Seismic signal classification research method, device and equipment

Abstract

The embodiment of the invention provides a seismic signal classification research method, device and equipment. The method comprises the following steps: acquiring original seismic signal sample data, and preprocessing the original seismic signal sample data to obtain seismic signal data; inputting the seismic signal data into a self-adaptive local iterative filtering model, outputting residual IMF components, and calculating the variance contribution rate of the residual IMF components; constructing a bidirectional long-short-term memory neural network model, inputting the variance contribution rate of each IMF component into the bidirectional long-short-term memory neural network model, and extracting seismic signal difference features; and classifying the seismic signal data according to the seismic signal difference characteristics. In this way, the generalization ability of the model can be enhanced, the requirements for high efficiency, accuracy, real-time performance and the like of earthquake time classification are improved, and meanwhile the labor cost is saved.

Description

Technical Field

[0001] The present invention generally relates to the field of seismic signal analysis and classification, and more specifically, to a seismic signal classification research method, device and apparatus. Background Art

[0002] With the expansion of the global earthquake monitoring network and the increase in engineering activities, seismic stations continue to collect various vibration signals, which may come from natural earthquakes or human activities, such as engineering blasting. Inaccurate identification of earthquake events leads to problems such as false positives and false negatives in earthquake catalogs, resulting in deviations in seismic activity analysis and errors in earthquake mechanism analysis. Manual methods have problems such as low identification efficiency, strong subjectivity and poor real-time performance.

[0003] In recent years, with the development of artificial intelligence, neural networks have been widely used in the research of earthquake event recognition. Neural network technology has received extensive attention in the field of earthquake signal classification due to its excellent feature extraction ability.

[0004] Existing deep learning methods using neural networks do not take into account the differences in frequency component characteristics in different seismic event signals and the noise and trend components in the seismic signals, resulting in the model's insufficient expression of signal characteristics and the inability to accurately and efficiently classify seismic signals. Summary of the invention

[0005] According to an embodiment of the present invention, a research scheme for earthquake signal classification is provided, which can enhance the generalization ability of the model, improve the requirements for high efficiency, accuracy and real-time earthquake time classification, and save labor costs.

[0006] In a first aspect of the present invention, a seismic signal classification research method is provided. The method comprises:

[0007] Acquire original seismic signal sample data, and preprocess the original seismic signal sample data to obtain seismic signal data;

[0008] Inputting the seismic signal data into an adaptive local iterative filtering model, outputting the residual IMF component, and then calculating the variance contribution rate of the residual IMF component;

[0009] A bidirectional long short-term memory neural network model is constructed, and the variance contribution rate of each IMF component is input into the bidirectional long short-term memory neural network model to extract the difference characteristics of seismic signals;

[0010] The seismic signal data is classified according to the difference characteristics of the seismic signals.

[0011] Furthermore, the adaptive local iterative filtering model includes an inner loop module and an outer loop module;

[0012] The inner loop module is used to traverse the seismic signal data and extract the effective IMF components;

[0013] The outer loop module is used to extract the remaining IMF components by cyclically calling the valid IMF components obtained by the inner loop module.

[0014] Furthermore, the traversing of the seismic signal data to extract effective IMF components includes:

[0015] Seismic signal data sequences are extracted from seismic signal data in turn, and the seismic signal data sequences are convolved with an adaptive filter function to obtain a sliding operator; a wave operator is then obtained based on the seismic signal data sequence and the sliding operator; finally, the correlation coefficient between the sliding operator and the wave operator is calculated, and when the correlation coefficient is greater than or equal to a threshold, the seismic signal data sequence is extracted as a valid IMF component and the next iteration is performed, otherwise the next iteration is performed directly until all seismic signal data are traversed.

[0016] Furthermore, the calculation formula of the sliding operator is as follows:

[0017]

[0018] The calculation formula of the wave operator is as follows:

[0019] v(t)=g(t)-μ(g(t))

[0020] The correlation coefficient calculation formula of the sliding operator and the fluctuation operator is as follows:

[0021]

[0022] Wherein, μ(g(t)) is the sliding operator, where g(t) is the seismic signal data sequence; i is the filtering interval; t is the time; τ is the delay time; ω(t) is the adaptive filtering function in the filtering interval i, which is composed of the solutions of all Fokker-Planck differential equations in the filtering interval i; d is the point integral symbol; v(t) is the wave operator; ρ μv is the correlation coefficient between the sliding operator and the fluctuation operator, that is, the basic statistic of regression analysis; T is the time period.

[0023] Furthermore, extracting the remaining IMF components from the valid IMF components obtained by cyclically calling the inner loop module comprises:

[0024] After each call to the inner loop module, the remaining IMF components are extracted according to the valid IMF components output by the inner loop module. When the outer loop module call time reaches the time period T, the outer loop is terminated and the remaining IMF components are output. Otherwise, the remaining IMF components are used as input data of the inner loop module to continue iterative extraction.

[0025] Furthermore, the extraction formula of the residual IMF component is as follows:

[0026] e(t)=g(t)-r(t)

[0027] Among them, e(t) is the residual IMF component; r(t) is the effective IMF component extracted by the inner loop module.

[0028] Furthermore, the variance contribution rate of the remaining IMF components is calculated, including:

[0029]

[0030] Among them, ψ i is the variance contribution rate; a is the residual IMF component index; S is the total number of residual IMF components; e c (t) is the cth residual IMF component; c is the seismic signal data index; N is the total number of seismic signal data.

[0031] Furthermore, the bidirectional long short-term memory neural network model is constructed, and the variance contribution rate of each IMF component is used as the input data of the bidirectional long short-term memory neural network model to extract the difference characteristics of the seismic signal, including:

[0032] The bidirectional long short-term memory neural network model includes: a forward LSTM module, a reverse LSTM module, and a difference feature extraction module;

[0033] The forward LSTM module comprises:

[0034] y f,t =g(W y'f h t-1 +ψ i e(t)+b y' )

[0035] The reverse LSTM module comprises:

[0036] y b,t =g(W y'b h t-1 +ψ i e(t)+b y ')

[0037] The difference feature extraction module comprises:

[0038] y′ t =g(W y'f y f,t +W y'b y b,t +b y ')

[0039] Among them, h t-1 is the result of the IMF at the previous moment; y' t is the difference feature of seismic signal; g is the activation function of output layer; W y'f is the forward weight matrix; y f,t is the output of the forward LSTM unit; W y'b is the inverse weight matrix; y b,t is the output of the reverse LSTM unit; b y' is the bias term.

[0040] In a second aspect of the present invention, a seismic signal classification research device is provided. The device comprises:

[0041] An acquisition module is used to acquire original seismic signal sample data, and pre-process the original seismic signal sample data to obtain seismic signal data;

[0042] A calculation module, used for inputting the seismic signal data into an adaptive local iterative filtering model, outputting a residual IMF component, and then calculating a variance contribution rate of the residual IMF component;

[0043] The feature extraction module is used to construct a bidirectional long short-term memory neural network model, input the variance contribution rate of each IMF component into the bidirectional long short-term memory neural network model, and extract the difference characteristics of the seismic signal;

[0044] The classification module is used to classify seismic signal data according to the difference characteristics of seismic signals.

[0045] In a third aspect of the present invention, an electronic device is provided. The electronic device has at least one processor; and a memory connected to the at least one processor in communication; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the method of the first aspect of the present invention.

[0046] It should be understood that the contents described in the summary of the invention are not intended to limit the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The above and other features, advantages and aspects of the embodiments of the present invention will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. In the accompanying drawings, the same or similar reference numerals represent the same or similar elements, wherein:

[0048] Figure 1 A flow chart showing a seismic signal classification research method according to an embodiment of the present invention;

[0049] Figure 2 shows a blasting signal waveform diagram according to an embodiment of the present invention;

[0050] Figure 3 A seismic signal waveform diagram according to an embodiment of the present invention is shown;

[0051] Figure 4 A flow chart showing a seismic signal classification research device according to an embodiment of the present invention is shown;

[0052] Figure 5 A block diagram of an exemplary electronic device capable of implementing embodiments of the present invention is shown;

[0053] Among them, 500 is an electronic device, 501 is a CPU, 502 is a ROM, 503 is a RAM, 504 is a bus, 505 is an I / O interface, 506 is an input unit, 507 is an output unit, 508 is a storage unit, and 509 is a communication unit. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] In addition, the term "and / or" in this article is only a description of the association relationship between the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0056] In the present invention, waveform data is collected from the seismic station, the raw data is preprocessed, and the variance contribution rate of each IMF (Intrinsic Mode Function) component is calculated by using an adaptive local iterative filtering model to decompose different seismic signals, and the difference characteristics of seismic events at different frequencies are extracted by a bidirectional long short-term memory neural network (BiLSTM, Bidirectional Long Short-Term Memory), and different seismic events are classified. The present invention classifies different seismic signals, considers incorporating external environmental factors and geological parameters into the model for training, identifies seismic signal types according to the waveform characteristic information of the signal, enhances the generalization ability of the model, and improves the requirements of high efficiency, accuracy and real-time performance of seismic time classification, while saving labor costs.

[0057] Figure 1 A flow chart of a seismic signal classification research method according to an embodiment of the present invention is shown.

[0058] The method includes:

[0059] S101, obtaining original seismic signal sample data, and preprocessing the original seismic signal sample data to obtain seismic signal data.

[0060] Specifically, the seismic signal sample data are high-quality natural earthquake waveform data and blasting waveform data containing noise and trend components collected and selected by the seismic station. At the same time, the waveform data contains waveforms of three components: vertical, east-west, and north-south.

[0061] In this embodiment, in order to avoid weight deviation caused by too large or too small values ​​and enhance the stability and convergence speed of the model, the original seismic signal sample data is preprocessed:

[0062] (1) The original seismic signal sample data from 1 second before the seismic phase to 50 seconds after the seismic phase are intercepted as preprocessed data, and the data sampling rate is 100 Hz.

[0063] (2) Perform normalization calculation on each preprocessed data:

[0064]

[0065] Among them, v i is the amplitude of the original seismic signal sample data; v' i is the amplitude of the seismic signal sample data after normalization.

[0066] S102, inputting the seismic signal data into an adaptive local iterative filtering model, outputting the residual IMF component, and then calculating the variance contribution rate of the residual IMF component.

[0067] In this embodiment, the adaptive local iterative filtering model includes an inner loop module and an outer loop module.

[0068] Specifically, the inner loop module is used to traverse the seismic signal data and extract the effective IMF components: extract the seismic signal data sequence from the seismic signal data in turn, convolve the seismic signal data sequence with the adaptive filter function to obtain the sliding operator; then obtain the wave operator based on the seismic signal data sequence and the sliding operator; finally, calculate the correlation coefficient between the sliding operator and the wave operator, and when the correlation coefficient is greater than or equal to a threshold, extract the seismic signal data sequence as the effective IMF component and perform the next iteration, otherwise directly perform the next iteration until all the seismic signal data are traversed.

[0069] Among them, the correlation coefficient mainly represents the correlation of the information amount of the sliding operator and the wave operator. The larger the value, the greater the correlation, and vice versa. The threshold is 1 / 10 of the maximum value in the correlation coefficient sequence. When the correlation coefficient is less than the threshold, the seismic signal data sequence extracted by the inner loop module is an invalid component; when the correlation coefficient is greater than or equal to the threshold, the seismic signal data sequence extracted by the inner loop module is a valid component.

[0070] The inner loop module extracts the IMF components in the signal more accurately through iteration, ensuring that the signal is decomposed into a series of IMF components with a single frequency, thereby more accurately capturing the multi-scale features in the seismic signal. The IMF components extracted by the inner loop module can better express the time-frequency characteristics of the seismic signal. The orderly arrangement of these components from high frequency to low frequency can well represent the low-frequency trend terms, thereby enhancing the model's ability to parse complex signals. At the same time, because the inner loop module can more effectively extract the time series dependency and multi-scale features in the seismic signal, it enhances the sensitivity of the fusion model to information in different frequency bands and improves the classification ability of complex seismic signals.

[0071] The calculation formula of the sliding operator is as follows:

[0072]

[0073] The calculation formula of the wave operator is as follows:

[0074] v(t)=g(t)-μ(g(t))

[0075] The correlation coefficient calculation formula of the sliding operator and the fluctuation operator is as follows:

[0076]

[0077] Wherein, μ(g(t)) is the sliding operator, where g(t) is the seismic signal data sequence; i is the filtering interval; t is the time; τ is the delay time; ω(t) is the adaptive filtering function in the filtering interval i, which is composed of the solutions of all Fokker-Planck differential equations (partial differential equations describing the evolution of probability density) in the filtering interval i; d is the point integral symbol; v(t) is the wave operator; ρ μv is the correlation coefficient between the sliding operator and the fluctuation operator, that is, the basic statistic of regression analysis; T is the time period.

[0078] Sliding operators can be used for time-frequency analysis of signals in seismic signal processing. By analyzing the sliding scanning signal, a sliding operator for extracting harmonics can be designed. The sliding operator can extract high-order harmonics from the original seismic record, thereby reducing the impact of harmonic distortion on the fundamental wave and improving the quality of seismic data. The wave operator can describe the propagation mechanism of effective waves. The wave operator implicitly performs summation operations during the imaging process, so that the diffraction energy of the effective diffraction point is enhanced, while the random noise energy cancels each other out, thereby improving the signal-to-noise ratio of the data. The correlation coefficient can be used to measure the similarity between different signals or different parts of the signal in seismic signal processing. By calculating the correlation characteristics of the fundamental wave and the harmonics, an operator for extracting harmonics can be designed to effectively separate and eliminate harmonic interference.

[0079] In this embodiment, the outer loop module is used to extract the remaining IMF components by cyclically calling the valid IMF components obtained by the inner loop module: after each call to the inner loop module, the remaining IMF components are extracted according to the valid IMF components output by the inner loop module. When the outer loop module call time reaches the time period T, the outer loop is terminated and the remaining IMF components are output. Otherwise, the remaining IMF components are used as input data of the inner loop module to continue iterative extraction.

[0080] Specifically, when the outer loop module calling time reaches the time period T, the trend of the remaining IMF components changing with time is formed (the signal becomes stable), that is, the remaining IMF components at this time are all trend components.

[0081] The outer loop module can further optimize and filter the IMF components extracted by the inner loop module, thereby improving the efficiency of the entire signal decomposition process. At the same time, the outer loop module also helps to identify and extract the main trends and periodic components in the signal, which is of great significance for fields such as time series analysis and prediction.

[0082] In this embodiment, the extraction formula of the residual IMF component is as follows:

[0083] e(t)=g(t)-r(t)

[0084] Among them, e(t) is the residual IMF component; r(t) is the effective IMF component extracted by the inner loop module.

[0085] The inner loop module is responsible for the detailed decomposition of each signal to ensure that each IMF component meets the definition conditions of the IMF, that is, in the entire time range, the number of extreme points is equal to or differs by 1 from the number of zero crossings. The outer loop module is responsible for the overall control and optimization of the results of the inner loop module to ensure the efficiency and accuracy of the entire decomposition process. At the same time, through the detailed decomposition of the inner loop module, the high-frequency to low-frequency components in the signal can be captured more accurately, and the outer loop module can further optimize and screen these components, thereby enhancing the model's ability to extract complex signal features.

[0086] In this embodiment, the adaptive filter function ω(t) is constructed using the Fokker-Planck differential equation, and can be adjusted according to the local characteristics of the input signal, effectively removing noise from the signal and avoiding over-smoothing and loss of details.

[0087] Specifically, the construction process of the adaptive filter function ω(t) is as follows:

[0088] (1) Determine whether the construction conditions are met. If so, jump to step (2).

[0089] Specifically, when constructing the filter function, the construction condition is:

[0090] Let f 1 (x) and f 2 (x) is a differentiable function, when x 1 <0 <x 2 When the following conditions are met:

[0091] 1)f 1 (x 1 ) = f 2 (x 2 )=0, and There is 1 (x)>0.

[0092] 2)f 2 (x 1 )<0 <f 2 (x 2 ).

[0093] Among them, x 1 、x 2 is the boundary value of the wave operator in a specific interval; x is the independent variable of the wave operator.

[0094] (2) Construct the adaptive filter function ω(t) through the Fokker-Planck differential equation.

[0095] In this embodiment, the Fokker-Planck differential equation is as follows:

[0096]

[0097] in, To solve the partial derivative sign; P(x, t) is the probability density function at position x and time t; t is time; x is position; A(x, t) is the drift term, which represents the deterministic trend of the system, and the gradient function is used to enhance the main characteristics of the signal; B(x, t) is the diffusion term, which represents the randomness or noise intensity of the system, and the local gradient function is used to adaptively adjust the smoothness; g(t) is the seismic signal data sequence; λ is the control coefficient; η is the diffusion coefficient.

[0098] Specifically, the drift term A(x, t) causes the probability density function P(x, t) to move in the positive and negative directions within the filter interval as the drift term changes. The diffusion term B(x, t) causes the probability density function P(x, t) to deviate from the center and diffuse to both ends within the filter interval.

[0099] when When , the Fokker-Planck differential equation is solved. If the Fokker-Planck differential equation has a non-zero solution P(x) and satisfies the following conditions, it is considered that the solutions P(x) of all differential equations exist in the filtering interval, and the solutions P(x) of all equations constitute the adaptive filtering function ω(t). The specific conditions that P(x) should satisfy are:

[0100] 1) P(x)≥0.

[0101] 2) P(x)=0.

[0102] In this embodiment, the variance contribution rate of the remaining IMF components is calculated, including:

[0103]

[0104] Among them, ψ i is the variance contribution rate; a is the residual IMF component index; S is the total number of residual IMF components; e c (t) is the cth residual IMF component; c is the seismic signal data index; N is the total number of seismic signal data.

[0105] Since the short-period surface waves generated by explosions tend to show more significant energy in higher-frequency IMFs, while the short-period surface waves of natural earthquakes may dominate in relatively low-frequency IMFs, by calculating the variance contribution rate of each IMF (i.e., the energy proportion of each IMF component in the entire signal), the contribution of each IMF component to the original signal fluctuation can be quantified, thereby identifying the components that have a greater impact on signal changes. Specifically, the variance contribution rate can guide the signal processing process. For example, when denoising or extracting features, those IMF components with higher variance contribution rates can be given priority because they have a greater impact on the overall structure and characteristics of the signal. In addition, in prediction tasks, the variance contribution rate can also help identify which IMF components contain more prediction information, thereby improving the accuracy of the prediction model.

[0106] S103, construct a bidirectional long short-term memory neural network model (BiLSTM), input the variance contribution rate of each IMF component into the bidirectional long short-term memory neural network model, and extract the difference characteristics of the seismic signal, including:

[0107] In this embodiment, the bidirectional long short-term memory neural network model includes: a forward LSTM module, a reverse LSTM module, and a difference feature extraction module;

[0108] The forward LSTM module comprises:

[0109] y f,t =g(W y'f h t-1 +ψ i e(t)+b y ')

[0110] The reverse LSTM module comprises:

[0111] y b,t =g(W y'b h t-1 +ψ i e(t)+b y' )

[0112] The difference feature extraction module comprises:

[0113] y′ t =g(W y'f y f,t +W y'b y b,t +b y' )

[0114] Among them, h t-1 is the result of the IMF at the previous moment; y' t is the difference feature of seismic signal; g is the activation function of output layer; W y'fis the forward weight matrix; y f,t is the output of the forward LSTM unit; W y'b is the inverse weight matrix; y b,t is the output of the reverse LSTM unit; b y' is the bias term.

[0115] Specifically, the BiLSTM neural network is a sequence processing model based on a bidirectional recurrent neural network (RNN) and an LSTM neural network (Long Short-Term Memor). By combining the advantages of bidirectional RNN and LSTM, it processes data simultaneously in the forward and backward directions of the seismic signal data sequence, thereby better understanding and processing complex sequence data.

[0116] Among them, the calculation formula of LSTM neural network is as follows:

[0117] f t =sig(W fy y t-1 +Wf x x t +b f )

[0118] i t =sig(W iy y t-1 +W ix x t +b i )

[0119] u t =tanh(W uy y t-1 +W ux x t +b u )

[0120] c t =i t ×u t +f t ×c t-1

[0121] o t =sig(W oy y t-1 +W ox x t +b o )

[0122] y t =o t ×tanh(c t )

[0123] Among them, f t is the output of the forget gate at time t; W fy is the weight matrix of the forget gate on the y-axis; t-1 is the output of the forget gate at time t-1; W fx is the weight matrix of the x-axis forget gate; t is the input data at time t; b f is the bias term of the forget gate; i t The output is the input; W iy is the weight matrix of the y-axis input gate; W ix is the weight matrix of the x-axis input gate; b i is the bias term of the input gate; u t is the output of the candidate memory; W uy is the weight matrix of the candidate state on the y-axis; W ux is the weight matrix of the candidate state on the x-axis; b u is the bias term of the output gate; c t is the current state of the memory unit, which is determined by the new information controlled by the input gate and the old information controlled by the forget gate; c t-1 is the state of the memory unit at the last moment, which is determined by the new information controlled by the input gate and the old information controlled by the forget gate; t is the output of the output gate; W oy is the weight matrix of the y-axis output gate; W ox is the weight matrix of the x-axis output gate; b o is the bias term of the candidate state; y t is the memory unit state at the current moment.

[0124] The BiLSTM neural network combines the outputs of the forward LSTM module and the reverse LSTM module to form a vector containing bidirectional information features, which fully captures the time series information features. Specifically, the BiLSTM neural network processes data from the forward (forward LSTM module) and backward (reverse LSTM module) of the residual IMF component at the same time, and improves the neural network's ability to capture asymmetric structures by processing information in both directions, thereby improving the accuracy of identifying trend changes in time series data.

[0125] S104. Classify the seismic signal data according to the difference characteristics of the seismic signals.

[0126] Specifically, explosions and natural earthquakes are distinguished based on their duration and high-frequency component amplitude characteristics:

[0127] (1) Explosion is a transient event caused by human beings. The energy is released quickly and concentratedly, which is manifested as strong instantaneous fluctuations, especially short-period surface waves with large amplitudes. Therefore, when the difference characteristics of seismic signals show the characteristics of short time and high frequency band concentration, it is a blasting signal, such as Figure 2 For example, when the normalized waveform amplitude of the signal exceeds 1 at a certain moment, and the amplitude of the seismic signal waveform disappears instantly within a few seconds after this moment, the seismic signal presents the characteristics of short time and high frequency band concentration.

[0128] (2) In earthquakes, energy is more dispersed than in explosions, and there are still obvious fluctuations in the medium and low frequency IMFs, indicating that natural earthquakes have complex source mechanisms, the rupture process is relatively continuous, and energy is distributed in different frequency ranges. Therefore, when the difference characteristics of seismic signals show the characteristics of long time and high frequency dispersion, it is a seismic signal, such as Figure 3 For example, when the amplitude of the normalized waveform of the seismic signal increases instantly but does not exceed 1 at a certain moment, and after this moment, the seismic signal waveform gradually disappears within tens of minutes, that is, the seismic signal presents the characteristics of long time and high-frequency dispersion.

[0129] According to the embodiments of the present invention, the present invention has the following advantages and effects compared with the prior art:

[0130] (1) The original seismic signals generated by natural earthquakes and explosions are decomposed using the adaptive local iterative filtering (ALIF) technology, which can efficiently extract the multi-frequency characteristics of the signals at different scales, thereby enhancing the model's expression of signal characteristics and improving its ability to analyze complex seismic signals.

[0131] (2) The ALIF-BiLSTM hybrid model can more effectively identify the temporal correlation and multi-scale characteristics in seismic signals, improve the sensitivity to signal information in different frequency bands, enhance the classification performance of complex seismic signals, and thus improve the accuracy of distinguishing natural earthquakes from explosion events.

[0132] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described order of actions, because according to the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0133] The above is an introduction to a method embodiment. The following is a further explanation of the scheme of the present invention through a device embodiment having the same inventive concept as the method in the aforementioned embodiment.

[0134] like Figure 4 As shown, the device 400 includes:

[0135] The acquisition module 410 is used to acquire original seismic signal sample data, pre-process the original seismic signal sample data, and obtain seismic signal data.

[0136] The calculation module 420 is used to input the seismic signal data into the adaptive local iterative filtering model, output the residual IMF component, and then calculate the variance contribution rate of the residual IMF component.

[0137] The feature extraction module 430 is used to construct a bidirectional long short-term memory neural network model, input the variance contribution rate of each IMF component into the bidirectional long short-term memory neural network model, and extract the difference characteristics of the seismic signal.

[0138] The classification module 440 is used to classify the seismic signal data according to the difference characteristics of the seismic signals.

[0139] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the described module can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0140] In the technical solution of the present invention, the acquisition, storage and application of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0141] According to an embodiment of the present invention, the present invention further provides an electronic device and a readable storage medium.

[0142] Figure 5 A schematic block diagram of an electronic device 500 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0143] The electronic device 500 includes a computing unit 501, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 502 or a computer program loaded from a storage unit 508 into a random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the electronic device 500 can also be stored. The computing unit 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0144] Multiple components in the electronic device 500 are connected to the I / O interface 505, including: an input unit 506, such as a keyboard, a mouse, etc.; an output unit 507, such as various types of displays, speakers, etc.; a storage unit 508, such as a disk, an optical disk, etc.; and a communication unit 509, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 509 allows the electronic device 500 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0145] The computing unit 501 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 501 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 501 performs the various methods and processes described above, such as methods S101 to S104. For example, in some embodiments, methods S101 to S104 may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 508. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 500 via the ROM 502 and / or the communication unit 509. When the computer program is loaded into the RAM 503 and executed by the computing unit 501, one or more steps of the methods S101 to S104 described above may be performed. Alternatively, in other embodiments, the computing unit 501 may be configured to execute methods S101 to S104 in any other appropriate manner (eg, by means of firmware).

[0146] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0147] The program code for implementing the method of the present invention can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code can be executed entirely on the machine, partially on the machine, partially on the machine as a stand-alone software package and partially on a remote machine, or entirely on a remote machine or server.

[0148] In the context of the present invention, a machine-readable medium may be a tangible medium that may contain or store a program for use by an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0149] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0150] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0151] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0152] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0153] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A seismic signal classification research method, characterized in that: include: Acquire original seismic signal sample data, and preprocess the original seismic signal sample data to obtain seismic signal data; Inputting the seismic signal data into an adaptive local iterative filtering model, outputting the residual IMF component, and then calculating the variance contribution rate of the residual IMF component; A bidirectional long short-term memory neural network model is constructed, and the variance contribution rate of each IMF component is input into the bidirectional long short-term memory neural network model to extract the difference characteristics of seismic signals; The seismic signal data is classified according to the difference characteristics of the seismic signals.

2. The method according to claim 1, characterized in that The adaptive local iterative filtering model includes an inner loop module and an outer loop module; The inner loop module is used to traverse the seismic signal data and extract the effective IMF components; The outer loop module is used to extract the remaining IMF components by cyclically calling the valid IMF components obtained by the inner loop module.

3. The method according to claim 2, characterized in that The traversing of the seismic signal data and extracting effective IMF components includes: Seismic signal data sequences are extracted from seismic signal data in turn, and the seismic signal data sequences are convolved with an adaptive filter function to obtain a sliding operator; a wave operator is then obtained based on the seismic signal data sequence and the sliding operator; finally, the correlation coefficient between the sliding operator and the wave operator is calculated, and when the correlation coefficient is greater than or equal to a threshold, the seismic signal data sequence is extracted as a valid IMF component and the next iteration is performed, otherwise the next iteration is performed directly until all seismic signal data are traversed.

4. The method according to claim 3, characterized in that The calculation formula of the sliding operator is as follows: The calculation formula of the wave operator is as follows: v(t)=g(t)-μ(g(t)) The correlation coefficient calculation formula of the sliding operator and the fluctuation operator is as follows: Wherein, μ(g(t)) is the sliding operator, where g(t) is the seismic signal data sequence; i is the filtering interval; t is the time; τ is the delay time; ω(t) is the adaptive filtering function in the filtering interval i, which is composed of the solutions of all Fokker-Planck differential equations in the filtering interval i; d is the point integral symbol; v(t) is the wave operator; ρ μv is the correlation coefficient between the sliding operator and the fluctuation operator, that is, the basic statistic of regression analysis; T is the time period.

5. The method according to claim 2, characterized in that: The effective IMF components obtained by cyclically calling the inner loop module and extracting the remaining IMF components include: After each call to the inner loop module, the remaining IMF components are extracted according to the valid IMF components output by the inner loop module. When the outer loop module call time reaches the time period T, the outer loop is terminated and the remaining IMF components are output. Otherwise, the remaining IMF components are used as input data of the inner loop module to continue iterative extraction.

6. The method according to claim 5, characterized in that The extraction formula of the residual IMF component is as follows: e(t)=g(t)-r(t) Among them, e(t) is the residual IMF component; r(t) is the effective IMF component extracted by the inner loop module.

7. The method according to claim 6, characterized in that Calculate the variance contribution of the remaining IMF components, including: Among them, ψ i is the variance contribution rate; a is the residual IMF component index; S is the total number of residual IMF components; e c (t) is the cth residual IMF component; c is the seismic signal data index; N is the total number of seismic signal data.

8. The method according to claim 1, characterized in that: The bidirectional long short-term memory neural network model is constructed, and the variance contribution rate of each IMF component is used as the input data of the bidirectional long short-term memory neural network model to extract the difference characteristics of the seismic signal, including: The bidirectional long short-term memory neural network model includes: a forward LSTM module, a reverse LSTM module, and a difference feature extraction module; The forward LSTM module comprises: y f,t =g(W y'f h t-1 +ψ i e(t)+b y' ) The reverse LSTM module comprises: y b,t =g(W y'b h t-1 +ψ i e(t)+b y' ) The difference feature extraction module comprises: y' t =g(W y'f y f,t +W y'b y b,t +b y' ) Among them, h t-1 is the result of the IMF at the previous moment; y' t is the difference feature of seismic signal; g is the activation function of output layer; W y'f is the forward weight matrix; y f,t is the output of the forward LSTM unit; W y'b is the inverse weight matrix; y b,t is the output of the reverse LSTM unit; b y' is the bias term.

9. A seismic signal classification research device, characterized in that: include: An acquisition module is used to acquire original seismic signal sample data, and pre-process the original seismic signal sample data to obtain seismic signal data; A calculation module, used for inputting the seismic signal data into an adaptive local iterative filtering model, outputting a residual IMF component, and then calculating a variance contribution rate of the residual IMF component; The feature extraction module is used to construct a bidirectional long short-term memory neural network model, input the variance contribution rate of each IMF component into the bidirectional long short-term memory neural network model, and extract the difference characteristics of the seismic signal; The classification module is used to classify seismic signal data according to the difference characteristics of seismic signals.

10. An electronic device comprising at least one processor; and a memory connected to the at least one processor in communication; characterized in that: The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 8.

Images