Seismic data processing method and device, equipment and storage medium

Abstract

The invention discloses a seismic data processing method and device, equipment and a storage medium, and belongs to the technical field of seismic exploration. According to the method, first seismic data is predicted through a first neural network model to obtain first prediction data, and based on the first seismic data and the first prediction data, part of aliasing noise is separated from the first seismic data to obtain second seismic data. And predicting the second seismic data through a second neural network model to obtain second prediction data, and further separating aliasing noise from the first seismic data based on the first seismic data and the second prediction data to obtain third seismic data. And so on, until the data after aliasing noise removal is obtained through the Nth neural network model in the target neural network model. According to the method, the aliasing noise is gradually separated through iterative prediction, and the limitation of a plurality of assumed conditions is overcome, so that the separation effect of the aliasing noise is improved.

Description

Technical Field

[0001] The present application relates to the field of seismic exploration technology, and in particular to a seismic data processing method, device, equipment and storage medium. Background Art

[0002] Traditional seismic data acquisition is to excite shot by shot in sequence, but this is inefficient. Currently, most of the aliasing acquisition technology is used, which is to continuously excite multiple shots within the excitation cycle of a single shot, thereby greatly improving the efficiency of seismic data acquisition. However, during aliasing acquisition, for each detector, the signals of adjacent shots excited continuously are aliased together to form aliasing noise, which interferes with each other and affects the subsequent seismic data processing. Therefore, it is necessary to separate the aliased adjacent shot signals.

[0003] The related technology mainly uses the inversion method to separate adjacent shot signals. This method extracts common detection point gather data from aliased seismic data, then estimates the aliasing noise through sparse inversion, and finally removes the aliasing noise from the aliased seismic data to obtain a valid signal.

[0004] However, the methods in related technologies have many assumptions in the process of removing aliasing noise, such as linear approximation of aliased seismic data and sparse distribution of data features, which limits the separation effect. Summary of the invention

[0005] The embodiments of the present application provide a seismic data processing method, device, equipment and storage medium, which can gradually separate aliasing noise from seismic data, thereby improving the separation effect of aliasing noise. The technical solution is as follows:

[0006] In one aspect, a method for processing seismic data is provided, the method comprising:

[0007] Acquire first seismic data, where the first seismic data is common detection point gather data of a target work area;

[0008] Inputting the first seismic data into a first neural network model in a target neural network model to obtain first prediction data; wherein the target neural network model is obtained through N iterative training, one iterative training obtains one neural network model, and N is an integer greater than 1;

[0009] Determine second seismic data based on the first seismic data and the first predicted data, where the second seismic data is data after a portion of aliasing noise is separated;

[0010] Inputting the second seismic data into a second neural network model in the target neural network model to obtain second prediction data;

[0011] Determine third seismic data based on the first seismic data and the second predicted data, wherein the third seismic data is data after aliasing noise is further separated;

[0012] By analogy, the step of inputting the seismic data into the corresponding neural network model in the target neural network model is repeated until the target prediction data is obtained through the Nth neural network model in the target neural network model, and the target prediction data is the data after removing aliasing noise.

[0013] In a possible implementation, determining second seismic data based on the first seismic data and the first prediction data includes:

[0014] Performing aliasing and pseudo separation on the first prediction data to determine first noise data;

[0015] The first noise data is separated from the first seismic data to obtain the second seismic data.

[0016] In another possible implementation, the first seismic data includes seismic data corresponding to a plurality of seismic traces;

[0017] The performing aliasing and pseudo separation on the first prediction data to determine first noise data includes:

[0018] Extracting the excitation time corresponding to the seismic data of each seismic trace from the first seismic data;

[0019] According to the excitation time corresponding to each seismic trace, the seismic data corresponding to the plurality of seismic traces are superimposed on the first prediction data to obtain fourth seismic data;

[0020] The first prediction data is subtracted from the fourth seismic data to obtain the first noise data.

[0021] In another possible implementation, the training process of the target neural network model includes:

[0022] Acquire first sample seismic data and first sample label data; wherein the first sample seismic data is common detection point gather data, and the first sample label data is data after aliasing noise is removed;

[0023] Performing a first iterative training based on the first sample seismic data and the first sample label data to obtain the first neural network model;

[0024] Determining second sample seismic data based on the first sample seismic data and first output data, wherein the first output data is output data of the first neural network model;

[0025] Based on the second sample seismic data and the first sample label data, performing a second iterative training to obtain the second neural network model;

[0026] By analogy, the steps of iterative model training are repeated until the Nth iterative training is performed to obtain the Nth neural network model.

[0027] In another possible implementation, performing a second iteration training based on the second sample seismic data and the first sample label data to obtain the second neural network model includes:

[0028] Based on the second sample seismic data and the first sample label data, a second iterative training is performed on the basis of the first neural network model to obtain the second neural network model.

[0029] In another possible implementation, the obtaining of the first sample seismic data and the first sample label data includes:

[0030] Acquire shot gather data collected by aliasing in the target work area;

[0031] Extracting common detection point gather data from the shot gather data to obtain the first sample seismic data;

[0032] A sparse inversion method is used to remove aliasing noise in the first sample seismic data to obtain the first sample label data.

[0033] In another possible implementation, the method further includes:

[0034] Acquire first verification seismic data and first verification label data; wherein the first verification seismic data is common detection point gather data, and the first verification label data is data after aliasing noise is removed;

[0035] Based on the first verification seismic data and the first verification label data, verifying the N neural network models included in the target neural network model;

[0036] When the verification is passed, the step of inputting the first seismic data into the first neural network model in the target neural network model is executed.

[0037] In another aspect, a seismic data processing device is provided, the device comprising:

[0038] A first acquisition module is used to acquire first seismic data, where the first seismic data is common detection point gather data of a target work area;

[0039] A first input module is used to input the first seismic data into a first neural network model in a target neural network model to obtain first prediction data; wherein the target neural network model is obtained through N iterative training, one iterative training obtains one neural network model, and N is an integer greater than 1;

[0040] A first determination module is used to determine second seismic data based on the first seismic data and the first prediction data, where the second seismic data is data after a portion of aliasing noise is separated;

[0041] A second input module, used for inputting the second seismic data into a second neural network model in the target neural network model to obtain second prediction data;

[0042] The second determination module is used to determine the third seismic data based on the first seismic data and the second prediction data, and the third seismic data is the data after further separation of aliasing noise; and so on, repeatedly executing the step of inputting the seismic data into the corresponding neural network model in the target neural network model until the target prediction data is obtained through the Nth neural network model in the target neural network model, and the target prediction data is the data after removing aliasing noise.

[0043] In a possible implementation, the first determination module is used to perform aliasing and pseudo-separation on the first prediction data to determine first noise data; and separate the first noise data from the first seismic data to obtain the second seismic data.

[0044] In another possible implementation, the first seismic data includes seismic data corresponding to a plurality of seismic traces;

[0045] The first determination module is used to extract the excitation time corresponding to the seismic data of each seismic channel from the first seismic data; superimpose the seismic data corresponding to the multiple seismic channels on the basis of the first prediction data according to the excitation time corresponding to each seismic channel to obtain fourth seismic data; and subtract the first prediction data from the fourth seismic data to obtain the first noise data.

[0046] In another possible implementation, the device further includes:

[0047] A second acquisition module is used to acquire first sample seismic data and first sample label data; wherein the first sample seismic data is common detection point gather data, and the first sample label data is data after aliasing noise is removed;

[0048] A first training module, configured to perform a first iterative training based on the first sample seismic data and the first sample label data to obtain the first neural network model;

[0049] A third determination module, configured to determine second sample seismic data based on the first sample seismic data and first output data, wherein the first output data is output data of the first neural network model;

[0050] The second training module is used to perform a second iterative training based on the second sample seismic data and the first sample label data to obtain the second neural network model; and so on, repeatedly perform the steps of iterative training of the model until the Nth iterative training is performed to obtain the Nth neural network model.

[0051] In another possible implementation, the second training module is used to perform a second iterative training on the basis of the first neural network model based on the second sample seismic data and the first sample label data to obtain the second neural network model.

[0052] In another possible implementation, the second acquisition module is used to acquire shot gather data acquired through aliasing in the target work area; extract common detection point gather data from the shot gather data to obtain the first sample seismic data; and use a sparse inversion method to remove aliasing noise in the first sample seismic data to obtain the first sample label data.

[0053] In another possible implementation, the device further includes:

[0054] A third acquisition module is used to acquire first verification seismic data and first verification label data; wherein the first verification seismic data is common detection point gather data, and the first verification label data is data after aliasing noise is removed;

[0055] A verification module, used for verifying the N neural network models included in the target neural network model based on the first verification seismic data and the first verification label data;

[0056] The first input module is used to input the first seismic data into the first neural network model in the target neural network model when verification is passed.

[0057] On the other hand, an electronic device is provided, comprising a processor and a memory, wherein the memory stores at least one program code, and the at least one program code is loaded and executed by the processor to implement any of the above-mentioned seismic data processing methods.

[0058] On the other hand, a computer-readable storage medium is provided, in which at least one program code is stored. The at least one program code is loaded and executed by a processor to implement any of the above-mentioned seismic data processing methods.

[0059] On the other hand, a computer program product is provided, wherein at least one program code is stored in the computer program product, and the at least one program code is loaded and executed by a processor to implement any of the above-mentioned seismic data processing methods.

[0060] The embodiment of the present application provides a method for processing seismic data, which first predicts the first seismic data through the first neural network model in the target neural network model to obtain the first predicted data, and then separates part of the aliasing noise from the first seismic data based on the first seismic data and the first predicted data to obtain the second seismic data. The second seismic data is predicted through the second neural network model in the target neural network model to obtain the second predicted data, and then further separates the aliasing noise from the first seismic data based on the first seismic data and the second predicted data to obtain the third seismic data. And so on, until the data after the aliasing noise is removed is obtained through the Nth neural network model in the target neural network model. It can be seen that this method utilizes the powerful extraction ability of the neural network model for nonlinear features, gradually separates the aliasing noise through iterative prediction, and finally obtains the data after the aliasing noise is removed, which overcomes the limitations of many assumptions and conditions, thereby improving the separation effect of the aliasing noise.

[0061] It is to be understood that the foregoing general description and the following detailed description are exemplary only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 It is a schematic diagram of an implementation environment of a seismic data processing method provided in an embodiment of the present application;

[0063] Figure 2 is a flow chart of a seismic data processing method provided in an embodiment of the present application;

[0064] Figure 3 is a schematic diagram of first seismic data before aliasing noise is removed provided by an embodiment of the present application;

[0065] Figure 4 is a schematic diagram of aliasing noise obtained after the first iterative prediction provided by an embodiment of the present application;

[0066] Figure 5 is a schematic diagram of target prediction data after aliasing noise is removed provided in an embodiment of the present application;

[0067] Figure 6is a schematic diagram of removing aliasing noise through a target neural network model provided by an embodiment of the present application;

[0068] Figure 7 is a flow chart of a training target neural network model provided in an embodiment of the present application;

[0069] Figure 8 It is a schematic diagram of iterative training based on sample seismic data and sample label data provided in an embodiment of the present application;

[0070] Fig. 9 is a structural schematic diagram of a seismic data processing device provided in an embodiment of the present application;

[0071] Fig.10 It is a structural block diagram of a terminal provided in an embodiment of the present application;

[0072] Fig.11 It is a structural block diagram of a server provided in an embodiment of the present application. DETAILED DESCRIPTION

[0073] In order to make the technical solutions and advantages of the present application clearer, the implementation methods of the present application are described in further detail below.

[0074] The terms "first", "second", "third" and "fourth" etc. in the specification and claims of the present application and the drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices.

[0075] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, stored data, displayed data, etc.) and signals involved in this application are all authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions. For example, the earthquake data, prediction data, aliasing noise, etc. involved in this application are all obtained with full authorization.

[0076] Figure 1 is a schematic diagram of an implementation environment of a seismic data processing method provided in an embodiment of the present application, see Figure 1The implementation environment includes: an electronic device, which can be provided as a terminal 101, or can be provided as a terminal 101 and a server 102, without specific limitation.

[0077] If the electronic device is provided as terminal 101 , the terminal 101 processes the seismic data.

[0078] If the electronic device is provided as a terminal 101 and a server 102 , the terminal 101 sends the seismic data to the server 102 , and the server 102 processes the seismic data and then returns the processing result to the terminal 101 .

[0079] The terminal 101 is at least one of a mobile phone, a tablet computer, a PC (Personal Computer), an intelligent voice interaction device, a vehicle-mounted terminal, etc. The server 102 can be at least one of a server, a server cluster consisting of multiple servers, a cloud server, a cloud computing platform, and a virtualization center.

[0080] Figure 2 is a flowchart of a seismic data processing method provided in an embodiment of the present application, which is executed by an electronic device, see Figure 2 , the method comprising:

[0081] Step 201: The electronic device acquires first seismic data.

[0082] The first seismic data is the common detection point gather data of the target work area, and the target work area is the area where the reservoir to be predicted is located.

[0083] The process of the electronic device acquiring the first seismic data may be: the electronic device acquires the shot gather data collected by aliasing in the target work area, and extracts a plurality of common detection point gather data from the shot gather data. For each common detection point gather data, the common detection point gather data is determined as the first seismic data. Each common detection point gather data, i.e., the first seismic data, includes seismic data corresponding to a plurality of seismic channels. Figure 3 , Figure 3 This is a schematic diagram corresponding to the first seismic data before removing aliasing noise.

[0084] Step 202: The electronic device inputs the first seismic data into the first neural network model in the target neural network model to obtain first prediction data.

[0085] Among them, the target neural network model is obtained through N iterative training, and one iterative training obtains a neural network model, that is, the target neural network model includes N neural network models, the first neural network model is a model obtained through one iterative training, and N is an integer greater than 1.

[0086] The electronic device inputs the first seismic data into the first neural network model in the target neural network model, and obtains the first prediction data through the first neural network model.

[0087] Among them, each neural network model in the target neural network model can be a convolutional neural network model U-Net and its variants, ResNet (residual network), a convolutional neural network using an attention mechanism, or other neural network models, without specific limitation. The training process of the target neural network model will be introduced in detail later and will not be described here.

[0088] Step 203: The electronic device determines second seismic data based on the first seismic data and the first prediction data.

[0089] The second seismic data is data after a portion of aliasing noise is separated.

[0090] This step can be achieved by following the steps (1) to (2), including:

[0091] (1) The electronic device performs aliasing and pseudo separation on the first prediction data to determine the first noise data.

[0092] The electronic device extracts the excitation time corresponding to the seismic data of each seismic channel from the first seismic data, and superimposes the seismic data corresponding to the multiple seismic channels on the basis of the first prediction data according to the excitation time corresponding to each seismic channel to obtain the fourth seismic data. The first prediction data is subtracted from the fourth seismic data to obtain the first noise data.

[0093] The electronic device sequentially superimposes the seismic data of the first seismic channel, the seismic data of the second seismic channel, the seismic data of the third seismic channel, on the basis of the first prediction data according to the excitation time corresponding to each seismic channel in the first seismic data, until the seismic data of the last seismic channel is superimposed to obtain the fourth seismic data. Then, the first prediction data is subtracted from the fourth seismic data to obtain the first noise data.

[0094] It should be noted that, according to the excitation time corresponding to each seismic channel, the first predicted data is superimposed with the seismic data corresponding to multiple seismic channels to form new data. Since the effective signal of the first predicted data remains unchanged after superposition, only aliasing noise is generated, the aliasing noise of the first iteration can be separated by subtracting the first predicted data from the new data. Figure 4 , Figure 4 Schematic diagram of the aliasing noise obtained after the first iterative prediction.

[0095] Among them, when the aliasing noise level in the predicted data changes, the noise data obtained by this step also changes accordingly. Therefore, each iteration will separate the neighboring shot aliasing noise of different intensities, that is, the neighboring shot aliasing noise is characterized in multiple levels. After N iterations, it means that a set of multi-level aliasing noise sets composed of N neighboring shot aliasing noises of different levels is generated. The neighboring shot multi-level aliasing noise is used to characterize the change of the aliasing noise level with the number of iterations during the iterative aliasing separation process. The introduction of neighboring shot multi-level aliasing noise can improve the accuracy of iterative aliasing separation, especially in the case of low signal-to-noise ratio and weak signal development, and can significantly improve the quality of aliasing data separation.

[0096] (2) The electronic device separates the first noise data from the first seismic data to obtain the second seismic data.

[0097] The electronic device subtracts the first noise data from the first seismic data to obtain the second seismic data with reduced aliasing noise, and uses the second seismic data as input data for the next iterative prediction.

[0098] Another point that needs to be explained is that, ideally, since only the aliasing noise is subtracted, both the effective signal and the weak signal are retained in the second seismic data, and even complex signals and weak signals will basically not cause effective signal damage. However, the noise of the actual seismic data itself and the aliasing noise of the adjacent shots are superimposed on each other, resulting in an insufficient learning process of the neural network. Therefore, there is still aliasing noise of the adjacent shots in the second seismic data, and it is necessary to continue iterative prediction to improve the separation effect of the aliasing noise of the adjacent shots.

[0099] Step 204: The electronic device inputs the second seismic data into the second neural network model in the target neural network model to obtain second prediction data.

[0100] The second neural network model is a model obtained through two iterative trainings.

[0101] The electronic device inputs the second earthquake data into the second neural network model and obtains the second prediction data through the second neural network model.

[0102] Step 205: The electronic device determines the third seismic data based on the first seismic data and the second prediction data. By analogy, the step of inputting the seismic data into the corresponding neural network model in the target neural network model is repeated until the target prediction data is obtained through the Nth neural network model in the target neural network model.

[0103] The electronic device performs aliasing and pseudo-separation on the second predicted data, determines second noise data, separates the second noise data from the first seismic data, and obtains third seismic data, which is data after further separation of aliasing noise.

[0104] Among them, the electronic device performs aliasing and pseudo-separation on the second predicted data, and the process of determining the second noise data is also based on the excitation time corresponding to each seismic channel, and the seismic data corresponding to multiple seismic channels are superimposed on the second predicted data to obtain the fifth seismic data, and the second predicted data is subtracted from the fifth seismic data to obtain the second noise data. This step is similar to step (1) in step 203 and will not be repeated here. The process of the electronic device separating the second noise data from the first seismic data is similar to step (2) in step 203 and will not be repeated here.

[0105] The electronic device inputs the third earthquake data into the third neural network model in the target neural network model, and obtains third prediction data through the third neural network model. The third neural network model is obtained through three iterations of training.

[0106] The electronic device performs aliasing and pseudo-separation on the third prediction data, determines the third noise data, separates the third noise data from the first seismic data, obtains the sixth seismic data, inputs the sixth seismic data into the next neural network model in the target neural network model, and obtains the fourth prediction data.

[0107] By analogy, the step of inputting seismic data into the corresponding neural network model in the target neural network model is repeated until the target prediction data is obtained through the Nth neural network model in the target neural network model, and the target prediction data is the data after removing the aliasing noise. Figure 5 , Figure 5 Schematic diagram of target prediction data after removing aliasing noise.

[0108] See also Figure 6 , Figure 6 Schematic diagram of removing aliasing noise through the target neural network model. The electronic device inputs the first seismic data into the first neural network model, performs the first iterative prediction, and obtains the first prediction data. According to the first prediction data and the first seismic data, the first noise data is determined, and then the data for the next iterative prediction input is constructed. It is determined whether the number of current iterative predictions is less than N. If it is less than N, the next iterative prediction is performed. If it is equal to N, the data obtained by the iterative prediction is output as the result.

[0109] This application combines the deep learning neural network's powerful ability to extract nonlinear features, the ability to characterize multi-level aliasing noise of different strengths at multiple levels, and the ability to adaptively separate strong and weak noises through iterative denoising. Since each iterative prediction only subtracts the aliasing noise of adjacent shots without damaging the effective signal, it can greatly improve the fidelity of the effective signal, especially for more effective protection of weak signals in seismic data, and provide high-quality aliasing-free data for the subsequent data processing process. This is mainly reflected in the following aspects:

[0110] (1) Compared with the methods of related technologies, the present application utilizes the powerful ability of neural networks to extract nonlinear features, which can effectively extract hyperbolic event axis features such as seismic reflection and scattering, overcome limitations such as local linear approximation assumptions, and thus achieve better separation effects.

[0111] (2) By using iterative training and iterative prediction, the aliasing noise is adaptively estimated and the corresponding neural network model is trained in a targeted manner. When removing the aliasing noise in the actual data, the neural network model corresponding to the current iteration is used for denoising at different iteration stages. This can better solve the problem of incomplete separation of aliasing noise and damage to effective signals, and maximize the retention of weak signal information in complex seismic data.

[0112] After the electronic device obtains the target prediction data, it can further process the target prediction data, such as removing random noise, removing multiple waves, etc., perform migration imaging based on the further processed data, and perform reservoir prediction based on the migration imaging results.

[0113] The embodiment of the present application provides a method for processing seismic data, which first predicts the first seismic data through the first neural network model in the target neural network model to obtain the first predicted data, and then separates part of the aliasing noise from the first seismic data based on the first seismic data and the first predicted data to obtain the second seismic data. The second seismic data is predicted through the second neural network model in the target neural network model to obtain the second predicted data, and then further separates the aliasing noise from the first seismic data based on the first seismic data and the second predicted data to obtain the third seismic data. And so on, until the data after the aliasing noise is removed is obtained through the Nth neural network model in the target neural network model. It can be seen that this method utilizes the powerful extraction ability of the neural network model for nonlinear features, gradually separates the aliasing noise through iterative prediction, and finally obtains the data after the aliasing noise is removed, which overcomes the limitations of many assumptions and conditions, thereby improving the separation effect of the aliasing noise.

[0114] The following describes the training process of the target neural network model. The target neural network model can be trained on an electronic device or on other devices and then deployed to the electronic device, without specific limitation. In the embodiments of the present application, only the target neural network model trained on an electronic device is used as an example for explanation.

[0115] Figure 7 is a flowchart of a training target neural network model provided by an embodiment of the present application, which is executed by an electronic device, see Figure 7 , the method comprising:

[0116] Step 701: The electronic device obtains first sample seismic data and first sample label data.

[0117] The process of the electronic device acquiring the first sample seismic data is the same as the process of acquiring the first seismic data, which is to extract multiple common detection point gather data from the shot gather data collected by the target work area in an overlapping manner to obtain the first sample seismic data.

[0118] For the first sample label data, the electronic device may use a method in the related art to remove the aliasing noise in the first sample seismic data, for example, using a sparse inversion method to remove the aliasing noise in the first sample seismic data to obtain the first sample label data. Of course, the electronic device may also use other methods such as filtering and denoising to remove the aliasing noise in the first sample seismic data to obtain the first sample label data, which is not specifically limited. The number of the first sample seismic data and the first sample label data is multiple.

[0119] For example, the electronic device extracts 300 common detection point gather data from the aliased shot gather data, each common detection point gather data has 201 seismic data, and randomly extracts 80 common detection point gather data from the 300 common detection point gather data, and randomly selects 80% of the data from the 80 common detection point gather data, that is, 64 common detection point gather data as the first sample seismic data. For these 64 common detection point gather data, the aliasing noise is removed by the sparse inversion method, and 64 sample label data, that is, the first sample label data, are obtained.

[0120] Step 702: The electronic device performs a first iterative training based on the first sample seismic data and the first sample label data to obtain a first neural network model.

[0121] The electronic device inputs the first sample seismic data and the first sample label data into the initial neural network model to perform the first iterative training. The model parameters of the initial neural network model can be initialized with random values, epochs are set, and then iterative training is performed.

[0122] In the embodiment of the present application, the neural network model can be represented by R, the model parameter is represented by θ, and the process of each iterative training can be represented by the following formula (1):

[0123]

[0124] Where j = 1, 2...N, N is the maximum number of iterations, and the first sample seismic data is expressed as After the first iteration, the first neural network model can be expressed as R1, its parameter is expressed as θ1, and the first output data is recorded as

[0125] The epoch may be set and changed as needed, for example, epoch=200 is set, and there is no specific limitation on this.

[0126] The electronic device obtains the first output data through the first neural network model, determines the difference between the first output data and the corresponding first sample label data, and then back-propagates the gradient and updates the neural network parameters, and uses the neural network parameters corresponding to the last epoch as the initial model parameters for the next iterative training.

[0127] Step 703: The electronic device determines second sample seismic data based on the first sample seismic data and the first output data.

[0128] The electronic device performs aliasing and pseudo separation on the first output data, determines first sample noise data, separates the first sample noise data from the first sample seismic data, and obtains second sample seismic data.

[0129] The process of the electronic device determining the first sample noise data may be as follows: for each common detection point gather data, the electronic device extracts the excitation time corresponding to the seismic data of each seismic channel from the common detection point gather data, and according to the excitation time corresponding to each seismic channel, superimposes the seismic data corresponding to multiple seismic channels on the basis of the first output data to obtain the third sample seismic data, and subtracts the first output data from the third sample seismic data to obtain the first sample noise data. This process is the same as step (1) in step 203, and will not be repeated here.

[0130] For example, each common detection point gather data has 201 seismic data. For 64 common detection point gather data, there are 64*201=12864 excitation times. The electronic device extracts 12864 excitation times from the 64 common detection point gather data.

[0131] Step 704: The electronic device performs a second iterative training based on the second sample seismic data and the first sample label data to obtain a second neural network model; and so on, repeating the steps of iterative training of the model until the Nth iterative training is performed to obtain the Nth neural network model.

[0132] The electronic device performs a second iterative training on the basis of the first neural network model based on the second sample seismic data and the first sample label data to obtain a second neural network model.

[0133] The electronic device obtains the second output data through the second neural network model, determines the difference between the second output data and the first sample label data corresponding to it, then back-propagates the gradient and updates the neural network parameters, and uses the neural network parameters corresponding to the last epoch as the initial model parameters for the next iterative training. The electronic device performs aliasing and pseudo-separation on the second output data, determines the second sample noise data, separates the second sample noise data from the first sample seismic data, and obtains the third sample seismic data. Based on the third sample seismic data and the first sample label data, a third iterative training is performed on the basis of the second neural network model to obtain a third neural network model.

[0134] The electronic device obtains the third output data through the third neural network model, determines the difference between the third output data and the first sample label data corresponding to it, then back-propagates the gradient and updates the neural network parameters, and uses the neural network parameters corresponding to the last epoch as the initial model parameters for the next iterative training. The electronic device performs aliasing and pseudo-separation on the third output data, determines the third sample noise data, separates the third sample noise data from the first sample seismic data, and obtains the fourth sample seismic data. Based on the third sample seismic data and the first sample label data, the next iterative training is performed. And so on, until the Nth iterative training is performed to obtain the Nth neural network model. The target neural network model includes the first neural network model, the second neural network model, the third neural network model, and the Nth neural network model.

[0135] See also Figure 8 The electronic device performs iterative training based on the first sample seismic data and the first sample label data. After each iterative training, it is determined whether the current number of iterations reaches the maximum number of iterations, that is, N. If N is reached, the iterative training is stopped, and the neural network models obtained by each iterative training are combined into a target neural network model. If N is not reached, the next iterative training is performed until N is reached, thereby obtaining the target neural network model.

[0136] Among them, N can be set and changed as needed, for example, if N=3, then three iterations of training are performed, and after three iterations of training, three neural network models trained for adjacent shot aliasing noise of different scales are obtained. In addition, for the first iteration, the model parameters can be initialized with random values, and for the second and subsequent iterations, the model parameters can be trained based on the model parameters obtained in the previous iteration training.

[0137] In the embodiment of the present application, after the electronic device obtains the target neural network model, it can immediately verify the target neural network model. After verification, the target neural network model is used to predict the actual data to remove aliasing noise. Alternatively, the electronic device can also verify the target neural network model when predicting the actual data, which is not specifically limited.

[0138] The process of the electronic device verifying the target neural network model may be: the electronic device obtains first verification seismic data and first verification label data, and based on the first verification seismic data and the first verification label data, verifies the N neural network models included in the target neural network model; if the verification is passed, execute the above step 202 again.

[0139] In this implementation, the electronic device inputs the first verification seismic data into the first neural network model in the target neural network model to obtain the first verification data. Based on the first verification seismic data and the first verification data, the second verification seismic data is determined, and the second verification seismic data is input into the second neural network model in the target neural network model to obtain the second verification data. Based on the first verification seismic data and the second verification data, the third verification seismic data is determined, and the third verification seismic data is input into the third neural network model in the target neural network model to obtain the third verification data; and so on, the step of inputting the verification seismic data into the corresponding neural network model in the target neural network model is repeated until the target verification data is obtained through the Nth neural network model in the target neural network model. Determine the difference between the target verification data and the first verification label data, and if the difference is within the preset threshold range, determine that the verification is passed. If the difference is not within the preset threshold range, determine that the verification is not passed.

[0140] If the verification fails, the electronic device can continue to perform iterative training, that is, increase the maximum number of iterations, until the verification passes.

[0141] The process of the electronic device determining the second verification seismic data based on the first verification seismic data and the first verification data is the same as step 203 and will not be repeated here.

[0142] For example, the electronic device uses the remaining 20% ​​of the 80 common detection point gather data, that is, 16 common detection point gather data, as the first verification seismic data. For these 16 common detection point gather data, the sparse inversion method is used to remove aliasing noise to obtain 16 verification label data, that is, the first verification label data.

[0143] The embodiment of the present application provides a method for training a target neural network model, which utilizes the powerful extraction capability of the neural network model for nonlinear features, adaptively estimates the noise through iterative training, and trains the corresponding neural network model in a targeted manner. In the iterative training stage, the neural network model corresponding to the current iterative stage is trained in view of the fact that different iterative stages contain different levels of adjacent shot aliasing noise. In this way, when the trained neural network model removes aliasing noise, the problem of incomplete separation of aliasing noise and damage to effective signals can be better solved, thereby retaining weak signal information in complex seismic data to the maximum extent.

[0144] Fig. 9 is a structural diagram of a seismic data processing device provided in an embodiment of the present application, see Fig. 9 , the device comprises:

[0145] A first acquisition module 901 is used to acquire first seismic data, where the first seismic data is common detection point gather data of a target work area;

[0146] A first input module 902 is used to input the first seismic data into a first neural network model in a target neural network model to obtain first prediction data; wherein the target neural network model is obtained through N iterative training, one iterative training obtains one neural network model, and N is an integer greater than 1;

[0147] A first determination module 903 is used to determine second seismic data based on the first seismic data and the first prediction data, where the second seismic data is data after a portion of aliasing noise is separated;

[0148] A second input module 904, used to input the second earthquake data into the second neural network model in the target neural network model to obtain second prediction data;

[0149] The second determination module 905 is used to determine the third seismic data based on the first seismic data and the second prediction data, and the third seismic data is the data after further separation of aliasing noise; and so on, repeatedly execute the step of inputting the seismic data into the corresponding neural network model in the target neural network model until the target prediction data is obtained through the Nth neural network model in the target neural network model, and the target prediction data is the data after removing aliasing noise.

[0150] In a possible implementation, the first determination module 903 is used to perform aliasing and pseudo-separation on the first prediction data to determine the first noise data; and separate the first noise data from the first seismic data to obtain the second seismic data.

[0151] In another possible implementation, the first seismic data includes seismic data corresponding to a plurality of seismic traces;

[0152] The first determination module 903 is used to extract the excitation time corresponding to the seismic data of each seismic channel from the first seismic data; according to the excitation time corresponding to each seismic channel, superimpose the seismic data corresponding to multiple seismic channels on the basis of the first prediction data to obtain the fourth seismic data; subtract the first prediction data from the fourth seismic data to obtain the first noise data.

[0153] In another possible implementation, the device further includes:

[0154] A second acquisition module is used to acquire first sample seismic data and first sample label data; wherein the first sample seismic data is common detection point gather data, and the first sample label data is data after aliasing noise is removed;

[0155] A first training module, used for performing a first iterative training based on first sample seismic data and first sample label data to obtain a first neural network model;

[0156] A third determination module is used to determine second sample seismic data based on the first sample seismic data and the first output data, where the first output data is output data of the first neural network model;

[0157] The second training module is used to perform a second iterative training based on the second sample seismic data and the first sample label data to obtain a second neural network model; and so on, the steps of iterative training of the model are repeated until the Nth iterative training is performed to obtain the Nth neural network model.

[0158] In another possible implementation, the second training module is used to perform a second iterative training on the basis of the first neural network model based on the second sample seismic data and the first sample label data to obtain a second neural network model.

[0159] In another possible implementation, the second acquisition module is used to acquire shot gather data collected by aliasing in the target work area; extract common detection point gather data from the shot gather data to obtain first sample seismic data; and use a sparse inversion method to remove aliasing noise in the first sample seismic data to obtain first sample label data.

[0160] In another possible implementation, the device further includes:

[0161] A third acquisition module is used to acquire first verification seismic data and first verification label data; wherein the first verification seismic data is common detection point gather data, and the first verification label data is data after aliasing noise is removed;

[0162] A verification module, used to verify N neural network models included in the target neural network model based on the first verification seismic data and the first verification label data;

[0163] The first input module 902 is used to input the first seismic data into the first neural network model in the target neural network model when the verification is passed.

[0164] The embodiment of the present application provides a seismic data processing device, which first predicts the first seismic data through the first neural network model in the target neural network model to obtain the first predicted data, and separates part of the aliasing noise from the first seismic data based on the first seismic data and the first predicted data to obtain the second seismic data. The second seismic data is predicted by the second neural network model in the target neural network model to obtain the second predicted data, and the aliasing noise is further separated from the first seismic data based on the first seismic data and the second predicted data to obtain the third seismic data. And so on, until the data after the aliasing noise is removed is obtained through the Nth neural network model in the target neural network model. It can be seen that the device uses the powerful extraction ability of the neural network model for nonlinear features, gradually separates the aliasing noise through iterative prediction, and finally obtains the data after the aliasing noise is removed, which overcomes the limitations of many assumptions and improves the separation effect of the aliasing noise.

[0165] refer to Fig.10 , Fig.10 The structure block diagram of a terminal 1000 provided by an exemplary embodiment of the present application is shown. The terminal 1000 may be a portable mobile terminal, such as a smart phone, a tablet computer, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 player (Moving Picture Experts Group Audio Layer IV), a laptop computer or a desktop computer. The terminal 1000 may also be called a user device, a portable terminal, a laptop terminal, a desktop terminal or other names.

[0166] Typically, the terminal 1000 includes a processor 1001 and a memory 1002 .

[0167] The processor 1001 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 1001 may be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor 1001 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 1001 may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 1001 may also include an AI (Artificial Intelligence) processor, which is used to process computing operations related to machine learning.

[0168] The memory 1002 may include one or more computer-readable storage media, which may be non-transitory. The memory 1002 may also include a high-speed random access memory, and a non-volatile memory, such as one or more disk storage devices, flash memory storage devices. In some embodiments, the non-transitory computer-readable storage medium in the memory 1002 is used to store at least one program code, which is used to be executed by the processor 1001 to implement the operation performed by the terminal in the seismic data processing method provided in the method embodiment of the present application.

[0169] In some embodiments, the terminal 1000 may further optionally include: a peripheral device interface 1003 and at least one peripheral device. The processor 1001, the memory 1002 and the peripheral device interface 1003 may be connected via a bus or a signal line. Each peripheral device may be connected to the peripheral device interface 1003 via a bus, a signal line or a circuit board. Specifically, the peripheral device includes: at least one of a radio frequency circuit 1004, a display screen 1005, a camera assembly 1006, an audio circuit 1007 and a power supply 1008.

[0170] The peripheral device interface 1003 may be used to connect at least one peripheral device related to I / O (Input / Output) to the processor 1001 and the memory 1002. In some embodiments, the processor 1001, the memory 1002, and the peripheral device interface 1003 are integrated on the same chip or circuit board; in some other embodiments, any one or two of the processor 1001, the memory 1002, and the peripheral device interface 1003 may be implemented on a separate chip or circuit board, which is not limited in this embodiment.

[0171] The radio frequency circuit 1004 is used to receive and transmit RF (Radio Frequency) signals, also known as electromagnetic signals. The radio frequency circuit 1004 communicates with the communication network and other communication devices through electromagnetic signals. The radio frequency circuit 1004 converts the electrical signal into an electromagnetic signal for transmission, or converts the received electromagnetic signal into an electrical signal. Optionally, the radio frequency circuit 1004 includes: an antenna system, an RF transceiver, one or more amplifiers, a tuner, an oscillator, a digital signal processor, a codec chipset, a user identity module card, etc. The radio frequency circuit 1004 can communicate with other terminals through at least one wireless communication protocol. The wireless communication protocol includes but is not limited to: the World Wide Web, a metropolitan area network, an intranet, various generations of mobile communication networks (2G, 3G, 4G and 5G), a wireless local area network and / or a WiFi (Wireless Fidelity) network. In some embodiments, the radio frequency circuit 1004 may also include circuits related to NFC (Near Field Communication), which is not limited in this application.

[0172] The display screen 1005 is used to display a UI (User Interface). The UI may include graphics, text, icons, videos, and any combination thereof. When the display screen 1005 is a touch display screen, the display screen 1005 also has the ability to collect touch signals on the surface or above the surface of the display screen 1005. The touch signal can be input to the processor 1001 as a control signal for processing. At this time, the display screen 1005 can also be used to provide virtual buttons and / or virtual keyboards, also known as soft buttons and / or soft keyboards. In some embodiments, the display screen 1005 can be one, arranged on the front panel of the terminal 1000; in other embodiments, the display screen 1005 can be at least two, respectively arranged on different surfaces of the terminal 1000 or in a folding design; in other embodiments, the display screen 1005 can be a flexible display screen, arranged on a curved surface or a folding surface of the terminal 1000. Even, the display screen 1005 can also be arranged as a non-rectangular irregular figure, that is, a special-shaped screen. The display screen 1005 can be made of materials such as LCD (Liquid Crystal Display) and OLED (Organic Light-Emitting Diode).

[0173] The camera assembly 1006 is used to capture images or videos. Optionally, the camera assembly 1006 includes a front camera and a rear camera. Typically, the front camera is arranged on the front panel of the terminal, and the rear camera is arranged on the back of the terminal. In some embodiments, there are at least two rear cameras, which are any one of a main camera, a depth of field camera, a wide-angle camera, and a telephoto camera, so as to realize the fusion of the main camera and the depth of field camera to realize the background blur function, the fusion of the main camera and the wide-angle camera to realize the panoramic shooting and VR (Virtual Reality) shooting function or other fusion shooting functions. In some embodiments, the camera assembly 1006 may also include a flash. The flash can be a monochrome temperature flash or a dual-color temperature flash. A dual-color temperature flash refers to a combination of a warm light flash and a cold light flash, which can be used for light compensation at different color temperatures.

[0174] The audio circuit 1007 may include a microphone and a speaker. The microphone is used to collect sound waves from the user and the environment, and convert the sound waves into electrical signals and input them into the processor 1001 for processing, or input them into the radio frequency circuit 1004 to achieve voice communication. For the purpose of stereo acquisition or noise reduction, there may be multiple microphones, which are respectively arranged at different parts of the terminal 1000. The microphone may also be an array microphone or an omnidirectional acquisition microphone. The speaker is used to convert the electrical signal from the processor 1001 or the radio frequency circuit 1004 into sound waves. The speaker may be a traditional film speaker or a piezoelectric ceramic speaker. When the speaker is a piezoelectric ceramic speaker, it can not only convert the electrical signal into sound waves audible to humans, but also convert the electrical signal into sound waves inaudible to humans for purposes such as ranging. In some embodiments, the audio circuit 1007 may also include a headphone jack.

[0175] The power supply 1008 is used to power various components in the terminal 1000. The power supply 1008 can be an alternating current, a direct current, a disposable battery, or a rechargeable battery. When the power supply 1008 includes a rechargeable battery, the rechargeable battery can be a wired rechargeable battery or a wireless rechargeable battery. A wired rechargeable battery is a battery charged through a wired line, and a wireless rechargeable battery is a battery charged through a wireless coil. The rechargeable battery can also be used to support fast charging technology.

[0176] In some embodiments, the terminal 1000 further includes one or more sensors 1009 . The one or more sensors 1009 include, but are not limited to: an acceleration sensor 1010 , a gyroscope sensor 1011 , a pressure sensor 1012 , an optical sensor 1013 , and a proximity sensor 1014 .

[0177] The acceleration sensor 1010 can detect the magnitude of acceleration on the three coordinate axes of the coordinate system established by the terminal 1000. For example, the acceleration sensor 1010 can be used to detect the components of gravity acceleration on the three coordinate axes. The processor 1001 can control the display screen 1005 to display the user interface in a horizontal view or a vertical view based on the gravity acceleration signal collected by the acceleration sensor 1010. The acceleration sensor 1010 can also be used for collecting game or user motion data.

[0178] The gyro sensor 1011 can detect the body direction and rotation angle of the terminal 1000, and the gyro sensor 1011 can cooperate with the acceleration sensor 1010 to collect the user's 3D actions on the terminal 1000. Based on the data collected by the gyro sensor 1011, the processor 1001 can implement the following functions: motion sensing (such as changing the UI based on the user's tilt operation), image stabilization during shooting, game control, and inertial navigation.

[0179] The pressure sensor 1012 can be set on the side frame of the terminal 1000 and / or the lower layer of the display screen 1005. When the pressure sensor 1012 is set on the side frame of the terminal 1000, it can detect the user's holding signal of the terminal 1000, and the processor 1001 performs left and right hand recognition or shortcut operation based on the holding signal collected by the pressure sensor 1012. When the pressure sensor 1012 is set on the lower layer of the display screen 1005, the processor 1001 controls the operability controls on the UI interface based on the user's pressure operation on the display screen 1005. The operability controls include at least one of a button control, a scroll bar control, an icon control, and a menu control.

[0180] The optical sensor 1013 is used to collect the ambient light intensity. In one embodiment, the processor 1001 can control the display brightness of the display screen 1005 based on the ambient light intensity collected by the optical sensor 1013. Specifically, when the ambient light intensity is high, the display brightness of the display screen 1005 is increased; when the ambient light intensity is low, the display brightness of the display screen 1005 is decreased. In another embodiment, the processor 1001 can also dynamically adjust the shooting parameters of the camera assembly 1006 based on the ambient light intensity collected by the optical sensor 1013.

[0181] The proximity sensor 1014, also called a distance sensor, is usually arranged on the front panel of the terminal 1000. The proximity sensor 1014 is used to collect the distance between the user and the front of the terminal 1000. In one embodiment, when the proximity sensor 1014 detects that the distance between the user and the front of the terminal 1000 is gradually decreasing, the processor 1001 controls the display screen 1005 to switch from the screen-on state to the screen-off state; when the proximity sensor 1014 detects that the distance between the user and the front of the terminal 1000 is gradually increasing, the processor 1001 controls the display screen 1005 to switch from the screen-off state to the screen-on state.

[0182] Those skilled in the art will understand that Fig.10 The structure shown in the figure does not constitute a limitation on the terminal 1000, and the terminal 1000 may include more or less components than those shown in the figure, or combine some components, or adopt a different component arrangement.

[0183] The server structure diagram can be found in Fig.11The server 1100 may have relatively large differences due to different configurations or performances, and may include a processor (Central Processing Units, CPU) 1101 and a memory 1102, wherein the memory 1102 stores at least one program code, which is loaded and executed by the processor 1101 to implement the operations performed by the server in the above-mentioned seismic data processing method. Of course, the server 1100 may also have components such as a wired or wireless network interface, a keyboard, and an input and output interface for input and output. The server 1100 may also include other components for implementing device functions, which will not be described in detail here.

[0184] In an exemplary embodiment, a computer-readable storage medium is also provided, which stores at least one program code, and the at least one program code is loaded and executed by a processor to implement the seismic data processing method in the above embodiment.

[0185] In an exemplary embodiment, a computer program product is also provided. The computer program product stores at least one program code, and the at least one program code is loaded and executed by a processor to implement the seismic data processing method in the above embodiment.

[0186] A person skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware or by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk or an optical disk, etc.

[0187] The above description is only for the purpose of facilitating those skilled in the art to understand the technical solution of the present application and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for processing seismic data, characterized in that: The method comprises: Acquire first seismic data, where the first seismic data is common detection point gather data of a target work area; Inputting the first seismic data into a first neural network model in a target neural network model to obtain first prediction data; wherein the target neural network model is obtained through N iterative training, one iterative training obtains one neural network model, and N is an integer greater than 1; Determine second seismic data based on the first seismic data and the first predicted data, where the second seismic data is data after a portion of aliasing noise is separated; Inputting the second seismic data into a second neural network model in the target neural network model to obtain second prediction data; Determine third seismic data based on the first seismic data and the second predicted data, wherein the third seismic data is data after aliasing noise is further separated; By analogy, the step of inputting the seismic data into the corresponding neural network model in the target neural network model is repeated until the target prediction data is obtained through the Nth neural network model in the target neural network model, and the target prediction data is the data after removing aliasing noise.

2. The method according to claim 1, characterized in that The determining of second seismic data based on the first seismic data and the first prediction data comprises: Performing aliasing and pseudo separation on the first prediction data to determine first noise data; The first noise data is separated from the first seismic data to obtain the second seismic data.

3. The method according to claim 2, characterized in that The first seismic data includes seismic data corresponding to a plurality of seismic traces; The performing aliasing and pseudo separation on the first prediction data to determine first noise data includes: Extracting the excitation time corresponding to the seismic data of each seismic trace from the first seismic data; According to the excitation time corresponding to each seismic trace, the seismic data corresponding to the plurality of seismic traces are superimposed on the first prediction data to obtain fourth seismic data; The first prediction data is subtracted from the fourth seismic data to obtain the first noise data.

4. The method according to claim 1, characterized in that: The training process of the target neural network model includes: Acquire first sample seismic data and first sample label data; wherein the first sample seismic data is common detection point gather data, and the first sample label data is data after aliasing noise is removed; Performing a first iterative training based on the first sample seismic data and the first sample label data to obtain the first neural network model; Determining second sample seismic data based on the first sample seismic data and first output data, wherein the first output data is output data of the first neural network model; Based on the second sample seismic data and the first sample label data, performing a second iterative training to obtain the second neural network model; By analogy, the steps of iterative model training are repeated until the Nth iterative training is performed to obtain the Nth neural network model.

5. The method according to claim 4, characterized in that The second iterative training is performed based on the second sample seismic data and the first sample label data to obtain the second neural network model, including: Based on the second sample seismic data and the first sample label data, a second iterative training is performed on the basis of the first neural network model to obtain the second neural network model.

6. The method according to claim 4, characterized in that The obtaining of the first sample seismic data and the first sample label data comprises: Acquire shot gather data collected by aliasing in the target work area; Extracting common detection point gather data from the shot gather data to obtain the first sample seismic data; A sparse inversion method is used to remove aliasing noise in the first sample seismic data to obtain the first sample label data.

7. The method according to claim 1, characterized in that The method further comprises: Acquire first verification seismic data and first verification label data; wherein the first verification seismic data is common detection point gather data, and the first verification label data is data after aliasing noise is removed; Based on the first verification seismic data and the first verification label data, verifying the N neural network models included in the target neural network model; When the verification is passed, the step of inputting the first seismic data into the first neural network model in the target neural network model is executed.

8. A seismic data processing device, characterized in that: The device comprises: A first acquisition module is used to acquire first seismic data, where the first seismic data is common detection point gather data of a target work area; A first input module is used to input the first seismic data into a first neural network model in a target neural network model to obtain first prediction data; wherein the target neural network model is obtained through N iterative training, one iterative training obtains one neural network model, and N is an integer greater than 1; A first determination module is used to determine second seismic data based on the first seismic data and the first prediction data, where the second seismic data is data after a portion of aliasing noise is separated; A second input module, used for inputting the second seismic data into a second neural network model in the target neural network model to obtain second prediction data; The second determination module is used to determine the third seismic data based on the first seismic data and the second prediction data, and the third seismic data is the data after further separation of aliasing noise; and so on, repeatedly executing the step of inputting the seismic data into the corresponding neural network model in the target neural network model until the target prediction data is obtained through the Nth neural network model in the target neural network model, and the target prediction data is the data after removing aliasing noise.

9. An electronic device, characterized in that: The electronic device includes a processor and a memory, wherein the memory stores at least one program code, and the at least one program code is loaded and executed by the processor to implement the seismic data processing method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores at least one program code, and the at least one program code is loaded and executed by a processor to implement the seismic data processing method according to any one of claims 1 to 7.

Images