S49, S410, S411, and S412:

[0053] Step S41. Divide multiple target earthquake data sets into training samples, test samples, and validation samples respectively;

[0054] Step S42. Obtain the initial range of each target parameter;

[0055] Step S43. Generation operation: Based on the initial range of each target parameter, randomly generate multiple initial parameters;

[0056] Step S44. Selection operation: Randomly select one initial parameter from the multiple initial parameters of each target parameter for combination to obtain multiple parameter combinations. Among them, at most two initial parameter values are the same in any two parameter combinations;

[0057] Step S45. Adjustment operation: Adjust the decision tree model based on each parameter combination to obtain multiple initial positioning models;

[0058] Step S46. Training operation: Bring the training samples into multiple initial positioning models for training respectively, and calculate the model metrics to obtain multiple initial model metrics;

[0059] Specifically, the grid search method can adjust three important parameter values in the LightGBM model. By setting the value range intervals of these three parameters, the grid search will search for the optimal parameters of the model within the selected interval. The final grid search result will select the values of the three parameters corresponding to the largest evaluation index value within the selected interval, so as to obtain the target positioning model based on the values of these three parameters.

[0060] This application focuses on optimizing the following three parameters: learning rate, which determines the magnitude of model parameter updates in each iteration. A larger learning rate will cause the model parameters to be updated significantly in each iteration, while a smaller learning rate will result in a reduced magnitude of parameter updates; maximum depth, which can be used to limit the depth of each tree, avoiding overfitting the details of the training data, thereby reducing the risk of overfitting. Limiting the tree depth can reduce the complexity of the model and improve the generalization ability of the model, which is particularly important when dealing with high-dimensional data. Deeper trees usually pay more attention to capturing the details of the training data, which may lead to a better fitting effect of the model on the training set, but also increases the sensitivity of the model to noise; number of iterations, increasing the number of iterations usually improves the performance of the model, especially for problems with complex objective functions. However, increasing the number of iterations may also lead to an increased risk of overfitting of the model on the training set but poor performance on the test set. At the same time, a larger number of iterations will also make the training time of the model longer and the training efficiency of the model lower.

[0061] Specifically, step S46 specifically includes steps S461, S462, S463, S464, S465, S466, S467, S468, S469, and S4610:

[0062] Step S461. Divide multiple training samples into multiple sets respectively to obtain multiple training sets;

[0063] Step S462. Determine an operation: randomly determine any one set from the multiple training sets as the target set;

[0064] Step S463. Substitute operation: substitute all training sets except the target set into the decision tree model for training to obtain a trained positioning model;

[0065] Step S464. First calculation operation: calculate the model metrics corresponding to each training set except the target set based on the trained positioning model to obtain multiple first metrics. The model metrics are used to characterize the fitting degree of the trained positioning model to the training set;

[0066] Step S465. Prediction operation: substitute the target set into the trained positioning model for prediction, and calculate the model metrics based on the prediction results to obtain a second metric;

[0067] Step S476. Second calculation operation: perform a summation calculation on all the first metrics to obtain a first summation result;

[0068] Step S467. Third calculation operation: perform a summation calculation on the first summation result and the second metric to obtain a second summation result;

[0069] Step S468. Fourth calculation operation: Calculate the ratio of the second summation result to the number of training sets to obtain the target ratio;

[0070] Step S469. Repeat the determination operation, substitution operation, first calculation operation, prediction operation, second calculation operation, third calculation operation, and fourth calculation operation until the target ratio of the number of training sets is obtained;

[0071] Step S4610. Sum up multiple target ratios and divide the summation result by the number of training sets to obtain the initial model metric;

[0072] Specifically, in this application, the training samples are divided into three parts, and triple cross-validation is used to bring the training set into model training and learning, so as to ensure that the trained model can better learn all the features in the training samples.

[0073] Step S47. Screening operation: Screen out the optimal parameter combination according to all the initial model metrics. The optimal parameter combination is the parameter combination corresponding to the best initial positioning model, and the best initial positioning model is the initial positioning model corresponding to the maximum value among multiple initial model metrics;

[0074] Step S48. Comparison operation: Compare each target parameter in the optimal parameter combination with the corresponding initial range. When there is a target parameter belonging to the boundary of the corresponding initial range, adjust the initial range to obtain the updated initial range;

[0075] Specifically, each target parameter has a preset initial range. The optimal parameter is found within the initial range as the model parameter. When the optimal parameter is at the boundary of the initial range, it is considered that the current initial range may limit the parameter value, so the initial range needs to be updated.

[0076] Specifically, step S48 specifically includes step S481, step S482, step S483, step S484, step S485, step S486, step S487, and step S488:

[0077] Step S481. Take the target parameter in the first parameter combination that belongs to the boundary of the corresponding initial range as the adjustment parameter;

[0078] Step S482. Determine the sub-optimal model metric from all the initial model metrics. The sub-optimal model metric is only less than the maximum model metric;

[0079] Step S483. Obtain the target parameter corresponding to the adjustment parameter in the target parameter combination corresponding to the sub-optimal model metric as the comparison parameter;

[0080] Step S484. Calculate the difference between the adjustment parameter and the comparison parameter to obtain the target difference;

[0081] Step S485. Compare the target difference with the second set threshold to obtain a parameter comparison result;

[0082] Step S486. Determine the magnification factor based on the parameter comparison result;

[0083] Step S487. Calculate the product of the target difference and the magnification factor to obtain the expansion range;

[0084] Step S488. Generate an updated initial range based on the initial range and the expansion range;

[0085] Specifically, determine the sub - optimal parameter combination that is second only to the optimal parameter combination from all parameter combinations. Take the parameter on the boundary of the initial range in the optimal parameter combination as the adjustment parameter, and determine the comparison parameter of the same category as the adjustment parameter from the sub - optimal parameter combination. Calculate the difference between the comparison parameter and the adjustment parameter. When this difference is large, a larger range of expansion of the initial range is required, while when this difference is small, only a smaller range of expansion of the initial range is needed. There are differences in the corresponding magnification factors in the two cases. Calculate the product of the magnification factor and the difference as the required expansion range. Finally, superimpose the required expansion range on the initial range to obtain the adjusted initial range.

[0086] Step S409. Repeat the generation operation, selection operation, adjustment operation, training operation, screening operation, and comparison operation until all target parameters in the optimal parameter combination belong to the corresponding initial range, and use the initial positioning model corresponding to the optimal parameter combination as the positioning model to be tested;

[0087] Specifically, re - determine the optimal parameter according to the updated range, and compare the relationship between this parameter and the updated range until all optimal parameters are within the initial range, then use this optimal parameter as the final parameter of the model.

[0088] Step S410. Bring the test samples and validation samples into the positioning model to be tested for prediction respectively, and calculate the model metrics based on the prediction results respectively to obtain the test model metrics and validation model metrics. The test model metrics are comprehensively generated by longitude metrics and latitude metrics;

[0089] Specifically, step S410 specifically includes step S4101, step S4102, step S4103, step S4104, step S4105, step S4106, step S4107, and step S4108:

[0090] Step S4101. Bring multiple test samples into the positioning model to be tested for prediction respectively to obtain multiple test prediction values. The test prediction pairs include predicted longitude differences and predicted latitude differences;

[0091] Step S4102. Calculate the difference between the epicenter longitude difference and the corresponding predicted longitude difference in multiple test samples, and calculate the square of each difference to obtain multiple first differences;

[0092] Step S4103. Calculate the sum of the epicenter longitude differences in multiple test samples to obtain a first result;

[0093] Step S4104. Calculate the difference between multiple predicted longitude differences and the first result, and calculate the square of each difference to obtain multiple second differences;

[0094] Step S4105. Calculate the sum of multiple first differences to obtain a second result;

[0095] Step S4106. Calculate the sum of multiple second differences to obtain a third result;

[0096] Step S4107. Calculate the ratio of the second result to the third result to obtain a fourth result;

[0097] Step S4108. Calculate the difference between the third set threshold and the fourth result as the longitude index;

[0098] Specifically, the model index calculation formula is:

[0099]

[0100] Among them, is the longitude and latitude prediction value of the prediction positioning model; y i is the actual longitude and latitude value in the test sample; y - is the average value of the actual longitude and latitude values in all test samples; n is the number of prediction samples;

[0101] Since the positioning model will output the epicenter longitude value and the epicenter latitude value, it is necessary to calculate the model index corresponding to the longitude value and the model index corresponding to the epicenter latitude value respectively, and comprehensively use the two model indexes to characterize the training situation of the current model, and determine whether the model needs to be trained again from the numerical value of the model index.

[0102] Step S411. Calculate the difference between the test model index and the validation model index to obtain the target difference;

[0103] Step S412. When the target difference is less than the first set threshold, use the positioning model to be tested as the target positioning model;

[0104] Specifically, for the model trained with training samples, validation samples and test samples are used to verify whether the model parameters are reasonable. The validation samples and test samples can test the performance of the model on unknown data. If the model metrics of the validation set are significantly smaller than those of the test set, it indicates that the model has the drawback of overfitting and certain adjustments need to be made to the model parameters.

[0105] Step S50. Input the predicted earthquake dataset into the target location model to obtain the target longitude and target latitude of the target epicenter.

[0106] Specifically, input the earthquake data to be predicted into the trained target location model for location, and relatively accurate epicenter longitude and latitude information can be obtained.

[0107] Embodiment 2:

[0108] As Figure 3 shown, this embodiment provides a positioning device for marine earthquakes. The device includes:

[0109] The first acquisition unit 10 is used to acquire multiple historical earthquake datasets. The historical earthquake datasets include information of multiple earthquake measurement stations that trigger alarms and epicenter information. The information of the earthquake measurement stations that trigger alarms includes the longitude of the measurement station, the latitude of the measurement station, and the triggering time of the measurement station. The epicenter information includes the longitude of the epicenter and the latitude of the epicenter.

[0110] The second acquisition unit 20 is used to acquire real-time earthquake datasets, and the real-time earthquake datasets include information of earthquake measurement stations that trigger alarms.

[0111] The preprocessing unit 30 is used to perform difference preprocessing on multiple historical earthquake datasets and real-time earthquake datasets to obtain multiple target earthquake datasets and predicted earthquake datasets. The difference preprocessing is used to represent the difference calculation of the information in the historical earthquake datasets.

[0112] The first training unit 40 is used to train a preset decision tree model based on the target earthquake datasets. When the preset model metrics meet the set conditions, stop training to obtain the target location model.

[0113] The input unit 50 is used to input the predicted earthquake dataset into the target location model to obtain the target longitude and target latitude of the target epicenter.

[0114] In a specific implementation manner disclosed in this application, the preprocessing unit 30 includes:

[0115] The first comparison unit is used to compare the triggering times of multiple measurement stations and use the measurement station corresponding to the minimum triggering time of the measurement station as the target measurement station.

[0116] A first calculation unit for calculating the trigger time differences between the remaining measurement stations and the target measurement station to obtain multiple measurement station trigger time differences;

[0117] A second calculation unit for calculating the longitude differences between the remaining measurement stations and the target measurement station to obtain multiple measurement station longitude differences;

[0118] A third calculation unit for calculating the latitude differences between the remaining measurement stations and the target measurement station to obtain multiple measurement station latitude differences;

[0119] A fourth calculation unit for calculating the difference between the epicenter longitude and the longitude of the target measurement station to obtain an epicenter longitude difference;

[0120] A fifth calculation unit for calculating the difference between the epicenter latitude and the latitude of the target measurement station to obtain an epicenter latitude difference;

[0121] A first acting unit for using the multiple measurement station trigger time differences, multiple measurement station longitude differences, multiple measurement station latitude differences, epicenter longitude difference, and epicenter latitude difference as a target earthquake data set.

[0122] In a specific implementation manner disclosed in the present application, the first training unit 40 includes:

[0123] A first partitioning unit for partitioning the multiple target earthquake data sets into training samples, test samples, and validation samples respectively;

[0124] A third obtaining unit for obtaining the initial range of each target parameter;

[0125] A generating unit for generating operations: randomly generating multiple initial parameters based on the initial range of each target parameter;

[0126] A selecting unit for selecting operations: randomly selecting one initial parameter from the multiple initial parameters of each target parameter for combination to obtain multiple parameter combinations, where at most two initial parameter values are the same in any two parameter combinations;

[0127] An adjusting unit for adjusting operations: adjusting the decision tree model based on each parameter combination to obtain multiple initial positioning models;

[0128] A second training unit for training operations: bringing the training samples into the multiple initial positioning models for training respectively and calculating model metrics to obtain multiple initial model metrics;

[0129] A screening unit for screening operations: screening the optimal parameter combination according to all the initial model metrics, where the optimal parameter combination is the parameter combination corresponding to the best initial positioning model, and the best initial positioning model is the initial positioning model corresponding to the maximum value among the multiple initial model metrics;

[0130] A second comparison unit for performing a comparison operation: comparing each target parameter in the optimal parameter combination with the corresponding initial range, and when there is a target parameter belonging to the boundary of the corresponding initial range, adjusting the initial range to obtain an updated initial range;

[0131] A first repetition unit for repeatedly performing the generation operation, selection operation, adjustment operation, training operation, screening operation, and comparison operation until all target parameters in the optimal parameter combination belong to the corresponding initial range, and using the initial positioning model corresponding to the optimal parameter combination as the positioning model to be tested;

[0132] A first input unit for respectively inputting the test sample and the verification sample into the positioning model to be tested for prediction, and respectively calculating model metrics based on the prediction results to obtain a test model metric and a verification model metric, where the test model metric is comprehensively generated from a longitude metric and a latitude metric;

[0133] A sixth calculation unit for calculating the difference between the test model metric and the verification model metric to obtain a target difference;

[0134] A second serving unit for, when the target difference is less than a first set threshold, using the positioning model to be tested as the target positioning model.

[0135] In a specific embodiment disclosed in the present application, the first input unit includes:

[0136] A second partitioning unit for respectively partitioning a plurality of training samples into a plurality of sets to obtain a plurality of training sets;

[0137] A first determination unit for performing a determination operation: randomly determining any one set from the plurality of training sets as the target set;

[0138] A second input unit for performing an input operation: inputting all training sets except the target set into a decision tree model for training to obtain a trained positioning model;

[0139] A seventh calculation unit for performing a first calculation operation: calculating, based on the trained positioning model, the model metric corresponding to each training set except the target set to obtain a plurality of first metrics, where the model metric is used to characterize the fitting degree of the trained positioning model to the training set;

[0140] A prediction unit for performing a prediction operation: inputting the target set into the trained positioning model for prediction, and calculating a model metric based on the prediction result to obtain a second metric;

[0141] An eighth calculation unit for performing a second calculation operation: performing a summation calculation on all the first metrics to obtain a first summation result;

[0142] The ninth calculation unit is used for the third calculation operation: summing the first summation result and the second index to obtain a second summation result;

[0143] The tenth calculation unit is used for the fourth calculation operation: calculating the ratio of the second summation result to the number of training sets to obtain a target ratio;

[0144] The second repetition unit is used to repeatedly perform the determination operation, substitution operation, first calculation operation, prediction operation, second calculation operation, third calculation operation, and fourth calculation operation until the target ratio of the number of training sets is obtained;

[0145] The summation calculation unit is used to sum multiple target ratios and divide the summation result by the number of training sets to obtain an initial model metric.

[0146] In a specific implementation manner disclosed in the present application, the second comparison unit includes:

[0147] The third assignment unit is used to use the target parameter belonging to the corresponding initial range boundary in the first parameter combination as an adjustment parameter;

[0148] The second determination unit is used to determine a sub-optimal model metric from all the initial model metrics, and the sub-optimal model metric is only smaller than the maximum model metric;

[0149] The fourth acquisition unit is used to acquire the target parameter corresponding to the adjustment parameter in the target parameter combination corresponding to the sub-optimal model metric as a comparison parameter;

[0150] The eleventh calculation unit is used to calculate the difference between the adjustment parameter and the comparison parameter to obtain a target difference;

[0151] The third comparison unit is used to compare the size of the target difference with a second set threshold to obtain a parameter comparison result;

[0152] The third determination unit is used to determine an expansion multiple based on the parameter comparison result;

[0153] The twelfth calculation unit is used to calculate the product of the target difference and the expansion multiple to obtain an expansion range;

[0154] The update unit is used to generate an updated initial range based on the initial range and the expansion range.

[0155] In a specific implementation manner disclosed in the present application, the first substitution unit includes:

[0156] The third substitution unit is used to substitute multiple test samples into the positioning model to be tested for prediction respectively to obtain multiple test prediction values, and the test prediction pair includes a predicted longitude difference and a predicted latitude difference;

[0157] A thirteenth calculation unit is configured to calculate the difference between the epicenter longitude differences and the corresponding predicted longitude differences in multiple test samples, and calculate the square of each difference to obtain multiple first differences.

[0158] A fourteenth calculation unit is configured to calculate the sum of the epicenter longitude differences in multiple test samples to obtain a first result.

[0159] A fifteenth calculation unit is configured to calculate the difference between multiple predicted longitude differences and the first result, and calculate the square of each difference to obtain multiple second differences.

[0160] A sixteenth calculation unit is configured to calculate the sum of multiple first differences to obtain a second result.

[0161] A seventeenth calculation unit is configured to calculate the sum of multiple second differences to obtain a third result.

[0162] An eighteenth calculation unit is configured to calculate the ratio of the second result to the third result to obtain a fourth result.

[0163] A nineteenth calculation unit is configured to calculate the difference between a third set threshold and the fourth result as a longitude index.

[0164] It should be noted that regarding the device in the above embodiments, the specific manners in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated herein.

[0165] Embodiment 3:

[0166] Corresponding to the above method embodiment, a positioning device for marine earthquakes is further provided in this embodiment. A positioning device for marine earthquakes described below can be correspondingly referred to the positioning method for marine earthquakes described above.

[0167] Figure 4 is a block diagram of a positioning device 800 for marine earthquakes shown according to an exemplary embodiment. As Figure 4 shown, the positioning device 800 for marine earthquakes may include: a processor 801, a memory 802. The positioning device 800 for marine earthquakes may further include one or more of a multimedia component 803, an I / O interface 804, and a communication component 805.

[0168] Among them, the processor 801 is used to control the overall operation of the positioning device 800 for marine seismic exploration to complete all or part of the steps in the above-mentioned positioning method for marine seismic exploration. The memory 802 is used to store various types of data to support the operation of the positioning device 800 for marine seismic exploration. These data may include, for example, instructions for any application or method operating on the positioning device 800 for marine seismic exploration, as well as application-related data, such as contact data, sent and received messages, pictures, audio, video, and so on. The memory 802 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disc. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signals may be further stored in the memory 802 or sent through the communication component 805. The audio component also includes at least one speaker for outputting audio signals. The I / O interface 804 provides an interface between the processor 801 and other interface modules, and the above-mentioned other interface modules may be a keyboard, a mouse, buttons, etc. These buttons may be virtual buttons or physical buttons. The communication component 805 is used for wired or wireless communication between the positioning device 800 for marine seismic exploration and other devices. Wireless communication, such as Wi-Fi, Bluetooth, near field communication (NFC), 2G, 3G or 4G, or a combination of one or more of them. Accordingly, the communication component 805 may include: a Wi-Fi module, a Bluetooth module, an NFC module.

[0169] In an exemplary embodiment, the positioning device 800 for marine seismic exploration can be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components, and is used to execute the above-mentioned method for positioning marine seismic exploration.

[0170] In another exemplary embodiment, a computer-readable storage medium including program instructions is further provided. When the program instructions are executed by a processor, the steps of the above-mentioned method for positioning marine seismic exploration are implemented. For example, the computer-readable storage medium can be the above-mentioned memory 802 including program instructions, and the above-mentioned program instructions can be executed by the processor 801 of the positioning device 800 for marine seismic exploration to complete the above-mentioned method for positioning marine seismic exploration.

[0171] Embodiment 4:

[0172] Corresponding to the above method embodiment, a readable storage medium is further provided in this embodiment. A readable storage medium described below can be correspondingly referred to with a method for positioning marine seismic exploration described above.

[0173] A readable storage medium stores a computer program. When the computer program is executed by a processor, the steps of the method for positioning marine seismic exploration in the above method embodiment are implemented.

[0174] Specifically, the readable storage medium can be various readable storage media such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc that can store program codes.

[0175] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

[0176] As described above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A method for locating marine earthquakes, characterized in that, Including: Obtain multiple historical earthquake datasets, where the historical earthquake datasets include information of multiple seismic measurement stations that trigger alarms and epicenter information. The information of the seismic measurement stations that trigger alarms includes the longitude of the measurement station, the latitude of the measurement station, and the trigger time of the measurement station. The epicenter information includes the longitude of the epicenter and the latitude of the epicenter; Obtain a real-time earthquake dataset, where the real-time earthquake dataset includes information of the seismic measurement station that triggers an alarm; Perform difference preprocessing on the multiple historical earthquake datasets and the real-time earthquake dataset to obtain multiple target earthquake datasets and a predicted earthquake dataset. The difference preprocessing is used to represent the calculation of differences in the information in the historical earthquake datasets; Train a preset decision tree model based on the target earthquake datasets. When the preset model metrics meet the set conditions, stop training to obtain a target positioning model; Input the predicted earthquake dataset into the target positioning model to obtain the target longitude and target latitude of the target epicenter.

2. The positioning method of marine seismic according to claim 1, characterized in that , Perform difference preprocessing on the multiple historical earthquake datasets to obtain multiple target earthquake datasets, including: Compare the trigger times of multiple measurement stations, and use the measurement station corresponding to the minimum trigger time of the measurement stations as the target measurement station; Calculate the trigger time differences between the remaining measurement stations and the target measurement station to obtain multiple trigger time differences of the measurement stations; Calculate the longitude differences between the remaining measurement stations and the target measurement station to obtain multiple longitude differences of the measurement stations; Calculate the latitude differences between the remaining measurement stations and the target measurement station to obtain multiple latitude differences of the measurement stations; Calculate the difference between the epicenter longitude and the longitude of the target measurement station to obtain an epicenter longitude difference; Calculate the difference between the epicenter latitude and the latitude of the target measurement station to obtain an epicenter latitude difference; Use the multiple trigger time differences of the measurement stations, the multiple longitude differences of the measurement stations, the multiple latitude differences of the measurement stations, the epicenter longitude difference, and the epicenter latitude difference as the target earthquake datasets.

3. The positioning method of marine seismic according to claim 1, characterized in that , Train a preset decision tree model based on the target earthquake datasets. When the preset model metrics meet the set conditions, stop training to obtain a target positioning model. The decision tree model includes three target parameters, including: Divide the multiple target earthquake datasets into training samples, test samples, and validation samples respectively; Obtain the initial range of each target parameter; Generate operation: Randomly generate multiple initial parameters based on the initial range of each target parameter; Selection operation: Randomly select one initial parameter from the multiple initial parameters of each target parameter for combination to obtain multiple parameter combinations. Among them, at most two initial parameter values are the same in any two parameter combinations; Adjustment operation: Adjust the decision tree model based on each parameter combination to obtain multiple initial positioning models; Training operation: Input the training samples into the multiple initial positioning models for training respectively, and calculate the model metrics to obtain multiple initial model metrics; Screening operation: Screen out the optimal parameter combination according to all the initial model metrics. The optimal parameter combination is the parameter combination corresponding to the best initial positioning model, and the best initial positioning model is the initial positioning model corresponding to the maximum value among the multiple initial model metrics; Comparison operation: Compare each target parameter in the optimal parameter combination with the corresponding initial range. When there is a target parameter belonging to the boundary of the corresponding initial range, adjust the initial range to obtain an updated initial range; Repeat the generation operation, the selection operation, the adjustment operation, the training operation, the screening operation, and the comparison operation until the target parameters in the optimal parameter combination all belong to the corresponding initial range, and use the initial positioning model corresponding to the optimal parameter combination as the positioning model to be tested; Bring the test sample and the verification sample into the positioning model to be tested for prediction respectively, and calculate the model metrics based on the prediction results respectively to obtain a test model metric and a verification model metric. The test model metric is comprehensively generated by a longitude metric and a latitude metric; Calculate the difference between the test model metric and the verification model metric to obtain a target difference; When the target difference is less than a first set threshold, use the positioning model to be tested as the target positioning model.

4. The positioning method of marine seismic according to claim 3, characterized in that, Training operation: Bring the training samples into multiple initial positioning models for training respectively, and calculate model metrics to obtain multiple initial model metrics, including: Divide multiple training samples into multiple sets respectively to obtain multiple training sets; Determination operation: Randomly determine any one set from multiple training sets as a target set; Bringing operation: Bring all training sets except the target set into the decision tree model for training to obtain a trained positioning model; First calculation operation: Calculate the model metrics corresponding to each training set except the target set based on the trained positioning model to obtain multiple first metrics. The model metric is used to characterize the fitting degree of the trained positioning model to the training set; Prediction operation: Bring the target set into the trained positioning model for prediction, and calculate the model metric based on the prediction result to obtain a second metric; Second calculation operation: Sum all the first metrics to obtain a first summation result; Third calculation operation: Sum the first summation result and the second metric to obtain a second summation result; Fourth calculation operation: Calculate the ratio of the second summation result to the number of training sets to obtain a target ratio; Repeat the determination operation, the bringing operation, the first calculation operation, the prediction operation, the second calculation operation, the third calculation operation, and the fourth calculation operation until the target ratio of the number of training sets is obtained; Sum multiple target ratios and divide the summation result by the number of training sets to obtain the initial model metric.

5. An ocean earthquake positioning device, characterized in that, Including: The first acquisition unit is used to acquire multiple historical earthquake data sets. The historical earthquake data set includes information of multiple earthquake measurement stations that trigger alarms and epicenter information. The information of the earthquake measurement stations that trigger alarms includes the longitude of the measurement station, the latitude of the measurement station, and the trigger time of the measurement station. The epicenter information includes the longitude of the epicenter and the latitude of the epicenter; The second acquisition unit is used to acquire real-time earthquake data sets. The real-time earthquake data set includes information of earthquake measurement stations that trigger alarms; A preprocessing unit for performing difference preprocessing on multiple historical earthquake datasets and the real-time earthquake dataset to obtain multiple target earthquake datasets and a predicted earthquake dataset, where the difference preprocessing is used to represent the calculation of differences in the information in the historical earthquake datasets; A first training unit for training a preset decision tree model based on the target earthquake datasets, and stopping the training when the preset model metrics meet the set conditions to obtain a target positioning model; An input unit for inputting the predicted earthquake dataset into the target positioning model to obtain the target longitude and target latitude of the target epicenter.

6. The positioning device for marine seismic according to claim 5, wherein The preprocessing unit includes: A first comparison unit for comparing the triggering times of multiple measuring stations and taking the measuring station corresponding to the minimum triggering time of the measuring stations as the target measuring station; A first calculation unit for calculating the triggering time differences between the remaining measuring stations and the target measuring station to obtain multiple measuring station triggering time differences; A second calculation unit for calculating the longitude differences between the remaining measuring stations and the target measuring station to obtain multiple measuring station longitude differences; A third calculation unit for calculating the latitude differences between the remaining measuring stations and the target measuring station to obtain multiple measuring station latitude differences; A fourth calculation unit for calculating the difference between the epicenter longitude and the longitude of the target measuring station to obtain an epicenter longitude difference; A fifth calculation unit for calculating the difference between the epicenter latitude and the latitude of the target measuring station to obtain an epicenter latitude difference; A first acting unit for taking the multiple measuring station triggering time differences, multiple measuring station longitude differences, multiple measuring station latitude differences, epicenter longitude difference, and epicenter latitude difference as the target earthquake datasets.

7. The positioning device for marine seismic according to claim 5, characterized in that, The first training unit includes: A first partitioning unit for partitioning multiple target earthquake datasets into training samples, test samples, and validation samples respectively; A third obtaining unit for obtaining the initial range of each target parameter; A generating unit for generating operations: randomly generating multiple initial parameters based on the initial range of each target parameter; A selecting unit for selecting operations: randomly selecting one initial parameter from the multiple initial parameters of each target parameter for combination to obtain multiple parameter combinations, where at most two initial parameter values are the same in any two parameter combinations; An adjusting unit for adjusting operations: adjusting the decision tree model based on each parameter combination to obtain multiple initial positioning models; A second training unit for training operations: bringing the training samples into the multiple initial positioning models for training respectively and calculating the model metrics to obtain multiple initial model metrics; A screening unit for screening operations: screening to obtain the optimal parameter combination according to all the initial model metrics, where the optimal parameter combination is the parameter combination corresponding to the best initial positioning model, and the best initial positioning model is the initial positioning model corresponding to the maximum value among the multiple initial model metrics; A second comparison unit for comparing operations: comparing each target parameter in the optimal parameter combination with the corresponding initial range, and when there is a target parameter belonging to the boundary of the corresponding initial range, adjusting the initial range to obtain the updated initial range; A first repeating unit for repeating the generating operation, the selecting operation, the adjusting operation, the training operation, the screening operation, and the comparing operation until all target parameters in the optimal parameter combination belong to the corresponding initial ranges, and taking the initial positioning model corresponding to the optimal parameter combination as the positioning model to be tested; A first inputting unit for respectively inputting the test samples and the validation samples into the positioning model to be tested for prediction, and respectively calculating the model metrics based on the prediction results to obtain a test model metric and a validation model metric, where the test model metric is comprehensively generated from a longitude metric and a latitude metric; A sixth calculating unit for calculating the difference between the test model metric and the validation model metric to obtain a target difference; A second serving as unit for taking the positioning model to be tested as the target positioning model when the target difference is less than a first set threshold; 8. The positioning device for marine seismic according to claim 7, wherein, The first inputting unit includes: A second dividing unit for respectively dividing a plurality of training samples into a plurality of sets to obtain a plurality of training sets; A first determining unit for a determining operation: randomly determining any one set from the plurality of training sets as a target set; A second inputting unit for an inputting operation: inputting all training sets except the target set into the decision tree model for training to obtain a trained positioning model; A seventh calculating unit for a first calculating operation: calculating, based on the trained positioning model, the model metric corresponding to each training set except the target set to obtain a plurality of first metrics, where the model metric is used to characterize the fitting degree of the trained positioning model to the training set; A predicting unit for a predicting operation: inputting the target set into the trained positioning model for prediction, and calculating the model metric based on the prediction result to obtain a second metric; An eighth calculating unit for a second calculating operation: performing a summation calculation on all the first metrics to obtain a first summation result; A ninth calculating unit for a third calculating operation: performing a summation calculation on the first summation result and the second metric to obtain a second summation result; A tenth calculating unit for a fourth calculating operation: calculating the ratio of the second summation result to the number of training sets to obtain a target ratio; A second repeating unit for repeating the determining operation, the inputting operation, the first calculating operation, the predicting operation, the second calculating operation, the third calculating operation, and the fourth calculating operation until the target ratio of the number of training sets is obtained; A summation calculating unit for performing a summation calculation on a plurality of target ratios, and dividing the summation result by the number of training sets to obtain the initial model metric; 9. An ocean earthquake positioning device, characterized in that, including: A memory for storing a computer program; A processor for implementing the steps of the positioning method for marine seismic waves according to any one of claims 1 to 4 when executing the computer program; 10. A readable storage medium, characterized in that, A computer program is stored on the readable storage medium, and when the computer program is executed by the processor, the steps of the positioning method for marine seismic waves according to any one of claims 1 to 4 are implemented.

Images