Seismic wave and sound wave multi-channel real-time fusion processing method and device

Abstract

The invention provides a seismic wave and sound wave multichannel real-time fusion processing method and device, and the method comprises the steps: obtaining a seismic data set collected by at least two collection points in multiple earthquakes, and dividing the seismic data set into a first sub-seismic data set and a second sub-seismic data set; inputting the first sub-seismic data set into a preset model for training to obtain a trained seismic information acquisition model; and acquiring a real-time seismic data set, and inputting the real-time seismic data set into the seismic information acquisition model to obtain real-time seismic information. The method has the beneficial effects that the accuracy and comprehensiveness of seismic information extraction are improved by fusing seismic wave and sound wave data, seismic data can be acquired and processed in real time, timely information support is provided for seismic early warning and disaster assessment, a confrontation training mechanism of the GAN model is adopted, model parameters are continuously optimized, the performance and robustness of the model are improved, and the method is suitable for large-scale popularization and application. A new technical means is provided, and the earthquake disaster coping capacity can be improved, and disaster loss can be reduced.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and in particular to a multi-channel real-time fusion processing method and device for seismic waves and sound waves. Background Art

[0002] In the field of seismic exploration and monitoring, it is crucial to obtain accurate and timely information on underground geological structures and seismic activity. Traditional seismic data processing methods mainly rely on a single type of data, such as seismic wave data, while ignoring other data sources that may contain valuable information, such as sound wave data. This single data source processing method often limits the accuracy and comprehensiveness of seismic information extraction.

[0003] With the advancement of science and technology and the continuous development of exploration technology, people have begun to try to fuse multiple types of data to improve the extraction of seismic information. As two different types of wave signals, seismic waves and sound waves have different propagation characteristics in underground media, but they both carry information about underground geological structures. Therefore, fusing seismic wave data and sound wave data is expected to improve the resolution and imaging effect of seismic exploration.

[0004] However, the fusion processing of multi-channel data is not an easy task. There may be differences and inconsistencies between different data sources. How to effectively fuse these data and extract accurate earthquake information is a major challenge in the field of seismic exploration. In addition, real-time earthquake monitoring and early warning have put forward higher requirements on the speed and accuracy of data processing, and traditional processing methods often cannot meet these requirements. Summary of the invention

[0005] The main purpose of the present invention is to provide a method and device for real-time fusion processing of multi-channel seismic waves and sound waves, aiming to solve the problem that it is impossible to fuse multi-channel data for real-time monitoring and early warning of earthquakes.

[0006] The present invention provides a multi-channel real-time fusion processing method for seismic waves and acoustic waves, comprising:

[0007] Acquire a seismic data set collected from at least two collection points during multiple earthquakes; wherein the seismic data set includes corresponding actual earthquake information, and seismic trace data and acoustic wave data from at least two collection points;

[0008] Dividing each group of the seismic data sets into a first sub-seismic data set and a second sub-seismic data set; wherein the first sub-seismic data set and the second sub-seismic data set both contain corresponding actual earthquake information;

[0009] The first sub-seismic data set is input into a generative model of a preset model, and the second sub-seismic data set is input into a discriminative model of a preset model, and adversarial training is performed to obtain a seismic information acquisition model after training; wherein the preset model adopts the GAN model principle, and the generative model and the discriminative model confront each other and continuously adjust their own parameters;

[0010] Acquire a real-time seismic data set currently being collected, and divide the real-time seismic data set into a first sub-real-time seismic data set and a second sub-real-time seismic data set;

[0011] The first sub-real-time seismic data set and the second sub-real-time seismic data set are input into the seismic information acquisition model to obtain real-time seismic information.

[0012] Furthermore, the step of inputting the first sub-seismic data set into a generation model of a preset model, and inputting the second sub-seismic data set into a discrimination model of a preset model, performing adversarial training, and obtaining a seismic information acquisition model after training includes:

[0013] The first sub-trace data v 1i Input into the generative model to obtain the best prediction value r i , the actual earthquake information r true Input into the generative model, through the formula Perform initial training on the generative model and obtain the temporary prediction value r after training j and a temporary generative model,

[0014] The second sub-trace data v 2i Input into the discriminant model, through the formula The discriminant model is initially trained to obtain a temporary discriminant model; wherein, θ represents the parameter set of the generative model, represents the parameter set of the discriminant model;

[0015] According to the formula The temporary generation model and the temporary discrimination model are trained twice, and the earthquake information acquisition model is obtained after the training is completed; wherein It means taking the minimum value of θ under the premise of satisfying the above formula and The maximum value of .

[0016] Furthermore, the step of obtaining seismic data sets collected from at least two collection points in multiple earthquakes includes:

[0017] Obtaining original data sets of seismic gathers from at least two acquisition points;

[0018] The seismic data set is obtained by performing data consistency processing and anisotropy correction processing on the original data set of the seismic gather.

[0019] Furthermore, before the step of inputting the first sub-seismic data set into a generation model of a preset model and inputting the second sub-seismic data set into a discrimination model of a preset model for adversarial training to obtain a seismic information acquisition model after training, the method further includes:

[0020] Determining whether there are seismic trace data collected by multiple directional collection points in the first sub-seismic data set;

[0021] If there are seismic trace data collected by multiple directional collection points, the preset weights of the collection points in each direction are obtained;

[0022] The data of each seismic channel are weighted according to the preset weights to obtain a weighted first sub-seismic data set for input into the generation model of the preset model.

[0023] Furthermore, the step of inputting the first sub-seismic data set into a generation model of a preset model comprises:

[0024] The seismic trace data and acoustic wave data in the first sub-seismic data set are input into the input channel of the generation model of the preset model, and the actual seismic information in the first sub-seismic data set is used as the output channel of the generation model of the preset model.

[0025] The present invention also provides a multi-channel real-time fusion processing device for seismic waves and acoustic waves, comprising:

[0026] A first acquisition module is used to acquire a seismic data set collected from at least two acquisition points in multiple earthquakes; wherein the seismic data set includes corresponding actual earthquake information, and seismic trace data and acoustic wave data of at least two acquisition points;

[0027] A division module, used for dividing each group of the seismic data sets into a first sub-seismic data set and a second sub-seismic data set; wherein the first sub-seismic data set and the second sub-seismic data set both contain corresponding actual earthquake information;

[0028] A first input module is used to input the first sub-seismic data set into a generation model of a preset model, and input the second sub-seismic data set into a discriminant model of a preset model, to perform adversarial training, and obtain a seismic information acquisition model after training; wherein the preset model adopts the GAN model principle, and the generation model and the discriminant model confront each other and continuously adjust their own parameters;

[0029] A second acquisition module is used to acquire a real-time seismic data set currently collected, and divide the real-time seismic data set into a first sub-real-time seismic data set and a second sub-real-time seismic data set;

[0030] The second input module is used to input the first sub-real-time seismic data set and the second sub-real-time seismic data set into the seismic information acquisition model to obtain real-time seismic information.

[0031] Furthermore, the first input module includes:

[0032] An input submodule is used to input the first sub-trace data v 1i Input into the generative model to obtain the best prediction value r i , the actual earthquake information r true Input into the generative model, through the formula Perform initial training on the generative model and obtain the temporary prediction value r after training j and a temporary generative model,

[0033] The second sub-trace data v 2i Input into the discriminant model, through the formula The discriminant model is initially trained to obtain a temporary discriminant model; wherein, θ represents the parameter set of the generative model, represents the parameter set of the discriminant model;

[0034] The secondary training submodule is used to The temporary generation model and the temporary discrimination model are trained twice, and the earthquake information acquisition model is obtained after the training is completed; wherein It means taking the minimum value of θ under the premise of satisfying the above formula and The maximum value of .

[0035] Furthermore, the first acquisition module includes:

[0036] The original data set acquisition submodule is used to acquire the original data sets of seismic gathers of at least two acquisition points;

[0037] The processing submodule is used to perform data consistency processing and anisotropy correction processing on the original data set of the seismic gather to obtain the seismic data set.

[0038] Furthermore, the multi-channel real-time fusion processing device for seismic waves and acoustic waves further includes:

[0039] A seismic trace data judging module, used for judging whether there are seismic trace data collected by multiple directional collection points in the first sub-seismic data set;

[0040] A weight acquisition module is used to obtain the preset weights of the acquisition points in each direction if there are seismic trace data collected by acquisition points in multiple directions;

[0041] The weighted processing module is used to perform weighted processing on each seismic channel data according to the preset weight to obtain a weighted first sub-seismic data set for input into the generation model of the preset model.

[0042] Furthermore, the first input module includes:

[0043] A data input submodule is used to input the seismic trace data and acoustic wave data in the first sub-seismic data set into the input channel of the generation model of the preset model, and to use the actual seismic information in the first sub-seismic data set as the output channel of the generation model of the preset model.

[0044] The beneficial effects of the present invention are as follows: by fusing seismic wave and sound wave data, the accuracy and comprehensiveness of earthquake information extraction are improved, and earthquake data can be acquired and processed in real time, providing timely information support for earthquake early warning and disaster assessment. The adversarial training mechanism of the GAN model is adopted to continuously optimize the model parameters, improve the performance and robustness of the model, and provide a new technical means to help improve the ability to respond to earthquake disasters and reduce disaster losses. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a flow chart of a multi-channel real-time fusion processing method of seismic waves and acoustic waves according to an embodiment of the present invention;

[0046] Figure 2 It is a schematic block diagram of the structure of a multi-channel real-time fusion processing device for seismic waves and acoustic waves according to an embodiment of the present invention;

[0047] Figure 3 A schematic block diagram of the structure of a computer device according to an embodiment of the present application.

[0048] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.

[0050] It should be noted that all directional indications in the embodiments of the present invention (such as up, down, left, right, front, back, etc.) are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly. The connection can be a direct connection or an indirect connection.

[0051] The term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there may be three relationships. For example, A and B can represent three situations: A exists alone, A and B exist at the same time, and B exists alone.

[0052] In addition, in the present invention, descriptions such as "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0053] Reference Figure 1 The present invention proposes a multi-channel real-time fusion processing method for seismic waves and acoustic waves, comprising:

[0054] S1: Acquire a seismic data set collected from at least two collection points during multiple earthquakes; wherein the seismic data set includes corresponding actual earthquake information, and seismic trace data and acoustic wave data of at least two collection points;

[0055] S2: dividing each group of the seismic data sets into a first sub-seismic data set and a second sub-seismic data set; wherein the first sub-seismic data set and the second sub-seismic data set both contain corresponding actual earthquake information;

[0056] S3: Input the first sub-seismic data set into the generative model of the preset model, and input the second sub-seismic data set into the discriminative model of the preset model, and perform adversarial training to obtain a seismic information acquisition model after training; wherein the preset model adopts the GAN model principle, and the generative model and the discriminative model confront each other and continuously adjust their own parameters;

[0057] S4: Acquire a currently acquired real-time seismic data set, and divide the real-time seismic data set into a first sub-real-time seismic data set and a second sub-real-time seismic data set;

[0058] S5: Inputting the first sub-real-time seismic data set and the second sub-real-time seismic data set into the seismic information acquisition model to obtain real-time seismic information.

[0059] As described in step S1 above, a seismic data set collected by at least two acquisition points in multiple earthquakes is obtained; wherein the seismic data set includes corresponding actual seismic information, and seismic trace data and acoustic wave data of at least two acquisition points. A seismic data set collected by at least two acquisition points in multiple earthquakes is obtained. These data sets include not only seismic trace data, but also acoustic wave data, and all correspond to actual seismic information. Wherein, the acquisition of seismic data and acoustic wave data can be acquired by a preset sensor, and the acquisition of seismic data can be a seismic detector, specifically a three-component detector (capable of simultaneously detecting seismic waves in the vertical direction, along the survey line direction and in the direction perpendicular to the survey line), and the sensor for acquiring acoustic wave data can be one or more of an acoustic wave sensor, an ultrasonic sensor and a distributed optical fiber acoustic wave sensor, and multiple acquisition points can be set, for example, three acquisition points are set, in a preferred embodiment, the three acquisition points cannot be located on the same straight line, and need to ensure a certain distance between them, so as to ensure the detection accuracy, it should be noted that each acquisition point can acquire a set of seismic data and a set of acoustic wave data, and a set of seismic data is three-dimensional seismic data.

[0060] As described in the above step S2, each group of the seismic data sets is divided into a first sub-seismic data set and a second sub-seismic data set; wherein, the first sub-seismic data set and the second sub-seismic data set both contain corresponding actual earthquake information; wherein, both sub-data sets contain corresponding actual earthquake information for subsequent model training and confrontation, and the division method can be an average distribution, or the size of each data can be pre-detected, and the group of data with the largest data is set as the first sub-seismic data set according to the size, and the rest are set as the second sub-seismic data set. Since the data is large, it means that an earthquake is likely to have occurred at this acquisition point. Therefore, using it as the input for subsequent model generation can effectively improve the accuracy of model prediction.

[0061] As described in step S3 above, the first sub-seismic data set is input into the generation model of the preset model, and the second sub-seismic data set is input into the discriminant model of the preset model, and adversarial training is performed to obtain a seismic information acquisition model after training; wherein, the preset model adopts the GAN model principle, and the generation model and the discriminant model are in conflict with each other and continuously adjust their own parameters. A preset model is constructed using the GAN model principle, and the model includes a generation model and a discriminant model. The first sub-seismic trace data set is input into the generation model, and the second sub-seismic trace data set is input into the discriminant model for adversarial training. During the adversarial training process, the generation model and the discriminant model are in conflict with each other, and their own parameters are continuously adjusted to improve the accuracy and robustness of the model. Finally, a seismic information acquisition model after training is obtained.

[0062] As described in the above steps S4-S5, the real-time seismic data set currently collected is obtained, and the real-time seismic data set is divided into a first sub-real-time seismic data set and a second sub-real-time seismic data set, and the first sub-real-time seismic data set and the second sub-real-time seismic data set are input into the seismic information acquisition model to obtain real-time seismic information. The real-time seismic trace data set collected by the two current acquisition points is obtained, and it is divided into a first sub-real-time seismic trace data set and a second sub-real-time seismic trace data set. The two sub-real-time seismic trace data sets are input into the seismic information acquisition model after training to extract and fuse real-time seismic information. By fusing seismic wave and sound wave data, the accuracy and comprehensiveness of seismic information extraction are improved, and seismic data can be obtained and processed in real time, providing timely information support for earthquake early warning and disaster assessment. The adversarial training mechanism of the GAN model is adopted to continuously optimize model parameters, improve the performance and robustness of the model, and provide a new technical means to help improve the ability to respond to earthquake disasters and reduce disaster losses.

[0063] In one embodiment, the step S3 of inputting the first sub-seismic data set into a generation model of a preset model, and inputting the second sub-seismic data set into a discrimination model of a preset model, performing adversarial training, and obtaining a seismic information acquisition model after training includes:

[0064] S301: The first sub-trace data v 1i Input into the generative model to obtain the best prediction value r i , the actual earthquake information r true Input into the generative model, through the formula Perform initial training on the generative model and obtain the temporary prediction value r after training j and a temporary generative model,

[0065] The second sub-trace data v 2iInput into the discriminant model, through the formula The discriminant model is initially trained to obtain a temporary discriminant model; wherein, θ represents the parameter set of the generative model, represents the parameter set of the discriminant model;

[0066] S302: According to the formula The temporary generation model and the temporary discrimination model are trained twice, and the earthquake information acquisition model is obtained after the training is completed; wherein It means taking the minimum value of θ under the premise of satisfying the above formula and The maximum value of .

[0067] As described in the above steps S301-S302, the training of the seismic information acquisition model is realized. Specifically, the training method is to use the stochastic gradient descent method for training and updating, that is, the training of the next sample is carried out after the training of the current sample is completed. After each training is completed, the parameter set is updated to complete the training of the initial generation model. That is, through the formula The generative model is initially trained by the formula The discriminant model is initially trained, and the parameter set is updated after each training is completed, so as to complete the training of the generative model and the discriminant model, and then according to the formula

[0068] The generation model and the discriminant model are trained twice. It should be noted that each training sample needs to be trained with the above three formulas, that is, in the training process of a set of training samples, the parameters need to be updated twice. Finally, the temporary generation model parameter set θ and the temporary discriminant model parameter set are obtained. In order to make the discrimination effect of the seismic information acquisition model better, the temporary generated model parameter set θ should be minimized as much as possible, and the temporary discrimination model parameter set Take the maximum value.

[0069] In one embodiment, the step S1 of acquiring seismic data sets collected from at least two collection points in multiple earthquakes includes:

[0070] S101: Acquire original data sets of seismic gathers from at least two acquisition points;

[0071] S102: Performing data consistency processing and anisotropy correction processing on the original data set of the seismic gather to obtain the seismic data set.

[0072] As described in the above steps S101-S102, collecting seismic data from different acquisition points is to obtain more comprehensive underground structural information. Since the propagation of seismic waves in different directions will be affected by different geological structures, data from multiple acquisition points can provide richer information for subsequent analysis. Specifically, seismic waves are recorded by seismic detectors (or seismographs) arranged at different acquisition points. These detectors may be located on the ground, on the seabed or in a well, depending on specific exploration needs. In order to eliminate data inconsistencies caused by instrument differences, environmental factors or improper operation. Specifically, steps such as data standardization, denoising, and time calibration are included to ensure that all data sets are compared and analyzed under the same standards. Differences in seismic wave propagation caused by the anisotropy of underground media (i.e., different physical properties of the medium in different directions). Specifically, mathematical models or physical simulations can be applied to estimate and correct such differences to improve the accuracy and reliability of the data.

[0073] In one embodiment, before step S3 of inputting the first sub-seismic data set into a generation model of a preset model and inputting the second sub-seismic data set into a discrimination model of a preset model for adversarial training to obtain a seismic information acquisition model after training, the method further includes:

[0074] S201: Determine whether there are seismic trace data collected by multiple directional collection points in the first sub-seismic data set;

[0075] S202: If there are seismic trace data collected by multiple directional collection points, obtain a preset weight of each directional collection point;

[0076] S203: performing weighted processing on each seismic trace data according to the preset weight to obtain a weighted first sub-seismic data set for input into the generation model of the preset model.

[0077] As described in the above steps S201-S203, weighted processing of each seismic channel data is implemented to facilitate input into the preset model. If there are seismic channel data collected by multiple directional collection points, the preset weights of the collection points in each direction are obtained, that is, preset weights are set in advance for the collection points in each direction. The preset weights are related to the size of the seismic channel data. For example, if the first sub-seismic data contains data in the due north and due east, the preset weights can be set according to the data size, or the weights can be set in proportion. Weighted processing is performed on each seismic channel data according to the preset weights to obtain a weighted first sub-seismic data set for input into the generation model of the preset model. Similarly, preset weights can also be set for the seismic data in the second sub-seismic data set.

[0078] In one embodiment, the step S3 of inputting the first sub-seismic data set into a generation model of a preset model comprises:

[0079] S311: The seismic trace data and acoustic wave data in the first sub-seismic data set are input into the input channel of the generation model of the preset model, and the actual seismic information in the first sub-seismic data set is used as the output channel of the generation model of the preset model.

[0080] As described in the above step S311, the setting of the input channel and output channel of the generation model is realized, that is, the seismic channel data and the acoustic wave data are input into the input channel of the generation model of the preset model, and the actual seismic information in the first sub-seismic data set is used as the output channel of the generation model of the preset model, thereby reducing the calculation amount of the generation model, improving the training efficiency of the generation model, and reducing its calculation amount.

[0081] Reference Figure 2 The present invention also provides a multi-channel real-time fusion processing device for seismic waves and acoustic waves, comprising:

[0082] A first acquisition module 10 is used to acquire a seismic data set collected from at least two acquisition points in multiple earthquakes; wherein the seismic data set includes corresponding actual earthquake information, and seismic trace data and acoustic wave data of at least two acquisition points;

[0083] A division module 20, for dividing each group of the seismic data sets into a first sub-seismic data set and a second sub-seismic data set; wherein the first sub-seismic data set and the second sub-seismic data set both contain corresponding actual seismic information;

[0084] The first input module 30 is used to input the first sub-seismic data set into the generation model of the preset model, and input the second sub-seismic data set into the discrimination model of the preset model, to perform adversarial training, and obtain a seismic information acquisition model after training; wherein the preset model adopts the GAN model principle, the generation model and the discrimination model confront each other and continuously adjust their own parameters;

[0085] A second acquisition module 40 is used to acquire a currently acquired real-time seismic data set, and divide the real-time seismic data set into a first sub-real-time seismic data set and a second sub-real-time seismic data set;

[0086] The second input module 50 is used to input the first sub-real-time seismic data set and the second sub-real-time seismic data set into the seismic information acquisition model to obtain real-time seismic information.

[0087] In one embodiment, the first input module 30 comprises:

[0088] An input submodule is used to input the first sub-trace data v 1i Input into the generative model to obtain the best prediction value r i , the actual earthquake information r true Input into the generative model, through the formula Perform initial training on the generative model and obtain the temporary prediction value r after training j and a temporary generative model,

[0089] The second sub-trace data v 2i Input into the discriminant model, through the formula The discriminant model is initially trained to obtain a temporary discriminant model; wherein, θ represents the parameter set of the generative model, represents the parameter set of the discriminant model;

[0090] The secondary training submodule is used to The temporary generation model and the temporary discrimination model are trained twice, and the earthquake information acquisition model is obtained after the training is completed; wherein It means taking the minimum value of θ under the premise of satisfying the above formula and The maximum value of .

[0091] In one embodiment, the first acquisition module 10 includes:

[0092] The original data set acquisition submodule is used to acquire the original data sets of seismic gathers of at least two acquisition points;

[0093] The processing submodule is used to perform data consistency processing and anisotropy correction processing on the original data set of the seismic gather to obtain the seismic data set.

[0094] In one embodiment, the multi-channel real-time fusion processing device for seismic waves and acoustic waves further includes:

[0095] A seismic trace data judging module, used for judging whether there are seismic trace data collected by multiple directional collection points in the first sub-seismic data set;

[0096] A weight acquisition module is used to obtain the preset weights of the acquisition points in each direction if there are seismic trace data collected by acquisition points in multiple directions;

[0097] The weighted processing module is used to perform weighted processing on each seismic channel data according to the preset weight to obtain a weighted first sub-seismic data set for input into the generation model of the preset model.

[0098] In one embodiment, the first input module 30 includes:

[0099] A data input submodule is used to input the seismic trace data and acoustic wave data in the first sub-seismic data set into the input channel of the generation model of the preset model, and to use the actual seismic information in the first sub-seismic data set as the output channel of the generation model of the preset model.

[0100] The beneficial effects of the present invention are as follows: by fusing seismic wave and sound wave data, the accuracy and comprehensiveness of earthquake information extraction are improved, and earthquake data can be acquired and processed in real time, providing timely information support for earthquake early warning and disaster assessment. The adversarial training mechanism of the GAN model is adopted to continuously optimize the model parameters, improve the performance and robustness of the model, and provide a new technical means to help improve the ability to respond to earthquake disasters and reduce disaster losses.

[0101] Reference Figure 3 In an embodiment of the present application, a computer device is also provided. The computer device may be a server, and its internal structure may be as follows: Figure 3 As shown. The computer device includes a processor, a memory, a network interface and a database connected through a system bus. Among them, the processor designed by the computer is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store various seismic trace data and sound wave data, etc. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the multi-channel real-time fusion processing method of seismic waves and sound waves described in any of the above embodiments can be implemented.

[0102] Those skilled in the art will understand that Figure 3 The structure shown in is merely a block diagram of a portion of the structure related to the present application solution and does not constitute a limitation on the computer device to which the present application solution is applied.

[0103] An embodiment of the present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the multi-channel real-time fusion processing method of seismic waves and sound waves described in any of the above embodiments can be implemented.

[0104] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided in this application and used in the embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0105] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, device, article or method including the element.

[0106] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0107] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0108] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention shall be included in the scope of the claims of the present invention.

Claims

1. A multi-channel real-time fusion processing method for seismic waves and acoustic waves, characterized in that: include: Acquire a seismic data set collected from at least two collection points during multiple earthquakes; wherein the seismic data set includes corresponding actual earthquake information, and seismic trace data and acoustic wave data from at least two collection points; Dividing each group of the seismic data sets into a first sub-seismic data set and a second sub-seismic data set; wherein the first sub-seismic data set and the second sub-seismic data set both contain corresponding actual earthquake information; The first sub-seismic data set is input into a generative model of a preset model, and the second sub-seismic data set is input into a discriminative model of a preset model, and adversarial training is performed to obtain a seismic information acquisition model after training; wherein the preset model adopts the GAN model principle, and the generative model and the discriminative model confront each other and continuously adjust their own parameters; Acquire a real-time seismic data set currently being collected, and divide the real-time seismic data set into a first sub-real-time seismic data set and a second sub-real-time seismic data set; The first sub-real-time seismic data set and the second sub-real-time seismic data set are input into the seismic information acquisition model to obtain real-time seismic information.

2. The multi-channel real-time fusion processing method of seismic waves and acoustic waves according to claim 1, characterized in that: The step of inputting the first sub-seismic data set into a generation model of a preset model, and inputting the second sub-seismic data set into a discrimination model of a preset model, performing adversarial training, and obtaining a seismic information acquisition model after training, comprises: The first sub-trace data v 1i Input into the generative model to obtain the best prediction value r i , the actual earthquake information r true Input into the generative model, through the formula Perform initial training on the generative model and obtain the temporary prediction value r after training j and a temporary generative model, The second sub-trace data v 2i Input into the discriminant model, through the formula The discriminant model is initially trained to obtain a temporary discriminant model; wherein, θ represents the parameter set of the generative model, represents the parameter set of the discriminant model; According to the formula The temporary generation model and the temporary discrimination model are trained twice, and the earthquake information acquisition model is obtained after the training is completed; wherein It means taking the minimum value of θ under the premise of satisfying the above formula and The maximum value of .

3. The multi-channel real-time fusion processing method of seismic waves and acoustic waves according to claim 1, characterized in that: The step of obtaining seismic data sets collected from at least two collection points in multiple earthquakes includes: Obtaining original data sets of seismic gathers from at least two acquisition points; The seismic data set is obtained by performing data consistency processing and anisotropy correction processing on the original data set of the seismic gather.

4. The multi-channel real-time fusion processing method of seismic waves and acoustic waves according to claim 1, characterized in that: Before the step of inputting the first sub-seismic data set into a generation model of a preset model and inputting the second sub-seismic data set into a discrimination model of a preset model to perform adversarial training to obtain a seismic information acquisition model after training, the method further includes: Determining whether there are seismic trace data collected by multiple directional collection points in the first sub-seismic data set; If there are seismic trace data collected by multiple directional collection points, the preset weights of the collection points in each direction are obtained; The data of each seismic channel are weighted according to the preset weights to obtain a weighted first sub-seismic data set for input into the generation model of the preset model.

5. The multi-channel real-time fusion processing method of seismic waves and acoustic waves according to claim 1, characterized in that: The step of inputting the first sub-seismic data set into a generation model of a preset model comprises: The seismic trace data and acoustic wave data in the first sub-seismic data set are input into the input channel of the generation model of the preset model, and the actual seismic information in the first sub-seismic data set is used as the output channel of the generation model of the preset model.

6. A multi-channel real-time fusion processing device for seismic waves and acoustic waves, characterized in that: include: A first acquisition module is used to acquire a seismic data set collected from at least two acquisition points in multiple earthquakes; wherein the seismic data set includes corresponding actual earthquake information, and seismic trace data and acoustic wave data of at least two acquisition points; A division module, used for dividing each group of the seismic data sets into a first sub-seismic data set and a second sub-seismic data set; wherein the first sub-seismic data set and the second sub-seismic data set both contain corresponding actual earthquake information; A first input module is used to input the first sub-seismic data set into a generation model of a preset model, and input the second sub-seismic data set into a discriminant model of a preset model, to perform adversarial training, and obtain a seismic information acquisition model after training; wherein the preset model adopts the GAN model principle, and the generation model and the discriminant model confront each other and continuously adjust their own parameters; A second acquisition module is used to acquire a currently acquired real-time seismic data set, and divide the real-time seismic data set into a first sub-real-time seismic data set and a second sub-real-time seismic data set; The second input module is used to input the first sub-real-time seismic data set and the second sub-real-time seismic data set into the seismic information acquisition model to obtain real-time seismic information.

7. The multi-channel real-time fusion processing device for seismic waves and acoustic waves according to claim 6, characterized in that: The first input module comprises: An input submodule is used to input the first sub-trace data v 1i Input into the generative model to obtain the best prediction value r i , the actual earthquake information r true Input into the generative model, through the formula Perform initial training on the generative model and obtain the temporary prediction value r after training j and a temporary generative model, The second sub-trace data v 2i Input into the discriminant model, through the formula The discriminant model is initially trained to obtain a temporary discriminant model; wherein, θ represents the parameter set of the generative model, represents the parameter set of the discriminant model; The secondary training submodule is used to The temporary generation model and the temporary discrimination model are trained twice, and the earthquake information acquisition model is obtained after the training is completed; wherein It means taking the minimum value of θ under the premise of satisfying the above formula and The maximum value of .

8. The multi-channel real-time fusion processing device for seismic waves and acoustic waves according to claim 6, characterized in that: The first acquisition module includes: The original data set acquisition submodule is used to acquire the original data sets of seismic gathers of at least two acquisition points; The processing submodule is used to perform data consistency processing and anisotropy correction processing on the original data set of the seismic gather to obtain the seismic data set.

9. The multi-channel real-time fusion processing device for seismic waves and acoustic waves according to claim 6, characterized in that: The multi-channel real-time fusion processing device for seismic waves and acoustic waves also includes: A seismic trace data judging module, used for judging whether there are seismic trace data collected by multiple directional collection points in the first sub-seismic data set; A weight acquisition module is used to obtain the preset weights of the acquisition points in each direction if there are seismic trace data collected by acquisition points in multiple directions; The weighted processing module is used to perform weighted processing on each seismic channel data according to the preset weight to obtain a weighted first sub-seismic data set for input into the generation model of the preset model.

10. The multi-channel real-time fusion processing device for seismic waves and acoustic waves according to claim 6, characterized in that: The first input module comprises: A data input submodule is used to input the seismic trace data and acoustic wave data in the first sub-seismic data set into the input channel of the generation model of the preset model, and to use the actual seismic information in the first sub-seismic data set as the output channel of the generation model of the preset model.

Images