High-precision imaging method and device for seismic data

Abstract

The present invention provides a high-precision seismic data imaging method and apparatus, which relates to the technical field of seismic exploration, and includes: obtaining a pre-stack seismic gather of the work area to be imaged, performing multi-azimuth pre-stack migration processing on the pre-stack seismic gather with a fixed migration aperture for each imaging line in the imaging space to obtain a set of multi-azimuth migration imaging profiles for each imaging line, and using a target neural network model to process the set of multi-azimuth migration imaging profiles of the target imaging line to obtain a high-precision migration imaging profile of the target imaging line. The target neural network model is a lightweight neural network composed of Transformer and CNN, and the training target is that the error between the predicted result of the output migration imaging profile and the actual high-precision migration imaging profile is less than a preset threshold. Therefore, it can ensure that the output migration imaging profile is a high-precision imaging result, effectively alleviating the technical problem of low imaging accuracy existing in the existing seismic data imaging method.

Description

Technical Field

[0001] The present invention relates to the technical field of seismic exploration, and in particular to a high-precision seismic data imaging method and device. Background Art

[0002] In seismic exploration, it has always been the goal of geophysicists to obtain underground complex structure information by performing migration processing on seismic observation data on the earth's surface. This process is generally referred to as seismic data imaging. For the practice of most migration processing methods, the selection of the migration aperture is an important task. A smaller migration aperture can reduce the migration calculation amount, but there is a risk of incorrect imaging of steeply dipping structures; an overly large migration aperture brings migration noise and a large migration calculation amount. A better migration aperture can suppress migration noise and improve the calculation efficiency of migration processing.

[0003] Due to the limitation of the implementation method of the migration algorithm, in a migration operation, usually a unified migration aperture is always selected. However, because the dip angle of the underground structure changes with the spatial position, the migration aperture suitable for a certain imaging point may be too large or too small for other imaging points. In addition, since it is difficult for people to accurately estimate the dip angle of the structure to be imaged before migration processing, the selection of the migration aperture can only adopt a conservative method, that is, to select a larger aperture. This results in the failure to suppress migration noise well, and further leads to an unclear imaging profile.

[0004] In summary, the existing seismic data imaging methods have the technical problem of low imaging accuracy. Summary of the Invention

[0005] The purpose of the present invention is to provide a high-precision seismic data imaging method and device to alleviate the technical problem of low imaging accuracy existing in the existing seismic data imaging methods.

[0006] In a first aspect, the present invention provides a method for high-precision seismic data imaging, including: obtaining a pre-stack seismic gather of the work area to be imaged; for each imaging line in the imaging space of the work area to be imaged, performing multi-azimuth pre-stack migration processing on the pre-stack seismic gather with a fixed migration aperture to obtain a multi-azimuth migration imaging profile set for each imaging line; using a target neural network model to process the multi-azimuth migration imaging profile set of the target imaging line to obtain a high-precision migration imaging profile of the target imaging line; wherein, the target neural network model is a lightweight neural network composed of a Transformer neural network and a CNN neural network, and the Transformer neural network is located between the encoding module and the decoding module of the CNN neural network; the training objective of the target neural network model is that the error between the predicted result of the output migration imaging profile and the actual high-precision migration imaging profile is less than a preset threshold; the target imaging line represents any imaging line in the imaging space; based on the high-precision migration imaging profiles of all imaging lines, determining the high-precision imaging result of the work area to be imaged.

[0007] Optionally, it further includes: obtaining a pre-stack seismic gather of the training work area; determining high-precision migration imaging profiles and corresponding multi-azimuth migration imaging profile sets of multiple typical imaging lines based on the pre-stack seismic gather of the training work area; wherein, the high-precision migration imaging profile is an imaging profile obtained by performing migration processing with the Fresnel zone of the imaging point as the migration aperture; using the multi-azimuth migration imaging profile set of each typical imaging line as the input data of the initial neural network model, and using the corresponding high-precision migration imaging profile as the training label, training the initial neural network model until the training objective is achieved to obtain the target neural network model.

[0008] Optionally, determining high-precision migration imaging profiles of multiple typical imaging lines based on the pre-stack seismic gather of the training work area includes: calculating a multi-azimuth dip-domain migration gather of the target typical imaging line based on the pre-stack seismic gather of the training work area; wherein, the target typical imaging line represents any imaging line among the multiple typical imaging lines; based on the multi-azimuth dip-domain migration gather, manually picking up the Fresnel zones of multiple typical imaging points; using the Fresnel zones of multiple typical imaging points as the migration aperture, performing pre-stack migration processing on the pre-stack seismic gather of the training work area to obtain the high-precision migration imaging profile of the target typical imaging line.

[0009] Optionally, the initial neural network model further includes: a normalization layer and an inverse normalization layer; using the multi-azimuth migration imaging profile set of each typical imaging line as the input data of the initial neural network model, and using the corresponding high-precision migration imaging profile as the training label, training the initial neural network model includes: the normalization layer respectively normalizes the multi-azimuth migration imaging profile set of the target typical imaging line and the corresponding high-precision migration imaging profile to obtain the multi-azimuth migration imaging profile set after amplitude normalization and the high-precision migration imaging profile after amplitude normalization; the encoding module encodes each local time window data in the multi-azimuth migration imaging profile set after amplitude normalization to extract the local features and profile position information of each local time window data; expanding the local features of all local time window data to construct a long sequence of local features; the Transformer neural network extracts the global features and effective features of the long sequence of local features to obtain a target feature sequence; folding the target feature sequence, and the decoding module decodes the folded features and profile position information to obtain a reconstructed migration imaging profile; calculating a loss function value based on the reconstructed migration imaging profile and the high-precision migration imaging profile after amplitude normalization to train the initial neural network model based on the loss function value; the inverse normalization layer performs inverse normalization processing on the reconstructed migration imaging profile based on the multi-azimuth migration imaging profile set of the target typical imaging line and the corresponding high-precision migration imaging profile to obtain a high-precision migration imaging profile after amplitude recovery.

[0010] Optionally, the normalization layer respectively normalizes the multi-azimuth migration imaging profile set of the target typical imaging line and the corresponding high-precision migration imaging profile, including: determining the amplitude value of each position point in the multi-azimuth migration imaging profile set of the target typical imaging line to obtain an amplitude value set corresponding to each migration imaging profile; adding the amplitude values with the same serial number in all amplitude value sets and taking the absolute value after addition to obtain a first target amplitude value set; using the maximum amplitude value in the first target amplitude value set as the normalization parameter of the multi-azimuth migration imaging profile set to normalize the multi-azimuth migration imaging profile set; determining the amplitude value of each position point in the high-precision migration imaging profile of the target typical imaging line and taking the absolute value to obtain a second target amplitude value set; using the maximum amplitude value in the second target amplitude value set as the normalization parameter of the high-precision migration imaging profile to normalize the high-precision migration imaging profile.

[0011] Optionally, the denormalization layer performs denormalization processing on the reconstructed migration imaging profile based on the multi-azimuth migration imaging profile set of the target typical imaging line and the corresponding high-precision migration imaging profile, including: determining the amplitude value of each position point in the multi-azimuth migration imaging profile set of the target typical imaging line, to determine the energy average value of the multi-azimuth migration imaging profile set based on the amplitude value, and obtaining the first energy; determining the amplitude value of each position point in the high-precision migration imaging profile of the target typical imaging line, to determine the energy average value of the high-precision migration imaging profile based on the amplitude value, and obtaining the second energy, and determining the maximum amplitude value in the high-precision migration imaging profile; determining the denormalization parameter of the reconstructed migration imaging profile based on the first energy, the second energy, and the maximum amplitude value; performing denormalization processing on the reconstructed migration imaging profile based on the denormalization parameter, to obtain the high-precision migration imaging profile with restored amplitude.

[0012] Optionally, the selection criterion for the typical imaging line is: at the position with drastic structural changes, select the imaging line based on the first interval, and at the position with gentle structure, select the imaging line based on the second interval, where the first interval is less than the second interval.

[0013] In a second aspect, the present invention provides a seismic data high-precision imaging device, including a first acquisition module, configured to acquire the pre-stack seismic gather of the work area to be imaged; a migration module, configured to perform multi-azimuth pre-stack migration processing on the pre-stack seismic gather with a fixed migration aperture for each imaging line in the imaging space of the work area to be imaged, to obtain the multi-azimuth migration imaging profile set of each imaging line; a processing module, configured to process the multi-azimuth migration imaging profile set of the target imaging line by using the target neural network model, to obtain the high-precision migration imaging profile of the target imaging line; where the target neural network model is a lightweight neural network composed of a Transformer neural network and a CNN neural network, and the Transformer neural network is located between the encoding module and the decoding module of the CNN neural network; the training objective of the target neural network model is that the error between the predicted result of the output migration imaging profile and the actual high-precision migration imaging profile is less than a preset threshold; the target imaging line represents any imaging line in the imaging space; a first determination module, configured to determine the high-precision imaging result of the work area to be imaged based on the high-precision migration imaging profiles of all imaging lines.

[0014] In a third aspect, the present invention provides an electronic device, including a memory and a processor, where a computer program is stored on the memory and can run on the processor, and when the processor executes the computer program, it implements the seismic data high-precision imaging method described in any one of the foregoing embodiments.

[0015] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions, which, when executed by a processor, implement the high-precision seismic data imaging method described in any one of the foregoing embodiments.

[0016] The present invention provides a high-precision seismic data imaging method. After obtaining the pre-stack seismic gathers of the work area to be imaged, first, for each imaging line in the imaging space, perform multi-azimuth pre-stack migration processing on the pre-stack seismic gathers with a fixed migration aperture to obtain a set of multi-azimuth migration imaging profiles for each imaging line. Then, use the target neural network model to process the set of multi-azimuth migration imaging profiles of the target imaging line to obtain a high-precision migration imaging profile of the target imaging line. The target neural network model is a lightweight neural network composed of a Transformer neural network and a CNN neural network, and the training target is that the error between the predicted result of the output migration imaging profile and the actual high-precision migration imaging profile is less than a preset threshold. Therefore, it can ensure that the output migration imaging profile is a high-precision imaging result, effectively alleviating the technical problem of low imaging accuracy existing in the existing seismic data imaging methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0018] Figure 1 It is a flowchart of a high-precision seismic data imaging method provided by an embodiment of the present invention;

[0019] Figure 2 It is a schematic diagram of the data processing flow of an initial neural network model provided by an embodiment of the present invention;

[0020] Figure 3 It is a schematic diagram of the conversion process from the output data of an encoding module to the input data of a decoding module provided by an embodiment of the present invention;

[0021] Figure 4 It is a schematic diagram of a set of multi-azimuth migration imaging profiles provided by an embodiment of the present invention;

[0022] Figure 5 It is a comparison diagram of the imaging results of multiple seismic data imaging methods provided by an embodiment of the present invention;

[0023] Figure 6Functional module diagram of a high-precision seismic data imaging device provided by an embodiment of the present invention;

[0024] Figure 7 Schematic diagram of an electronic device provided by an embodiment of the present invention. Specific embodiments

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Usually, the components of the embodiments of the present invention described and illustrated in the accompanying drawings here can be arranged and designed in various different configurations.

[0026] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed present invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0027] The following will describe in detail some embodiments of the present invention with reference to the accompanying drawings. Without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0028] Embodiment 1

[0029] In the prior art, before seismic migration imaging, if a conservative method is used to select a large migration aperture, the migration noise cannot be well suppressed, resulting in an unclear imaging profile. Research shows that during the migration imaging process, if the Fresnel zone of the imaging point is used as the migration aperture for migration calculation, both correct imaging of steep dips and suppression of migration noise can be taken into account. Strictly speaking, for each imaging point, its Fresnel zone needs to be determined based on the full (multiple) azimuth dip-domain migration gathers. However, the number of imaging points in the entire work area is huge, and determining the Fresnel zones of all imaging points will consume huge computing and storage resources.

[0030] In order to reduce the computational load and requirements for storage resources, generally, first, by constructing dip-domain migration gathers along the inline direction and crossline direction during the migration process, the Fresnel zones are directly and vividly displayed in the above-mentioned migration gathers; second, for each imaging point, a rectangle is constructed by regarding the dip values corresponding to the Fresnel zone boundaries in the inline direction and crossline direction picked interactively based on the dip-domain migration gathers as the long side and short side of the rectangle, and the constructed rectangle is approximately regarded as the Fresnel zone area of the imaging point, i.e., the migration aperture; then, the obtained migration aperture is applied to the migration process for reprocessing to obtain a migration imaging result that takes into account steep-dip imaging and migration noise suppression.

[0031] However, in the above prestack migration technology based on the Fresnel zones of imaging points as the migration aperture, on the one hand, the practice of approximately regarding the rectangle constructed by the dip values corresponding to the Fresnel zone boundaries in the inline direction and crossline direction picked interactively as the Fresnel zone area of the imaging point is based on the basic assumption that the underground three-dimensional structure changes gently in directions other than the inline and crossline directions. This approximation has a problem that when the underground three-dimensional structure changes relatively complexly, such as when there are drastic changes in multiple azimuths, the accuracy of the set migration aperture is not high, resulting in the loss of steep-dip structures in the underground part of the azimuth in the migration imaging result. On the other hand, as a key link in the implementation process of this existing technology, even if only the Fresnel zone boundaries in the inline and crossline directions of the imaging points are picked interactively based on the dip-domain migration gathers to determine the migration aperture, the time consumed by data processing personnel is huge, which greatly limits the popularization and application of this technology. That is to say, the above prestack migration technology based on the Fresnel zones in the inline and crossline directions of imaging points as the migration aperture still has problems of low imaging accuracy and huge time consumption for data processing personnel.

[0032] In view of this, the embodiments of the present invention provide a high-precision seismic data imaging method to solve the technical problems involved above. Figure 1 The flowchart of a high-precision seismic data imaging method provided by the embodiments of the present invention is as Figure 1 shown, and the method specifically includes the following steps:

[0033] Step S102, obtaining the prestack seismic gathers of the work area to be imaged.

[0034] Step S104, for each imaging line in the imaging space of the work area to be imaged, performing multi-azimuth prestack migration processing on the prestack seismic gathers with a fixed migration aperture to obtain a set of multi-azimuth migration imaging profiles for each imaging line.

[0035] In the embodiments of the present invention, the source data used for high-precision imaging of the imaging work area is the seismic pre-stack gather of this work area, and the seismic pre-stack gather can be obtained by professional seismic data acquisition equipment. Next, low-precision imaging is first performed on each imaging line in this work area. Specifically, multi-azimuth pre-stack migration processing is performed on the seismic pre-stack gather with a fixed migration aperture to obtain a set of multi-azimuth migration imaging profiles for each imaging line. That is to say, for each imaging line, multiple migration imaging profiles are output, and each migration imaging profile is obtained by migrating the seismic data input for the corresponding azimuth. In the embodiments of the present invention, the fixed migration aperture is not less than the maximum dip value (known parameter) of the subsurface structure to be imaged.

[0036] If the imaging space contains imaging lines, each imaging line contains CDPs (Common Depth Points), each CDP contains imaging points, and the coordinates of any one imaging point are , then its multi-azimuth migration imaging result is expressed as: , . Among them, represents the number of azimuth angle intervals, represents the preset azimuth angle interval length (interval), represents the two-way travel time of seismic waves, and respectively represent the travel times of seismic waves from the shot point propagating to the imaging point and from this imaging point to the receiving point , represents taking the first derivative of the seismic data sequence of the jth seismic trace, k represents the number of seismic traces of all input seismic pre-stack gathers, i represents the azimuth angle number of the seismic trace, , where θ represents the azimuth angle of the seismic trace and is calculated by the following formula: .

[0037] is the weight function, indicating the application of a fixed migration aperture during the migration process, and its expression is: , among which, , respectively represent the maximum formation dips in the x and y directions of the preset work area (input migration parameters, which are constants), , , represents the root-mean-square velocity of the imaging point (which is also a known migration parameter).

[0038] Based on the expression, it can be known that is a four-dimensional data volume, and the dimension sizes are respectively , , , . Among them, if , , remain unchanged, and only changes, that is, the single azimuth imaging result at the same CDP, which is called a single azimuth imaging seismic trace; if , remains unchanged, , changes, that is, the single azimuth imaging result at the same imaging line, which is called a single azimuth imaging profile; if only remains unchanged, , , all change, that is, the multi-azimuth imaging result at the same imaging line, which is called a multi-azimuth migration imaging profile set.

[0039] Step S106: Use the target neural network model to process the multi-azimuth migration imaging profile set of the target imaging line to obtain the high-precision migration imaging profile of the target imaging line.

[0040] It is known that the imaging accuracy of the migration imaging profile obtained after prestack migration processing with a fixed migration aperture is relatively low. Therefore, in the embodiments of the present invention, a target neural network model is pre-trained. Among them, the target neural network model is a lightweight neural network composed of a Transformer neural network and a CNN neural network, and the Transformer neural network is located between the encoding module and the decoding module of the CNN neural network; the training target of the target neural network model is that the error between the predicted result of the output migration imaging profile and the actual high-precision migration imaging profile is less than a preset threshold; optionally, the actual high-precision migration imaging profile is obtained by calibrating the Fresnel zone of the imaging point based on the multi-azimuth dip domain migration gather and then performing migration calculation using it as the migration aperture.

[0041] That is to say, after inputting the multi-azimuth migration imaging profile set of the target imaging line into the target neural network model, the predicted result of the migration imaging profile output by the target neural network model is the high-precision migration imaging profile of the target imaging line, where the target imaging line represents any imaging line in the imaging space.

[0042] The target neural network model combines the advantages of the CNN neural network and the Transformer neural network, significantly reducing the video memory requirements and training time for neural network training, thereby improving the computational efficiency. Moreover, the application of the target neural network model in the prestack migration imaging process enables seismic data processors to perform multi-azimuth migration processing on the input seismic data (i.e., seismic prestack gathers) with a fixed migration aperture, achieving both correct imaging of subsurface steep-dip structures and suppression of migration noise. This not only avoids the huge workload of determining the spatially variant migration aperture in the imaging space but also improves the accuracy of migration imaging.

[0043] Step S108: Based on the high-precision migration imaging profiles of all imaging lines, determine the high-precision imaging result of the area to be imaged.

[0044] In the embodiment of the present invention, the set of high-precision migration imaging profiles of all imaging lines is used as the high-precision imaging result of the area to be imaged.

[0045] The embodiment of the present invention provides a method for high-precision imaging of seismic data. After obtaining the seismic prestack gathers of the area to be imaged, first, for each imaging line in the imaging space, perform multi-azimuth prestack migration processing on the seismic prestack gathers with a fixed migration aperture to obtain a set of multi-azimuth migration imaging profiles for each imaging line. Then, use the target neural network model to process the set of multi-azimuth migration imaging profiles of the target imaging line to obtain the high-precision migration imaging profile of the target imaging line. The target neural network model is a lightweight neural network composed of a Transformer neural network and a CNN neural network, and the training objective is that the error between the predicted result of the output migration imaging profile and the actual high-precision migration imaging profile is less than a preset threshold. Therefore, it can ensure that the output migration imaging profile is a high-precision imaging result, effectively alleviating the technical problem of low imaging accuracy existing in the existing seismic data imaging methods.

[0046] The above text introduced that the embodiment of the present invention uses a trained target neural network model for high-precision imaging of seismic data. Next, a specific introduction will be given to the method of obtaining the target neural network model through training. In an optional embodiment, the embodiment of the present invention further includes the following steps:

[0047] Step S201: Obtain the seismic prestack gathers of the training area.

[0048] Step S202: Based on the seismic prestack gathers of the training area, determine the high-precision migration imaging profiles and the corresponding sets of multi-azimuth migration imaging profiles of multiple typical imaging lines; among them, the high-precision migration imaging profile is the imaging profile obtained by performing migration processing with the Fresnel zone of the imaging point as the migration aperture.

[0049] Step S203: Use the multi-azimuth migration imaging profile sets of each typical imaging line as the input data of the initial neural network model, and use the corresponding high-precision migration imaging profiles as training labels to train the initial neural network model until the training objective is achieved, obtaining the target neural network model.

[0050] Specifically, the embodiments of the present invention do not specifically limit the number and location of training work areas. If the number of training work areas is 1, the training work area is the work area to be imaged; if the number of training work areas is multiple, it is not required that they must include the above-mentioned work area to be imaged. In theory, the more the number of training work areas, the better the diversity of the training samples of the model, and thus the stronger the generalization ability of the target neural network.

[0051] After obtaining the pre-stack seismic gathers of the training work area, the method in step S104 above can be referred to to determine the multi-azimuth migration imaging profile sets of multiple typical imaging lines based on the pre-stack seismic gathers of the training work area. The high-precision migration imaging profile of each typical imaging line is the imaging profile obtained after performing migration processing with the Fresnel zone of the imaging point as the migration aperture. To ensure the accuracy of the training labels, the Fresnel zone of the imaging point is calibrated based on the multi-azimuth dip-domain migration gathers. In the embodiments of the present invention, the number of azimuths included in the multi-azimuth must be greater than 2, and the number of azimuths included in the multi-azimuth is , , indicating the length of the preset azimuth angle interval.

[0052] In an optional embodiment, the selection criterion for typical imaging lines is: at positions with drastic structural changes, select imaging lines based on the first interval; at positions with gentle structures, select imaging lines based on the second interval, and the first interval is less than the second interval.

[0053] Specifically, at positions with drastic structural changes, the geological structure is complex, and denser imaging lines (the first interval) are required to capture detailed geological information to ensure the accuracy and reliability of the imaging results. At positions with gentle structures, the geological structure is relatively simple, and sparser imaging lines (the second interval) can be used to reduce the data volume and processing time while still being able to obtain sufficient geological information. Moreover, by using the first interval to select imaging lines in areas with drastic structural changes, limited resources can be utilized more effectively to ensure detailed imaging of key areas. By using the second interval to select imaging lines in areas with gentle structures, unnecessary data processing can be reduced, saving time and costs. The above flexible selection criterion for typical imaging lines can be adjusted according to different geological conditions to ensure high-quality imaging results in various geological environments and optimize resource utilization.

[0054] After obtaining the high-precision migration imaging profiles of multiple typical imaging lines and the corresponding set of multi-azimuth migration imaging profiles, use all the sets of multi-azimuth migration imaging profiles as the input data set for the initial neural network training, and use all the high-precision migration imaging profiles as the label data set for the initial neural network training. Based on the correspondence between the input data and the label data, train the initial neural network model until the error between the predicted result of the migration imaging profile output by the model and the training label is less than the preset threshold, then the training can be ended to obtain the target neural network model.

[0055] In an optional implementation manner, in the above step S202, determining the high-precision migration imaging profiles of multiple typical imaging lines based on the seismic pre-stack gathers in the training area specifically includes the following steps:

[0056] Step S2021, calculate the multi-azimuth dip-domain migration gathers of the target typical imaging line based on the seismic pre-stack gathers in the training area; wherein, the target typical imaging line represents any one of the multiple typical imaging lines.

[0057] Step S2022, interactively pick the Fresnel zones of multiple typical imaging points based on the multi-azimuth dip-domain migration gathers.

[0058] Step S2023, use the Fresnel zones of multiple typical imaging points as the migration aperture, and perform pre-stack migration processing on the seismic pre-stack gathers in the training area to obtain the high-precision migration imaging profile of the target typical imaging line.

[0059] In the existing pre-stack migration technology that uses the Fresnel zone of the imaging point as the migration aperture, the Fresnel zones in the Inline direction and the Crossline direction picked interactively are used as the migration aperture. Compared with the existing technology, in the embodiment of the present invention, when determining the training label, a "multi-azimuth (i.e., all-round)" dip-domain migration gather is constructed, and then the Fresnel zones of typical imaging points are picked interactively. Compared with the method of picking based on the two-azimuth (i.e., Inline direction and Crossline direction) dip-domain migration gather in the existing technology, the obtained Fresnel zone is more accurate, thus ensuring the accuracy of the pre-stack migration processing result. The embodiment of the present invention does not specifically limit the method for selecting typical imaging points, and the selection criteria of typical imaging lines can be referred to, that is, in the positions where the structure changes violently, the selected imaging points are relatively dense, and in the positions where the structure is gentle, the selected imaging points are relatively sparse.

[0060] In an alternative embodiment, the initial neural network model further includes: a normalization layer and an inverse normalization layer; in step S203, the set of multi-azimuth offset imaging profiles of each typical imaging line is used as the input data of the initial neural network model, and the corresponding high-precision offset imaging profile is used as the training label to train the initial neural network model, which specifically includes the following steps:

[0061] Step S2031: The normalization layer respectively performs normalization processing on the set of multi-azimuth offset imaging profiles of the target typical imaging line and the corresponding high-precision offset imaging profile to obtain the set of multi-azimuth offset imaging profiles after amplitude normalization and the high-precision offset imaging profile after amplitude normalization.

[0062] Specifically, considering that the amplitude range of seismic data is too large, possibly up to the million level, it is not conducive to calculation if directly input into the neural network. Therefore, in the embodiments of the present invention, normalization is used to narrow the value range of the input data and label data, so as to facilitate the training of the initial neural network model. In addition, considering that the amplitude (value range) of seismic data varies greatly, using the same normalization parameters for the input data and label data for normalization will result in a large error between the predicted amplitude and the actual value when the trained neural network is in inference. In view of this, in the embodiments of the present invention, the normalization parameters of the set of multi-azimuth offset imaging profiles and the normalization parameters of the high-precision offset imaging profile need to be calculated separately, and then the two are respectively normalized.

[0063] Step S2032: The encoding module encodes each local time window data in the set of multi-azimuth offset imaging profiles after amplitude normalization to extract the local features and profile position information of each local time window data.

[0064] Step S2033: Expand the local features of all local time window data to construct a long sequence of local features.

[0065] Step S2034: The Transformer neural network extracts the global features and effective features of the long sequence of local features to obtain the target feature sequence.

[0066] Step S2035: Fold the target feature sequence, and the decoding module decodes the folded features and profile position information to obtain the reconstructed offset imaging profile.

[0067] Step S2036: Based on the reconstructed offset imaging profile and the high-precision offset imaging profile after amplitude normalization, calculate the loss function value to train the initial neural network model based on the loss function value.

[0068] Step S2037: The denormalization layer performs denormalization processing on the reconstructed migration imaging profile based on the multi-azimuth migration imaging profile set of the target typical imaging line and the corresponding high-precision migration imaging profile, and obtains the high-precision migration imaging profile with restored amplitude.

[0069] In the embodiment of the present invention, the neural network model has multiple inputs and a single output. Its input is the multi-azimuth migration imaging profile set, and the output is the high-precision migration profile reconstructed after the neural network extracts and fuses the features of the multi-azimuth migration imaging profile set. Figure 2 It is a schematic diagram of the data processing flow of an initial neural network model provided by an embodiment of the present invention. Among them, the convolutional neural network (CNN) includes an encoding module (encoder block) and a decoding module (decoder block). The embodiment of the present invention does not specifically limit the structures of the encoding module and the decoding module, as long as the above-mentioned encoding function and decoding function can be realized.

[0070] In an optional embodiment, in the encoding module, each layer contains 2 3×3 convolutional operations for feature extraction and feature mapping of data; its activation layer uses ReLU as the activation function to increase non-linear mapping; its pooling layer is 2×2 max pooling for feature extraction and downsampling of data, reducing the parameters of the neural network to facilitate the training of the subsequent Transformer model. In the decoding module, each layer contains 2 3×3 convolutional operations, an activation layer with ReLU as the activation function, and an upsampling operation with a size of 2×2 for restoring the sampling size of the data and reducing the number of channels. In addition, its last layer adds a 1×1 convolutional kernel for fusing and reconstructing the high-precision migration result data.

[0071] Figure 3 It is a schematic diagram of the conversion process from the output data of the encoding module to the input data of the decoding module provided by an embodiment of the present invention. Through Figure 3It can be seen that the data output by the encoding module is essentially multiple two-dimensional feature image blocks. By flattening the elements at the same position in the multiple two-dimensional feature image blocks into one dimension (i.e., feature expansion), multiple one-dimensional sequences can be obtained. The above multiple one-dimensional sequences are connected end to end in a specified order to construct a local feature long sequence for inputting into the Transformer neural network. The Transformer neural network can capture the global features and effective features contained in the input data to obtain a target feature sequence. Among them, the global feature represents the features possessed by the sample data as a whole; the effective feature represents the features possessed by the effective data in the sample excluding interference such as noise. Next, the target feature sequence is folded to construct multiple two-dimensional feature image blocks with the same data structure as the output data of the encoding module. The decoding module can obtain the reconstructed migration imaging profile by decoding it.

[0072] Next, according to the reconstructed migration imaging profile and the high-precision migration imaging profile after amplitude normalization, the loss function value can be calculated, and then the initial neural network model can be trained based on the loss function value. The embodiments of the present invention do not specifically calculate the calculation method of the loss function value, and users can design it according to the actual situation, as long as it can ensure that the greater the error between the reconstructed migration imaging profile and the high-precision migration imaging profile after amplitude normalization, the greater the loss function value. When training the model, batch processing can also be used for training.

[0073] The prediction results of the neural network are generally normalized to [-1, 1] or [0, 1]. Therefore, after the decoding module outputs the prediction results (i.e., the reconstructed migration imaging profile), the prediction results need to be amplitude restored (i.e., "inverse normalization") to obtain effective seismic amplitude values. The function of amplitude restoration is implemented by the inverse normalization layer. After learning the amplitude relationship between the input data (multi-azimuth migration imaging profile set) and the label data (high-precision migration imaging profile), the inverse normalization layer can perform inverse normalization processing on the reconstructed migration imaging profile to obtain the high-precision migration imaging profile after amplitude restoration.

[0074] The embodiments of the present invention do not specifically limit the structure of the Transformer neural network, and users can choose according to actual needs. Optionally, for the Transformer neural network model, each layer is composed of an efficient multi-head attention (EMHA) and a multi-layer perceptron (MLP) block.

[0075] Among them, in the multi-head self-attention mechanism (EMHA), the input sequence first passes through three different linear transformation layers to obtain Query, Key, and Value respectively. Then, these transformed vectors are divided into several "heads", and each head has its own independent Query, Key, and Value matrices. For each head, a scaled dot-product attention operation is performed. Finally, the outputs of all heads are concatenated together and fused through a linear layer to obtain the final attention output vector. In this way, the multi-head self-attention mechanism (EMHA) can perform attention processing on the input sequence from different perspectives in parallel, improving the ability of the Transformer neural network model to understand and capture complex dependencies. Or, the multi-head self-attention mechanism (EMHA) obtains the attention distributions of different subspaces of the input sequence by running multiple independent attention mechanisms in parallel, thereby more comprehensively capturing various potential feature associations in the sequence.

[0076] The multi-layer perceptron block (MLP) is a component for further processing information after the multi-head self-attention mechanism. Through two linear transformations and a non-linear activation function operation, the Transformer neural network model can effectively process and represent complex input data while capturing dependencies at different positions in the sequence.

[0077] In an optional implementation manner, in step S2031, the normalization layer normalizes the multi-azimuth offset imaging profile set of the target typical imaging line and the corresponding high-precision offset imaging profile respectively, which specifically includes the following steps:

[0078] Step S20311: Determine the amplitude value of each position point in the multi-azimuth offset imaging profile set of the target typical imaging line to obtain the amplitude value set corresponding to each offset imaging profile.

[0079] Step S20312: Superimpose the amplitude values with the same serial number in all amplitude value sets, and take the absolute value after superimposition to obtain the first target amplitude value set.

[0080] Step S20313: Use the maximum amplitude value in the first target amplitude value set as the normalization parameter for the multi-azimuth offset imaging profile set to normalize the multi-azimuth offset imaging profile set.

[0081] Step S20314: Determine the amplitude value of each position point in the high-precision offset imaging profile of the target typical imaging line, and take the absolute value to obtain the second target amplitude value set.

[0082] Step S20315: Use the maximum amplitude value in the second target amplitude value set as the normalization parameter for the high-precision migration imaging profile to perform normalization processing on the high-precision migration imaging profile.

[0083] Specifically, the set of amplitude values corresponding to all migration imaging profiles in the multi-azimuth migration imaging profile set of the target typical imaging line can be expressed as: , where represents the set of amplitude values of all position points on the migration imaging profile with azimuth angle number i, or can also be embodied in the form of a two-dimensional array, and the positions of the array elements correspond to the pixel positions. Based on the description of the normalization processing of the multi-azimuth migration imaging profile set in the above text, the normalization parameter of the multi-azimuth migration imaging profile set is: , and the normalization processing result of the multi-azimuth migration imaging profile set is expressed as: .

[0084] If the set of amplitude values of all position points on the high-precision migration imaging profile is L, then the normalization parameter of the high-precision migration imaging profile is , and the normalization processing result of the high-precision migration imaging profile is expressed as: .

[0085] In an optional implementation manner, in the above step S2037, the denormalization layer performs denormalization processing on the reconstructed migration imaging profile based on the multi-azimuth migration imaging profile set of the target typical imaging line and the corresponding high-precision migration imaging profile, which specifically includes the following steps:

[0086] Step S20371: Determine the amplitude value of each position point in the multi-azimuth migration imaging profile set of the target typical imaging line, and based on the amplitude value, determine the energy average value of the multi-azimuth migration imaging profile set to obtain the first energy.

[0087] Step S20372: Determine the amplitude value of each position point in the high-precision migration imaging profile of the target typical imaging line, and based on the amplitude value, determine the energy average value of the high-precision migration imaging profile to obtain the second energy, and determine the maximum amplitude value in the high-precision migration imaging profile.

[0088] Step S20373: Based on the first energy, the second energy, and the maximum amplitude value, determine the denormalization parameter of the reconstructed migration imaging profile.

[0089] Step S20374: Perform denormalization processing on the reconstructed migration imaging profile based on the denormalization parameter to obtain the high-precision migration imaging profile with restored amplitude.

[0090] In the embodiments of the present invention, the denormalization parameter is expressed as , where , , , represents the first energy, represents the second energy, and the energy at each position point in the imaging profile is the square of its amplitude. represents the maximum amplitude value in the high-precision migration imaging profile. The anti-normalization parameter is used to perform anti-normalization processing on the reconstructed migration imaging profile, that is, amplitude recovery processing, which is expressed as: , where represents the set of amplitude values at all position points in the reconstructed migration imaging profile, represents the set of amplitude values at all position points in the high-precision migration imaging profile after amplitude recovery.

[0091] In one embodiment, when training the initial neural network, the azimuth interval ∆θ = 45 degrees is specified in advance, that is, the multi-azimuth is specifically azimuths. In the neural network, the CNN layer of the network is set to 5 layers, the initial convolution kernel is set to 32, and the number of transformer heads is set to 8. The size of the input data is 128×128×8, the initial learning rate is set to 0.001, and it is halved every 200 training epochs. The Adam optimizer and the smooth L1 loss function are used to train the neural network model. During the training process, first calculate the total losses of the training dataset and the validation dataset, and then divide them by their respective occupied storage spaces. After 400 training epochs, the loss tends to be stable. For 8.6G of training data, the neural network training can be completed in 4.8 hours.

[0092] In order to verify the performance of the embodiments of the present invention, Figure 4 is a schematic diagram of a set of multi-azimuth migration imaging profiles provided by the embodiments of the present invention. Figure 4 In the views from a to h, they respectively correspond to the following azimuth ranges: (a) , (b) , (c) , (d) , (e) , (f) , (g) , (h) . Figure 5 is a comparison diagram of the imaging results of various seismic data imaging methods provided by the embodiments of the present invention. Figure 5 In the a view, it is the result of directly superimposing the multi-azimuth migration profile set. Figure 5The middle b view is the result of post - migration stacking after manual interactive picking of Fresnel zones based on the dip - domain migration gathers in two azimuths (Inline direction and Crossline direction). Figure 5 The middle c view is the prediction result of applying the target neural network model in the embodiments of the present invention.

[0093] Figure 5 The middle a view is the directly stacked profile of the multi - azimuth migration profile set, which is equivalent to the migration profile with a fixed migration aperture. In this result image, the signal - to - noise ratio is very low, and the underground geological structure is not clearly depicted: the continuity of the horizontal strata is damaged, and the breakpoints of the faults are blurred. Compared with the a view, Figure 5 In the middle b view, the signal - to - noise ratio is significantly improved, especially the random noise is significantly suppressed. The continuity of the horizontal strata and the depiction of the breakpoints of the faults are significantly improved compared with the a view. However, there are obvious regular noises in the time window of CDP850 - 1050 and Time1.60S - 3.10S, which is not conducive to the accurate depiction of the large faults from top to bottom in this area. It can be seen from the investigation that Figure 5 On the profile shown in the middle c view, both random noise and regular noise are significantly suppressed, and the complex structure of the underground medium is depicted more clearly: the imaging continuity of the horizontal strata is good, the fault breakpoints are distinct, and the depiction is accurate.

[0094] Embodiment 2

[0095] The embodiments of the present invention also provide a high - precision seismic data imaging device, which is mainly used to execute the high - precision seismic data imaging method provided in the above Embodiment 1. The following is a specific introduction to the high - precision seismic data imaging device provided in the embodiments of the present invention.

[0096] Figure 6 It is a functional module diagram of a high - precision seismic data imaging device provided in the embodiments of the present invention. As Figure 6 shown, the device mainly includes: a first acquisition module 10, a migration module 20, a processing module 30, and a first determination module 40, where:

[0097] The first acquisition module 10 is used to acquire the pre - stack seismic gathers of the work area to be imaged.

[0098] The migration module 20 is used to perform multi - azimuth pre - stack migration processing on the pre - stack seismic gathers with a fixed migration aperture for each imaging line in the imaging space of the work area to be imaged, and obtain a multi - azimuth migration imaging profile set for each imaging line.

[0099] A processing module 30 is configured to process a multi-azimuth migration imaging profile set of a target imaging line by using a target neural network model to obtain a high-precision migration imaging profile of the target imaging line. The target neural network model is a lightweight neural network composed of a Transformer neural network and a CNN neural network, and the Transformer neural network is located between the encoding module and the decoding module of the CNN neural network. The training objective of the target neural network model is that the error between the predicted result of the migration imaging profile output and the actual high-precision migration imaging profile is less than a preset threshold. The target imaging line represents any imaging line in the imaging space.

[0100] A first determination module 40 is configured to determine a high-precision imaging result of the to-be-imaged work area based on the high-precision migration imaging profiles of all imaging lines.

[0101] An embodiment of the present invention provides a seismic data high-precision imaging device. After obtaining the pre-stack seismic gather of the to-be-imaged work area, first, for each imaging line in the imaging space, perform multi-azimuth pre-stack migration processing on the pre-stack seismic gather with a fixed migration aperture to obtain a multi-azimuth migration imaging profile set of each imaging line. Then, use the target neural network model to process the multi-azimuth migration imaging profile set of the target imaging line to obtain a high-precision migration imaging profile of the target imaging line. The target neural network model is a lightweight neural network composed of a Transformer neural network and a CNN neural network, and the training objective is that the error between the predicted result of the migration imaging profile output and the actual high-precision migration imaging profile is less than a preset threshold. Therefore, it can ensure that the output migration imaging profile is a high-precision imaging result, effectively alleviating the technical problem of low imaging accuracy existing in the existing seismic data imaging method.

[0102] Optionally, the device further includes:

[0103] A second acquisition module is configured to acquire the pre-stack seismic gather of the training work area.

[0104] A second determination module is configured to determine the high-precision migration imaging profiles and corresponding multi-azimuth migration imaging profile sets of multiple typical imaging lines based on the pre-stack seismic gather of the training work area. The high-precision migration imaging profile is an imaging profile obtained by performing migration processing with the Fresnel zone of the imaging point as the migration aperture.

[0105] A training module is configured to use the multi-azimuth migration imaging profile set of each typical imaging line as the input data of the initial neural network model, and use the corresponding high-precision migration imaging profile as the training label to train the initial neural network model until the training objective is reached to obtain the target neural network model.

[0106] Optionally, the second determination module includes:

[0107] A calculation unit for calculating a multi - azimuth dip - domain migration gather of a target typical imaging line based on a pre - stack seismic gather of a training work area; wherein, the target typical imaging line represents any one of a plurality of typical imaging lines.

[0108] A picking unit for interactively picking Fresnel zones of a plurality of typical imaging points based on the multi - azimuth dip - domain migration gather.

[0109] A migration unit for using the Fresnel zones of a plurality of typical imaging points as a migration aperture to perform pre - stack migration processing on the pre - stack seismic gather of the training work area, and obtaining a high - precision migration imaging profile of the target typical imaging line.

[0110] Optionally, the initial neural network model further includes: a normalization layer and an inverse - normalization layer; the training module includes:

[0111] A normalization unit for respectively performing normalization processing on the multi - azimuth migration imaging profile set of the target typical imaging line and the corresponding high - precision migration imaging profile through the normalization layer, to obtain a multi - azimuth migration imaging profile set with amplitude normalization and a high - precision migration imaging profile with amplitude normalization.

[0112] An encoding unit for encoding each local time - window data in the multi - azimuth migration imaging profile set with amplitude normalization through an encoding module, to extract local features and profile position information of each local time - window data.

[0113] An unfolding unit for unfolding the local features of all local time - window data to construct a long sequence of local features.

[0114] An extraction unit for extracting global features and effective features of the long sequence of local features through a Transformer neural network to obtain a target feature sequence.

[0115] A folding unit for folding the target feature sequence, and a decoding module for decoding the folded features and profile position information to obtain a reconstructed migration imaging profile.

[0116] A calculation unit for calculating a loss function value based on the reconstructed migration imaging profile and the high - precision migration imaging profile with amplitude normalization, to train the initial neural network model based on the loss function value.

[0117] An inverse - normalization unit for performing inverse - normalization processing on the reconstructed migration imaging profile through the inverse - normalization layer based on the multi - azimuth migration imaging profile set of the target typical imaging line and the corresponding high - precision migration imaging profile, to obtain a high - precision migration imaging profile with amplitude recovery.

[0118] Optionally, the normalization unit is specifically used for:

[0119] Determine the amplitude value of each position point in the multi - azimuth migration imaging profile set of the target typical imaging line, and obtain the amplitude value set corresponding to each migration imaging profile.

[0120] Superimpose the amplitude values with the same serial number in all amplitude value sets, and take the absolute value after superimposition to obtain the first target amplitude value set.

[0121] Take the maximum amplitude value in the first target amplitude value set as the normalization parameter of the multi - azimuth migration imaging profile set to perform normalization processing on the multi - azimuth migration imaging profile set.

[0122] Determine the amplitude value of each position point in the high - precision migration imaging profile of the target typical imaging line, and take the absolute value to obtain the second target amplitude value set.

[0123] Take the maximum amplitude value in the second target amplitude value set as the normalization parameter of the high - precision migration imaging profile to perform normalization processing on the high - precision migration imaging profile.

[0124] Optionally, the denormalization unit is specifically used for:

[0125] Determine the amplitude value of each position point in the multi - azimuth migration imaging profile set of the target typical imaging line to determine the energy average value of the multi - azimuth migration imaging profile set based on the amplitude value, and obtain the first energy.

[0126] Determine the amplitude value of each position point in the high - precision migration imaging profile of the target typical imaging line to determine the energy average value of the high - precision migration imaging profile based on the amplitude value, obtain the second energy, and determine the maximum amplitude value in the high - precision migration imaging profile.

[0127] Based on the first energy, the second energy, and the maximum amplitude value, determine the denormalization parameter of the reconstructed migration imaging profile.

[0128] Perform denormalization processing on the reconstructed migration imaging profile based on the denormalization parameter to obtain the high - precision migration imaging profile with restored amplitude.

[0129] Optionally, the selection criterion for the typical imaging line is: at the position with drastic structural changes, select the imaging line based on the first interval, and at the position with gentle structure, select the imaging line based on the second interval, where the first interval is less than the second interval.

[0130] Embodiment III

[0131] See Figure 7, an embodiment of the present invention provides an electronic device, which includes: a processor 60, a memory 61, a bus 62, and a communication interface 63, and the processor 60, the communication interface 63, and the memory 61 are connected through the bus 62; the processor 60 is configured to execute an executable module stored in the memory 61, such as a computer program.

[0132] Among them, the memory 61 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 63 (which can be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. can be used.

[0133] The bus 62 may be an ISA bus, a PCI bus, an EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of convenience of representation, Figure 7 only a bidirectional arrow is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0134] Among them, the memory 61 is used to store a program, and after receiving an execution instruction, the processor 60 executes the program. The method executed by the device defined by the process disclosed in any embodiment of the foregoing embodiments of the present invention can be applied to the processor 60 or implemented by the processor 60.

[0135] The processor 60 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in the processor 60 or the instructions in the form of software. The above-mentioned processor 60 may be a general-purpose processor, including a central processing unit (CPU for short), a network processor (NP for short), etc.; it may also be a digital signal processor (DSP for short), an application specific integrated circuit (ASIC for short), a field-programmable gate array (FPGA for short), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. It can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of the present invention. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the embodiments of the present invention can be directly embodied as being executed and completed by the hardware decoding processor, or executed and completed by a combination of the hardware and software modules in the decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. This storage medium is located in the memory 61, and the processor 60 reads the information in the memory 61 and combines its hardware to complete the steps of the above method.

[0136] A computer program product of a high-precision seismic data imaging method and device provided by an embodiment of the present invention includes a computer-readable storage medium storing non-volatile program code executable by a processor. The instructions included in the program code can be used to execute the method described in the foregoing method embodiments. For specific implementation, reference can be made to the method embodiments and will not be elaborated here.

[0137] In addition, in each embodiment of the present invention, the functional units may be integrated in one processing unit, or each unit may exist physically alone, or two or more units may be integrated in one unit.

[0138] When the above-mentioned function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile computer-readable storage medium executable by a processor. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs that can store program codes.

[0139] It should be noted that: similar reference numerals and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0140] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by terms such as "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship in which the product of this invention is usually placed during use. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention. In addition, terms such as "first", "second", "third", etc. are only used for descriptive distinction and cannot be understood as indicating or implying relative importance.

[0141] In addition, terms such as "horizontal", "vertical", "hanging", etc. do not mean that the components are required to be absolutely horizontal or hanging, but can be slightly inclined. For example, "horizontal" only means that its direction is more horizontal relative to "vertical", and does not mean that this structure must be completely horizontal, but can be slightly inclined.

[0142] In the description of the present invention, it should also be noted that unless otherwise clearly specified and limited, the terms "set", "install", "connect", "couple" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0143] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for high-precision imaging of seismic data, characterized in that: include: Obtain seismic pre-stack gathers of the area to be imaged; For each imaging line in the imaging space of the work area to be imaged, multi-azimuth pre-stack migration processing is performed on the seismic pre-stack gathers with a fixed migration aperture to obtain a multi-azimuth migration imaging section set for each imaging line; The target neural network model is used to process the multi-azimuth offset imaging profile set of the target imaging line to obtain the high-precision offset imaging profile of the target imaging line; wherein the target neural network model is a lightweight neural network composed of a Transformer neural network and a CNN neural network, and the Transformer neural network is located between the encoding module and the decoding module of the CNN neural network; the training goal of the target neural network model is that the error between the output offset imaging profile prediction result and the actual high-precision offset imaging profile is less than a preset threshold; the target imaging line represents any imaging line in the imaging space; the actual high-precision offset imaging profile is an imaging result obtained by determining the Fresnel zone of the imaging point based on the multi-azimuth dip domain offset gather and using it as the offset aperture for offset calculation; the number of azimuths included in the multi-azimuth is greater than 2; Based on the high-precision offset imaging sections of all imaging lines, a high-precision imaging result of the work area to be imaged is determined.

2. The high-precision imaging method for seismic data according to claim 1, characterized in that: Also includes: Obtain the seismic pre-stack gathers in the training area; Determine a high-precision migration imaging section of a plurality of typical imaging lines and a corresponding multi-azimuth migration imaging section set based on the seismic pre-stack gathers of the training work area; wherein the high-precision migration imaging section is an imaging section obtained by performing migration processing by using the Fresnel zone of the imaging point as the migration aperture; The multi-azimuth offset imaging section set of each typical imaging line is used as input data of the initial neural network model, and the corresponding high-precision offset imaging section is used as a training label to train the initial neural network model until the training goal is achieved, thereby obtaining the target neural network model.

3. The high-precision seismic data imaging method according to claim 2, characterized in that: Determine a high-precision migration imaging profile of multiple typical imaging lines based on the seismic pre-stack gathers in the training area, including: Calculating a multi-azimuth dip domain migration gather of a target typical imaging line based on the seismic pre-stack gathers of the training work area; wherein the target typical imaging line represents any one of the multiple typical imaging lines; Based on the multi-azimuth dip-domain migration gathers, human-computer interaction is used to pick up Fresnel zones of multiple typical imaging points; The Fresnel zones of the multiple typical imaging points are used as migration apertures, and pre-stack migration processing is performed on the seismic pre-stack gathers in the training area to obtain a high-precision migration imaging profile of the target typical imaging line.

4. The method for high-precision imaging of seismic data according to claim 3, characterized in that: The initial neural network model further includes: a normalization layer and a denormalization layer; taking the multi-azimuth migration imaging section set of each typical imaging line as input data of the initial neural network model, and taking the corresponding high-precision migration imaging section as a training label, training the initial neural network model, including: The normalization layer normalizes the multi-azimuth migration imaging section set and the corresponding high-precision migration imaging section of the target typical imaging line, respectively, to obtain the amplitude-normalized multi-azimuth migration imaging section set and the amplitude-normalized high-precision migration imaging section; The encoding module encodes each local time window data in the amplitude-normalized multi-azimuth migration imaging section set to extract local features and section position information of each local time window data; Expanding the local features of all the local time window data to construct a long sequence of local features; The Transformer neural network extracts global features and effective features of the long sequence of local features to obtain a target feature sequence; The target feature sequence is feature folded, and the decoding module decodes the folded features and the section position information to obtain a reconstructed offset imaging section; Calculating a loss function value based on the reconstructed migration imaging section and the amplitude-normalized high-precision migration imaging section, so as to train the initial neural network model based on the loss function value; The denormalization layer performs denormalization processing on the reconstructed offset imaging section based on the multi-azimuth offset imaging section set of the target typical imaging line and the corresponding high-precision offset imaging section to obtain the high-precision offset imaging section after amplitude restoration.

5. The method for high-precision imaging of seismic data according to claim 4, characterized in that: The normalization layer performs normalization processing on the multi-azimuth migration imaging section set of the target typical imaging line and the corresponding high-precision migration imaging section, respectively, including: Determine the amplitude value of each position point in the multi-azimuth offset imaging section set of the typical imaging line of the target, and obtain the amplitude value set corresponding to each offset imaging section; The amplitude values ​​of the same sequence number in all the amplitude value sets are superimposed, and the absolute value is taken after superposition to obtain a first target amplitude value set; using the maximum amplitude value in the first target amplitude value set as a normalization parameter of the multi-azimuth migration imaging section set, so as to perform normalization processing on the multi-azimuth migration imaging section set; Determine the amplitude value of each position point in the high-precision offset imaging section of the target typical imaging line, and take the absolute value to obtain a second target amplitude value set; The maximum amplitude value in the second target amplitude value set is used as a normalization parameter of the high-precision migration imaging section to perform normalization processing on the high-precision migration imaging section.

6. The method for high-precision imaging of seismic data according to claim 4, characterized in that: The denormalization layer performs denormalization processing on the reconstructed migration imaging section based on the multi-azimuth migration imaging section set of the target typical imaging line and the corresponding high-precision migration imaging section, including: Determine the amplitude value of each position point in the multi-azimuth offset imaging section set of the target typical imaging line, and determine the energy average value of the multi-azimuth offset imaging section set based on the amplitude value to obtain a first energy; Determine the amplitude value of each position point in the high-precision offset imaging section of the target typical imaging line, determine the energy average value of the high-precision offset imaging section based on the amplitude value to obtain a second energy, and determine the maximum amplitude value in the high-precision offset imaging section; Determining an inverse normalization parameter of the reconstructed migration imaging section based on the first energy, the second energy and the maximum amplitude; The reconstructed migration imaging section is subjected to a denormalization process based on the denormalization parameter to obtain a high-precision migration imaging section after amplitude restoration.

7. The method for high-precision imaging of seismic data according to claim 2, characterized in that: The selection criteria of the typical imaging lines are: in a location with a drastic structural change, the imaging lines are selected based on a first interval, and in a location with a gentle structural change, the imaging lines are selected based on a second interval, and the first interval is smaller than the second interval.

8. A high-precision imaging device for seismic data, characterized in that: The first acquisition module is used to acquire the seismic pre-stack gathers of the work area to be imaged; A migration module, for performing multi-azimuth pre-stack migration processing on the seismic pre-stack gathers with a fixed migration aperture for each imaging line in the imaging space of the work area to be imaged, so as to obtain a multi-azimuth migration imaging section set for each imaging line; A processing module, used for processing a multi-azimuth offset imaging profile set of a target imaging line using a target neural network model to obtain a high-precision offset imaging profile of the target imaging line; wherein the target neural network model is a lightweight neural network based on a Transformer neural network and a CNN neural network, and the Transformer neural network is located between an encoding module and a decoding module of the CNN neural network; a training target of the target neural network model is that an error between an output offset imaging profile prediction result and an actual high-precision offset imaging profile is less than a preset threshold; the target imaging line represents any imaging line in the imaging space; the actual high-precision offset imaging profile is an imaging result obtained by determining the Fresnel zone of an imaging point based on a multi-azimuth dip domain offset gather and using it as an offset aperture for offset calculation; the number of azimuths included in the multi-azimuths is greater than 2; The first determination module is used to determine the high-precision imaging result of the work area to be imaged based on the high-precision offset imaging profiles of all imaging lines.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, wherein: When the processor executes the computer program, the method for high-precision imaging of seismic data according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the method for high-precision imaging of seismic data according to any one of claims 1 to 7 is implemented.

Images