A method and apparatus for deconvolution of seismic data

Abstract

The application provides a deconvolution method and device for seismic data, and the method comprises the following steps: obtaining to-be-measured seismic data and historical seismic data; establishing a seismic sample database based on the historical seismic data; obtaining seismic data training samples from the seismic sample database; the seismic data samples comprise common shot data and corresponding sample category labels; a target consistency deconvolution network is constructed based on the common shot data and the corresponding sample category labels; the to-be-measured seismic data is input into the consistency deconvolution network to obtain a corresponding predicted category. An intelligent surface consistency deconvolution deep neural network with a self-attention mechanism is established, and a surface-oriented deconvolution operator is extracted based on the deep neural network, so that the processing of pre-stack seismic consistency deconvolution is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of geophysical exploration, and in particular to a seismic data deconvolution method and device. BACKGROUND

[0002] In the process of seismic data processing, deconvolution processing is a particularly important link, and plays a role in the whole process. Deconvolution mainly compresses the basic seismic wavelet in the seismic record, suppresses mixed reverberation and short-period multiple waves, thereby improving the vertical resolution.

[0003] Although at present, the deconvolution technology has made certain development at home and abroad, and in recent years, with the continuous development of artificial intelligence, the seismic data processing technology based on deep learning has also achieved certain application effect, but the method based on deep neural network to realize the consistency processing of seismic data to better serve the seismic data preprocessing consistency deconvolution process has not been proposed. SUMMARY

[0004] The present application provides a seismic data deconvolution method and device, which realizes the extraction of surface-oriented deconvolution operators based on deep neural networks, improves the network precision, and realizes the processing of pre-stack seismic data consistency deconvolution.

[0005] In a first aspect, the present application provides a seismic data deconvolution method, comprising:

[0006] Obtaining the measured seismic data and historical seismic data;

[0007] Establishing a seismic sample database based on the historical seismic data;

[0008] Obtaining seismic data training samples from the seismic sample database; the seismic data samples include common shot gather data and corresponding sample class labels;

[0009] Based on the common shot gather data and the corresponding sample class labels, a target consistency deconvolution network is constructed;

[0010] The measured seismic data is input into the consistency deconvolution network to obtain the corresponding predicted class.

[0011] Optionally, the sample class label includes common receiver domain data and common offset domain data; establishing a seismic sample database based on the historical seismic data comprises:

[0012] The historical seismic data is sorted to obtain two types of data including common receiver domain data and common offset domain data;

[0013] According to the data type of the historical seismic data, the corresponding data class label is labeled;

[0014] According to the historical seismic data and the corresponding data category label, a seismic sample database is constructed.

[0015] Optionally, based on the common shot data and the corresponding sample category label, a target consistency deconvolution network is constructed, comprising:

[0016] All common shots of the common shot data are input into the consistency deconvolution network to generate a corresponding sample category.

[0017] According to the sample category label and the sample category, a training error is determined, and based on the training error, the consistency deconvolution network is adjusted to obtain optimal network parameters, and the target consistency deconvolution network is generated using the optimal network parameters.

[0018] Optionally, before all common shots of the common shot data are input into the consistency deconvolution network to generate a corresponding sample category, the method further comprises:

[0019] The common shot data is blocked, cropped or resampled according to a preset size.

[0020] In a second aspect, the present application provides a deconvolution device for seismic data, comprising:

[0021] A data acquisition module is configured to acquire to-be-measured seismic data and historical seismic data.

[0022] A database establishment module is configured to establish a seismic sample database based on the historical seismic data.

[0023] A sample acquisition module is configured to acquire seismic data training samples from the seismic sample database; the seismic data samples comprise common shot data and corresponding sample category labels.

[0024] A construction module is configured to construct a target consistency deconvolution network based on the common shot data and the corresponding sample category labels.

[0025] A prediction module is configured to input the to-be-measured seismic data into the consistency deconvolution network to obtain a corresponding prediction category.

[0026] Optionally, the sample category label comprises common receiver domain data and common offset domain data; and the database establishment module comprises:

[0027] A sorting sub-module is configured to sort the historical seismic data to obtain two data types including common receiver domain data and common offset domain data.

[0028] A labeling sub-module is configured to label corresponding data category labels according to the data types of the historical seismic data.

[0029] The constructing module is configured to construct the seismic sample database according to the historical seismic data and the corresponding sample category label.

[0030] Optionally, the constructing module comprises:

[0031] The input sub-module is configured to input all common shot data of the common shot data into a consistent deconvolution network to generate a corresponding sample category.

[0032] The network generating sub-module is configured to determine a training error according to the sample category label and the sample category, and adjust the consistent deconvolution network based on the training error to obtain optimal network parameters, and generate a target consistent deconvolution network by using the optimal network parameters.

[0033] Optionally, the constructing module further comprises:

[0034] The preprocessing sub-module is configured to block cut or resample the common shot data according to a preset size.

[0035] In a third aspect, the present application provides an electronic device, comprising a processor and a memory, wherein the memory stores computer readable instructions, and when the computer readable instructions are executed by the processor, the steps in the method provided in the first aspect are executed.

[0036] In a fourth aspect, the present application provides a storage medium, which stores a computer program, and when the computer program is executed by a processor, the steps in the method provided in the first aspect are executed.

[0037] From the above technical solutions, the present application has the following advantages:

[0038] The present application provides a deconvolution method and device for seismic data, which comprises the following steps: obtaining to-be-measured seismic data and historical seismic data; establishing a seismic sample database based on the historical seismic data; obtaining a seismic data training sample from the seismic sample database; the seismic data sample comprises common shot data and a corresponding sample category label; constructing a target consistent deconvolution network based on the common shot data and the corresponding sample category label; inputting the to-be-measured seismic data into the consistent deconvolution network to obtain a corresponding predicted category. An intelligent surface consistent deconvolution deep neural network with a self-attention mechanism is established, and a surface-oriented deconvolution operator is extracted based on the deep neural network, thereby realizing the processing of pre-stack seismic consistent deconvolution. BRIEF DESCRIPTION OF DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required by the embodiments or the prior art description. Obviously, the drawings in the following description only constitute some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained on the basis of these drawings without creative labor.

[0040] Figure 1 A flow chart of a deconvolution method for seismic data according to an embodiment of the present application;

[0041] Figure 2 A flow chart of a deconvolution method for seismic data according to an embodiment of the present application;

[0042] Figure 3 A flow chart of a deconvolution method for seismic data according to an embodiment of the present application;

[0043] Figure 4 A flow chart of a deconvolution method for seismic data according to an embodiment of the present application;

[0044] Figure 5 A network structure diagram of a deconvolution method for seismic data according to an embodiment of the present application;

[0045] Figure 6 A pre-processing data diagram of a deconvolution method for seismic data according to an embodiment of the present application;

[0046] Figure 7 A commercial software processing effect diagram of a deconvolution method for seismic data according to an embodiment of the present application;

[0047] Figure 8 A de-noising effect diagram of a deconvolution method for seismic data according to an embodiment of the present application;

[0048] Figure 9 A structural block diagram of a deconvolution method device for seismic data according to an embodiment of the present application. DETAILED DESCRIPTION

[0049] The embodiments of the present application provide a deconvolution method and device for seismic data, which realizes extraction of a surface-oriented deconvolution operator based on a deep neural network, improves network precision, and realizes processing of pre-stack data seismic consistency deconvolution.

[0050] In order to make the application purposes, features and advantages of the present application more obvious and easy to understand, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the following described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0051] Embodiment one, please refer to Figure 1 , Figure 1 The flow step diagram of the geological target-oriented VTI medium reverse time migration imaging method embodiment one of the present application comprises:

[0052] Step S101, obtaining the to-be-measured seismic data and historical seismic data;

[0053] Step S102, establishing a seismic sample database based on the historical seismic data;

[0054] Step S103, obtaining seismic data training samples from the seismic sample database; the seismic data samples comprise common shot data and corresponding sample category labels;

[0055] Step S104, constructing a target consistency deconvolution network based on the common shot data and the corresponding sample category labels;

[0056] Step S105, inputting the to-be-measured seismic data into the consistency deconvolution network to obtain a corresponding predicted category.

[0057] The present application discloses a seismic data deconvolution method, comprising: obtaining to-be-measured seismic data and historical seismic data; establishing a seismic sample database based on the historical seismic data; obtaining seismic data training samples from the seismic sample database; the seismic data samples comprise common shot data and corresponding sample category labels; constructing a target consistency deconvolution network based on the common shot data and the corresponding sample category labels; inputting the to-be-measured seismic data into the consistency deconvolution network to obtain a corresponding predicted category. An intelligent surface consistency deconvolution deep neural network introducing a self-attention mechanism is established, and the extraction of a surface-oriented deconvolution operator is realized based on the deep neural network, thereby realizing the processing of pre-stack seismic consistency deconvolution.

[0058] Embodiment two, please refer to Figure 2 , Figure 2 The flow step diagram of the seismic data deconvolution method embodiment two of the present application comprises:

[0059] Step S201, obtaining to-be-measured seismic data and historical seismic data;

[0060] Step S202, sorting the historical seismic data to obtain two data types including common receiver domain data and common offset domain data;

[0061] Step S203, labeling the corresponding data class label according to the data type of the historical seismic data;

[0062] Step S204, constructing the seismic sample database according to the historical seismic data and the corresponding data class label;

[0063] Step S205, obtaining seismic data training samples from the seismic sample database; the seismic data samples include common shot data and corresponding sample class labels;

[0064] Step S206, constructing a target consistency deconvolution network based on the common shot data and the corresponding sample class labels;

[0065] Step S207, inputting the to-be-tested seismic data into the consistency deconvolution network to obtain the corresponding predicted class.

[0066] The embodiment of the application discloses a seismic data deconvolution method, which labels two data types including common receiver domain data and common offset domain data obtained by sorting historical seismic data, uses historical seismic data as training samples, and ensures the authenticity and accuracy of the samples.

[0067] Embodiment three, please refer to Figure 3 , Figure 3 is a flow step chart of a seismic data deconvolution method embodiment three of the application, and the method comprises:

[0068] Step S301, obtaining to-be-tested seismic data and historical seismic data;

[0069] Step S302, establishing a seismic sample database based on the historical seismic data;

[0070] Step S303, obtaining seismic data training samples from the seismic sample database; the seismic data samples include common shot data and corresponding sample class labels;

[0071] Step S304, inputting all common shots of the common shot data into a consistency deconvolution network to generate corresponding sample classes;

[0072] Step S305, determining a training error according to the sample class label and the sample class, and adjusting the consistency deconvolution network based on the training error to obtain optimal network parameters, and generating a target consistency deconvolution network by using the optimal network parameters;

[0073] Step S306, inputting the to-be-tested seismic data into the consistency deconvolution network to obtain a corresponding prediction category.

[0074] The embodiment of the present application discloses a seismic data deconvolution method, which comprises the following steps:

[0075] Embodiment four, please refer to Figure 4 , Figure 4 The flow chart of the fourth embodiment of the seismic data deconvolution method of the present application is shown in the figure, and the method comprises the following steps:

[0076] Step S401, obtaining to-be-tested seismic data and historical seismic data;

[0077] Step S402, sorting the historical seismic data to obtain two types of data including common receiver domain data and common offset domain data;

[0078] Step S403, labeling a corresponding data category label according to the data type of the historical seismic data;

[0079] Step S404, constructing a seismic sample database according to the historical seismic data and the corresponding data category label;

[0080] Step S405, obtaining a seismic data training sample from the seismic sample database; the seismic data sample comprises common shot data and a corresponding sample category label;

[0081] In the embodiment of the present application, for the consistency deconvolution processing of historical seismic data, the effect of the consistency processing of the common shot data with large consistency deconvolution demand in the actual work area is evaluated, the historical seismic data is filtered, the data with good processing effect is selected, and then the selected common shot data is sorted to obtain corresponding common receiver domain data and common offset domain data as sample data for constructing a seismic sample database.

[0082] Step S406, block cutting or resampling the common shot data according to a preset size;

[0083] In the embodiment of the present application, the selected seismic data sample is subjected to data block cutting, each input common shot data and the corresponding sample category label are cut or resampled according to the size of 256*256 for data enhancement, so as to improve the sample diversity and enhance the network generalization ability.

[0084] Step S407, inputting all common shots of the common shot data into the consistency deconvolution network to generate a corresponding sample category;

[0085] The network structure adopted by the embodiment of the present application is as shown in the following figure Figure 5 The network structure diagram of the fourth embodiment of the deconvolution method of seismic data of the present application introducing the self-attention mechanism is shown in the following figure. In actual training, the network input data can be common shot data of multiple work areas, but the common shot data of the same work area contains corresponding shot domain, receiver domain and offset domain gather data. The output of the network is the result after the corresponding gather consistency deconvolution processing. The whole network is a UNET network structure of a fully convolutional neural network, which is divided into three parts of downsampling, upsampling and self-attention network block. The basic structure unit in the downsampling path is two convolution layers and one maximum pooling layer, a total of three groups. The channel number of the convolution layer is the same, the convolution kernel size is 3x3, the maximum pooling layer is 2x2, the step is 2x2, the convolution layer channel number rises from 64 to 512, and the channel number of each group is twice that of the previous group. Under the action of the pooling layer, the image downsampling is completed, and the horizontal and vertical sampling point numbers of the image become half of the previous group. In the upsampling path, the image upsampling is performed by using the deconvolution algorithm, and the method of cutting and copying is further used to splice the picture with higher resolution on the left side in the downsampling, so as to ensure the richness and resolution of the information in the upsampling process. The channel number is continuously reduced in the upsampling convolution process, and finally a 1x1 convolution is performed to obtain the final output result. In addition, the introduced self-attention mechanism network block is mainly composed of three parallel Q, K and V convolution kernels with a convolution layer of 1x1, wherein the channel number of Q and K is 128, and the channel number of V is 1024. After the tensors after the Q and K two convolution layers are correlated by inner product, the result is activated by a softmax function, and then the result after the V convolution is inner multiplied. The initial input of the self-attention network block is weighted and summed as the final output and connected to the upsampling part of the whole network.

[0086] In a specific implementation, the designed consistency deconvolution network is a single-channel input and single-channel output structure, and the size of the input and output data remains consistent. In order to ensure that the energy and amplitude of the spatial domain data before and after processing remain unchanged, the loss function used is L2 norm, and the Adam optimizer is used for iterative update.

[0087] Step S408, determining a training error according to the sample category label and the sample category, and adjusting the consistency deconvolution network based on the training error to obtain optimal network parameters, and generating a target consistency deconvolution network by using the optimal network parameters;

[0088] In the embodiment of the present application, the result after consistency processing in the training process is used as a label, different domain data is used as input data to input the consistency deconvolution network for training, and finally the trained consistency deconvolution network is obtained, and the trained network is saved;

[0089] Then the saved network is reloaded, the surface consistency processing is performed on the test data, and the effect of the consistency deconvolution network after training is verified. Since the network of the method is a full convolution network, the input data can be directly input for testing in the inference test stage without the need for cropping and other operations, and the consistency deconvolution network after training that passes the test is defined as the target consistency deconvolution network.

[0090] In a specific implementation, the input data size of this training is 256x256, and 90,000 samples are prepared. The data sample label data is prepared in the form of TFRecord. The network is established, the learning rate is set to 0.0001, the batch_size is set to 10, and the learning round number is set to 50 rounds. In addition, the cluster GPU used for this test is RTX 2080Ti, the memory size is 11G, the number of GPU cards used for training is four, and the training efficiency is improved by using data parallelism.

[0091] In the embodiment of the application, the network is trained based on the above parameters, and the consistency deconvolution network is saved. In the test process, the data size is 266*2001, and the pre-processing data input is as shown in the pre-processing data of Embodiment Four of the deconvolution method of seismic data of the application. Figure 6 The processing effect of the commercial software based on Embodiment Four of the deconvolution method of seismic data of the application is shown in the processing effect diagram of the commercial software based on Embodiment Four of the deconvolution method of seismic data of the application. Figure 7 The de-noising data after using Embodiment Four of the deconvolution method of seismic data of the application is shown in the de-noising effect diagram of Embodiment Four of the deconvolution method of seismic data of the application. Figure 8 As can be seen from the figure, the commercial software and the method of the application have good effects in consistency processing, and the consistency of the processed result wavelet is obviously enhanced. The processing effect of the application is comparable to that of the commercial software, and the comprehensive processing efficiency based on the method is 3 times that of the commercial software. Therefore, it can be concluded that the processing effect based on the application is comparable to that of the commercial software, but the processing efficiency is greatly improved.

[0092] In step S409, the to-be-tested seismic data is input into the consistency deconvolution network to obtain a corresponding prediction category.

[0093] The embodiment of the application discloses a deconvolution method of seismic data. By establishing a sample label set of multiple data domains and introducing a self-attention mechanism to fully mine the influence factors of different data domains on surface consistency deconvolution processing, the extraction of a surface-oriented deconvolution operator is realized based on a deep neural network, the network precision is improved, and the processing of pre-stack seismic consistency deconvolution is realized.

[0094] Embodiment Five, please refer toFigure 9 , Figure 9 is a structural block diagram of an embodiment of a deconvolution device for seismic data of the present application, comprising:

[0095] The data acquisition module 501 is configured to acquire seismic data to be measured and historical seismic data.

[0096] The database establishment module 502 is configured to establish a seismic sample database based on the historical seismic data.

[0097] The sample acquisition module 503 is configured to acquire seismic data training samples from the seismic sample database; the seismic data samples comprise common shot gather data and corresponding sample category labels.

[0098] The construction module 504 is configured to construct a target consistency deconvolution network based on the common shot gather data and the corresponding sample category labels.

[0099] The prediction module 505 is configured to input the seismic data to be measured into the consistency deconvolution network to obtain a corresponding predicted category.

[0100] In an optional embodiment, the sample category labels comprise common receiver domain data and common offset domain data; and the database establishment module 502 comprises:

[0101] The sorting sub-module is configured to sort the historical seismic data to obtain two types of data including common receiver domain data and common offset domain data.

[0102] The labeling sub-module is configured to label corresponding data category labels according to the data types of the historical seismic data.

[0103] The construction sub-module is configured to construct the seismic sample database according to the historical seismic data and the corresponding data category labels.

[0104] In an optional embodiment, the construction module 504 comprises:

[0105] The input sub-module is configured to input all common shots of the common shot gather data into the consistency deconvolution network to generate corresponding sample categories.

[0106] The network generation sub-module is configured to determine a training error according to the sample category labels and the sample categories, and adjust the consistency deconvolution network based on the training error to obtain optimal network parameters, and generate a target consistency deconvolution network using the optimal network parameters.

[0107] In an optional embodiment, the construction module 504 further comprises:

[0108] The preprocessing submodule is configured to block, clip or resample the common shot data according to a preset size.

[0109] The embodiment of the present application further provides an electronic device, including a memory and a processor, the memory stores a computer program, and the processor executes the computer program to perform the steps of the deconvolution method of seismic data according to any one of the above-mentioned embodiments.

[0110] The embodiment of the present application further provides a computer storage medium, which stores a computer program, and the computer program is executed by the processor to perform the steps of the deconvolution method of seismic data according to any one of the above-mentioned embodiments.

[0111] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the system, device and unit described above can refer to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0112] In several embodiments provided in the present application, it should be understood that the disclosed method, device, electronic device and storage medium can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interface, device or unit, and can be electrical, mechanical or other forms.

[0113] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on a plurality of network units. According to actual needs, part or all of the units can be selected to achieve the purpose of the embodiment scheme.

[0114] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of software functional unit.

[0115] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer-readable storage medium. Based on such understanding, the technical solutions of the present application or all or part of the technical solutions that essentially contribute to the prior art can be embodied in the form of a software product. The computer software product is stored in a readable storage medium, including a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned readable storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0116] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacements for some technical features; and these modifications or replacements do not make the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method of deconvolution of seismic data, characterized in that, The method comprises the following steps: acquiring to-be-detected seismic data and historical seismic data; establishing a seismic sample database based on the historical seismic data; acquiring seismic data training samples from the seismic sample database; the seismic data samples comprise common shot gather data and corresponding sample category labels; constructing a target consistency deconvolution network based on the common shot gather data and the corresponding sample category labels; inputting the to-be-detected seismic data into the consistency deconvolution network to obtain corresponding predicted categories.

2. The deconvolution method of seismic data according to claim 1, characterized in that, The sample category labels comprise common receiver domain data and common offset domain data; the step of establishing a seismic sample database based on the historical seismic data comprises the following steps: sorting the historical seismic data to obtain two types of data including common receiver domain data and common offset domain data; labeling corresponding data category labels according to the data types of the historical seismic data; constructing the seismic sample database according to the historical seismic data and the corresponding data category labels.

3. The deconvolution method for seismic data of claim 1, wherein, The step of constructing a target consistency deconvolution network based on the common shot gather data and the corresponding sample category labels comprises the following steps: inputting all common shots of the common shot gather data into the consistency deconvolution network to generate corresponding sample categories; determining training errors according to the sample category labels and the sample categories, and adjusting the consistency deconvolution network based on the training errors to obtain optimal network parameters, and generating a target consistency deconvolution network by using the optimal network parameters.

4. The deconvolution method of seismic data according to claim 3, characterized in that, Before the step of inputting all common shots of the common shot gather data into the consistency deconvolution network to generate corresponding sample categories, the step further comprises the following step: blocking, cropping or resampling the common shot gather data according to a preset size.

5. An apparatus for deconvolution of seismic data, characterized by The method comprises the following steps: a data acquisition module is configured to acquire to-be-detected seismic data and historical seismic data; a database establishment module is configured to establish a seismic sample database based on the historical seismic data; a sample acquisition module is configured to acquire seismic data training samples from the seismic sample database; the seismic data samples comprise common shot gather data and corresponding sample category labels; a construction module is configured to construct a target consistency deconvolution network based on the common shot gather data and the corresponding sample category labels; a prediction module is configured to input the to-be-detected seismic data into the consistency deconvolution network to obtain corresponding predicted categories.

6. The deconvolution apparatus for seismic data of claim 5, wherein, The sample category labels comprise common receiver domain data and common offset domain data; the database establishment module comprises the following modules: a sorting sub-module is configured to sort the historical seismic data to obtain two types of data including common receiver domain data and common offset domain data; a labeling sub-module is configured to label corresponding data category labels according to the data types of the historical seismic data; a construction sub-module is configured to construct the seismic sample database according to the historical seismic data and the corresponding data category labels.

7. The deconvolution apparatus for seismic data of claim 5, wherein, The construction module comprises the following modules: an input sub-module is configured to input all common shots of the common shot gather data into the consistency deconvolution network to generate corresponding sample categories; The network generation submodule is configured to determine a training error according to the sample category label and the sample category, adjust the consistency deconvolution network based on the training error to obtain optimal network parameters, and generate a target consistency deconvolution network by using the optimal network parameters.

8. The deconvolution apparatus for seismic data of claim 7, wherein, The construction module further comprises: The preprocessing submodule is configured to block, crop or resample the common shot data according to a preset size.

9. An electronic device, comprising: The computer program is executed by the processor to run the method according to any one of claims 1-4.

10. A storage medium having stored thereon a computer program, characterized in that The computer program is executed by the processor to run the method according to any one of claims 1-4.

Images