CN113722893B - Seismic record inversion method, device, equipment and storage medium - Google Patents
Seismic record inversion method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN113722893B CN113722893B CN202110925389.0A CN202110925389A CN113722893B CN 113722893 B CN113722893 B CN 113722893B CN 202110925389 A CN202110925389 A CN 202110925389A CN 113722893 B CN113722893 B CN 113722893B
- Authority
- CN
- China
- Prior art keywords
- inversion
- seismic record
- loss function
- generator
- wgan
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 78
- 230000006870 function Effects 0.000 claims abstract description 140
- 230000009977 dual effect Effects 0.000 claims abstract description 59
- 238000003062 neural network model Methods 0.000 claims abstract description 10
- 125000004122 cyclic group Chemical group 0.000 claims abstract description 8
- 238000012549 training Methods 0.000 claims description 15
- 238000005457 optimization Methods 0.000 claims description 13
- 238000010276 construction Methods 0.000 claims description 2
- 238000010586 diagram Methods 0.000 description 17
- 230000008569 process Effects 0.000 description 9
- 238000013528 artificial neural network Methods 0.000 description 8
- 238000004590 computer program Methods 0.000 description 7
- 230000004913 activation Effects 0.000 description 5
- 238000013135 deep learning Methods 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000002372 labelling Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 238000011176 pooling Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000008034 disappearance Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000013016 damping Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000000802 evaporation-induced self-assembly Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000002194 synthesizing effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
The embodiment of the invention provides a seismic record inversion method, a seismic record inversion device, seismic record inversion equipment and a storage medium, wherein the method comprises the following steps: the method comprises the steps of obtaining a small sample seismic record, a corresponding small sample wave impedance and a large sample label-free seismic record, constructing a dual WGAN model based on a neural network model, wherein the dual WGAN model comprises inversion WGAN and forward-modeling WGAN, optimizing the dual WGAN model through the small sample seismic record, the small sample wave impedance and the large sample label-free seismic record, inputting the large sample label-free seismic record into an optimized inversion generator to generate a corresponding large sample wave impedance, constructing the dual WGAN model, realizing high-precision inversion of the small sample seismic record based on cyclic consistency of dual learning, and replacing a cross entropy loss function in the traditional GAN by using a Wasserstein distance, so that the stability of seismic record inversion is improved.
Description
Technical Field
The present invention relates to the field of energy development and exploration, and in particular, to a seismic record inversion method, apparatus, device, and storage medium.
Background
Wave impedance inversion is a post-stack inversion technique that estimates elastic parameters of an underground medium. According to a convolution formula, the amplitude of the same reflection point of the post-stack seismic trace set is related to the elastic properties of the medium on the reflection interface and the medium on the reflection interface, inversion of the post-stack seismic trace is based on the relation, and an inversion method is used for calculating elastic parameters, namely wave impedance, of the underground medium according to post-stack seismic data.
The existing inversion method is a new method for directly establishing a velocity model from an original seismic record based on a supervised depth full convolution neural network by applying the convolution neural network to wave impedance inversion. During the training phase, the network builds a nonlinear projection from multi-shot seismic data to the corresponding velocity model. In the prediction phase, the trained network may be used to estimate a new velocity model of the incoming seismic data.
However, the method is difficult to train and optimize a large number of unlabeled seismic records to reach the required accuracy of inversion when facing a large number of seismic records and a small number of corresponding wave impedances.
Disclosure of Invention
The embodiment of the invention provides a seismic record inversion method, a device, equipment and a storage medium, which are used for improving inversion accuracy of a large number of unlabeled seismic records under a small sample condition.
In a first aspect, an embodiment of the present invention provides a seismic record inversion method, including:
acquiring a small sample seismic record, small sample wave impedance corresponding to the small sample seismic record and a large sample label-free seismic record;
constructing a dual WGAN model based on a neural network model, wherein the dual WGAN model comprises an inversion WGAN and a forward WGAN, the inversion WGAN comprises an inversion generator and an inversion discriminator, and the forward WGAN comprises a forward generator and a forward discriminator;
optimizing the dual WGAN model by the small sample seismic record, the small sample wave impedance, and the large sample unlabeled seismic record;
and inputting the large sample unlabeled seismic records into an optimized inversion generator, and generating corresponding large sample wave impedance.
Optionally, the inversion generator is a one-dimensional U-net generator which is input into two channels under the constraint of the condition information, and the inversion discriminator is a one-dimensional Markovian discriminator which is input into two channels under the constraint of the condition information;
the forward generator is a one-dimensional U-net generator, and the forward arbiter is a one-dimensional AlexNet arbiter.
Optionally, optimizing the dual WGAN model by the small sample seismic record, the small sample wave impedance, and the large sample unlabeled seismic record includes:
Determining an inversion discriminant loss function, a forward discriminant loss function, and a generator loss function from the small sample seismic record, the small sample wave impedance, and the large sample unlabeled seismic record;
optimizing the inversion discriminator, the forward decision maker and the generator according to the inversion discriminator loss function, the forward decision maker loss function and the generator loss function;
wherein the generator comprises the forward generator and the inversion generator.
Optionally, determining an inversion discriminator loss function, a forward discriminator loss function, and a generator loss function from the small sample seismic record, the small sample wave impedance, and the large sample unlabeled seismic record, comprising:
inputting a low-frequency model, the small sample wave impedance and a seismic record into the inversion generator to obtain a wave impedance predicted value, wherein the seismic record comprises a small sample seismic record and a large sample label-free seismic record;
inputting the wave impedance predicted value and the small sample wave impedance into the inversion discriminator to obtain the inversion discriminator loss function;
Inputting the wave impedance predicted value into the forward generator to obtain a reconstructed seismic record;
inputting the small sample wave impedance into the forward generator to obtain a seismic record predicted value;
inputting the seismic record predicted value and the seismic record into the forward model arbiter to obtain the forward model arbiter loss function;
inputting the seismic record predicted value into the inversion generator to obtain reconstructed wave impedance;
and determining the generator loss function according to the wave impedance predicted value, the seismic record predicted value, the reconstructed seismic record and the reconstructed wave impedance.
Optionally, the inversion discriminator loss function is:
wherein D is inver Inversion discriminant for inverting WGAN, G inver For inversion generator of inversion WGAN, AI is wavelet impedance of small sample, S is small sample seismic record corresponding to AI, (A|B) is expressed as A input network under B condition constraint, gp 1 As gradient penalty term, λ is gradient penaltyThe term coefficient, m, is the number of small samples for training;
the forward arbiter loss function is:
wherein D is forward Inversion discriminant for inverting WGAN, G forward Inversion generator for inverting WGAN, gp 2 Is a gradient penalty term.
Optionally, the generator loss function includes generating a loss function for the challenge portion and a loss function for the loop consistency portion; the generating the loss function of the countermeasure section includes generating the loss function of the countermeasure inversion section and generating the loss function of the countermeasure forward section; the loss function of the loop consistency part comprises the loss function of the loop consistency open-loop part and the loss function of the loop consistency closed-loop part;
the generating a loss function against the inversion portion is:
wherein S is * AI for large sample unlabeled seismic recording low Is a wave impedance low-frequency model;
the generating a loss function against the forward part is:
the loss function of the loop-consistent open-loop part is:
the loss function of the cyclic consistency closed loop portion is:
the generator loss function is:
g_loss＝g_inver_loss+g_forward_loss+α×(loss_3+loss_4+loss_5)+β×(loss_1+loss_2)
wherein, alpha and beta are constraint coefficients respectively controlling the contribution of an open loop part and a closed loop part, and M is the number of large samples for training.
In a second aspect, an embodiment of the present invention provides a seismic record inversion apparatus, the apparatus comprising:
the acquisition module is used for acquiring a small sample seismic record, small sample wave impedance corresponding to the small sample seismic record and a large sample label-free seismic record;
The system comprises a building module, a determining module and a determining module, wherein the building module is used for building a dual WGAN model based on a neural network model, the dual WGAN network model comprises an inversion WGAN and a forward WGAN, the inversion WGAN comprises an inversion generator and an inversion discriminator, and the forward WGAN comprises a forward generator and a forward discriminator;
an optimization module for optimizing the dual WGAN model by the small sample seismic record, the small sample wave impedance, and the large sample unlabeled seismic record;
the generation module is used for inputting the large sample unlabeled seismic records into an optimized inversion generator and generating corresponding large sample wave impedance.
In a third aspect, an embodiment of the present invention provides a seismic record inversion apparatus, including:
a memory and at least one processor;
the memory stores computer-executable instructions;
the at least one processor executing computer-executable instructions stored in the memory causes the at least one processor to perform the seismic record inversion method of any of the first aspect above.
In a fourth aspect, embodiments of the present invention provide a computer-readable storage medium having stored therein computer-executable instructions that, when executed by a processor, are configured to implement the seismic record inversion method according to any one of the first aspect above.
In a fifth aspect, embodiments of the present invention provide a computer program product comprising a computer program which when executed by a processor implements the seismic record inversion method of any of the first aspects above.
According to the seismic record inversion method, the device, the equipment and the storage medium, the dual WGAN model is built based on the neural network model by acquiring the small sample seismic record, the small sample wave impedance corresponding to the small sample seismic record and the large sample label-free seismic record, the dual WGAN model comprises inversion WGAN and forward WGAN, wherein the inversion WGAN comprises an inversion generator and an inversion discriminator, the forward WGAN comprises the forward generator and the forward discriminator, the dual WGAN model is optimized through the small sample seismic record, the small sample wave impedance and the large sample label-free seismic record, the large sample label-free seismic record is input into the optimized inversion generator, the corresponding large sample wave impedance is generated, the dual WGAN model is built, the high-precision inversion of the small sample seismic record can be realized based on the cyclic consistency of dual learning, the problem of gradient disappearance and collapse mode in the conventional GAN inversion process is avoided by using the Wassess distance to replace the cross entropy loss function, and the stability of the inversion seismic record is improved.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.
Fig. 1 is a schematic view of an application scenario provided in an embodiment of the present invention;
FIG. 2 is a schematic flow chart of a seismic record inversion method according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a network structure of an inversion generator according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of a network structure of an inversion discriminator according to the embodiment of the invention;
fig. 5 is a schematic diagram of a network structure of a forward generator according to an embodiment of the present invention;
fig. 6 is a schematic diagram of a network structure of a forward performance arbiter according to an embodiment of the present invention;
FIG. 7 is a schematic diagram of a framework for a dual WGAN model provided in accordance with an embodiment of the invention;
FIG. 8 is a flow chart of another seismic record inversion method according to an embodiment of the invention;
FIG. 9 is a schematic diagram of a Marmousi2 wave impedance model according to an embodiment of the invention;
FIG. 10 is a schematic representation of wavelets for a synthetic seismic record according to an embodiment of the invention;
FIG. 11 is a cross-sectional view of a synthetic seismic record provided in an embodiment of the invention;
FIG. 12 is a graph comparing the results of a single pass inversion with actual results provided by the practice of the present invention;
FIG. 13A is a two-dimensional cross-sectional view of a real impedance value according to an embodiment of the present invention;
FIG. 13B is a two-dimensional cross-sectional view of a low frequency model according to an embodiment of the present invention;
FIG. 13C is a two-dimensional cross-sectional view of a dual WGAN inversion result in accordance with an embodiment of the invention;
FIG. 14 is a schematic structural diagram of an apparatus for inversion of seismic records according to an embodiment of the present invention;
fig. 15 is a schematic structural diagram of an earthquake record inversion apparatus according to an embodiment of the present invention.
Specific embodiments of the present invention have been shown by way of the above drawings and will be described in more detail below. The drawings and the written description are not intended to limit the scope of the inventive concepts in any way, but rather to illustrate the inventive concepts to those skilled in the art by reference to the specific embodiments.
Detailed Description
Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numbers in different drawings refer to the same or similar elements, unless otherwise indicated. The implementations described in the following exemplary examples do not represent all implementations consistent with the invention. Rather, they are merely examples of apparatus and methods consistent with aspects of the invention as detailed in the accompanying claims.
The following describes the technical scheme of the present invention and how the technical scheme of the present invention solves the above technical problems in detail with specific embodiments. The following embodiments may be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.
An application scenario provided by the embodiment of the present invention is explained below: the scheme provided by the embodiment of the invention relates to inversion of a large sample label-free seismic record to obtain corresponding wave impedance. In practical seismic exploration inversion, the problems of more seismic records and less corresponding wave impedance exist, so that inversion of large-sample unlabeled seismic records is needed to be realized through small-sample seismic records, and the wave impedance corresponding to the large-sample unlabeled seismic records is obtained.
In some technologies, linear approximation is performed on the branching mapping from the seismic record to the wave impedance based on a generalized linear inversion method, so that generalized linear inversion of the post-stack seismic record is realized through the linear inversion method, but the method may have the problem of poor inversion accuracy.
In other techniques, a self-adaptive damping term algorithm can be provided according to strong reflection coefficient sparse characteristics through two different regularization criteria, so that constrained sparse pulse inversion of wave impedance is realized. However, the bandwidth of the seismic recordings recorded according to the convolution model is often limited, and the bandwidth of the inversion target is relatively wide, resulting in underdetermined seismic inversion problems.
In other techniques, a broadband seismic inversion method may be implemented by sparse constraint methods as well constraint methods. And the inversion method based on geostatistics proposes to take the seismic parameters as random variables and perform seismic data inversion based on a Bayesian framework, so as to obtain the optimal solution of statistics. In the method, all uncertainties can be described by using multidimensional Gaussian probability density, environmental noise, discretization errors, theoretical errors and priori information of an underground medium model are taken into calculation information, inversion of earth surface seismic data is achieved through a Bayesian framework, and the method possibly has the problem of low inversion precision.
Among other techniques, since the development of the deep learning technique has been successfully applied to various fields, the deep learning-based inversion method has received a lot of attention and research in recent years. The method can be based on a new method for directly establishing a velocity model from an original seismic record by using a supervised deep full convolution neural network, in a training stage, the network establishes a nonlinear projection to multi-shot seismic data to a corresponding velocity model, and in a prediction stage, the trained network can be used for estimating the velocity model of new input seismic data, but the method has no clear physical meaning.
In other technologies, aiming at a small sample data learning method in seismic inversion, a method based on closed-loop convolution neural network wave impedance inversion under the constraint of logging data is provided, wherein backwave-CNN and forward-CNN in the closed-loop neural network are forward methods for realizing seismic inversion from seismic record to wave impedance and seismic record from wave impedance to seismic record respectively, but the required accuracy of inversion is still difficult to achieve.
Therefore, the embodiment of the invention provides a seismic record inversion method, which constructs a dual WGAN (Wasserstein GAN (Generative Adversarial Networks, generating an countermeasure network)) model based on a neural network model, wherein the method comprises inverting the WGAN and forward-modeling the WGAN, inputting the obtained small sample seismic record, the corresponding small sample wave impedance and the large sample seismic record into the network model for optimization training, inputting the large sample untagged record into an optimized inversion generator to generate the corresponding large sample wave impedance, and improving inversion accuracy of the large sample untagged seismic record under the condition of the small sample seismic record by utilizing a dual learning idea in deep learning through the dual WGAN model.
Fig. 1 is a schematic diagram of an application scenario provided in an embodiment of the present invention. As shown in fig. 1, the small sample wave impedance, small sample seismic record, and large sample unlabeled seismic record are input into the dual WGAN model for optimal training of the model.
Some embodiments of the present invention are described in detail below with reference to the accompanying drawings. In the case where there is no conflict between the embodiments, the following embodiments and features in the embodiments may be combined with each other.
Fig. 2 is a schematic flow chart of a seismic record inversion method according to an embodiment of the invention. As shown in fig. 2, the method in this embodiment may include:
The seismic record may also be referred to as post-stack seismic record, and the small sample wave impedance may be referred to as log data, or as subsurface medium elasticity parameters. The small sample wave impedance and post-stack seismic recording are corresponding. A large sample unlabeled seismic record may be a large number of unlabeled datasets.
Where small samples refer to a relatively small number. Large samples refer to a large number. An unlabeled seismic record refers to a seismic record without a corresponding wave impedance.
The seismic records may be two-dimensional data and the wavelet impedance may be corresponding one-dimensional data in the two-dimensional small sample seismic records.
Alternatively, the seismic record may be approximated by a convolution model, specifically by calculating the reflection coefficient from a known impedance model:
Wherein Z is 2 Is the impedance of the lower layer wave, Z 1 Is the upper layer wave impedance and R is the reflection coefficient. Then, according to the convolution formula, the seismic record is obtained:
S(t)＝W(t)*R(t)
where S is the seismic record, R is the reflection coefficient, and W is the wavelet.
Alternatively, the small sample wave impedance may be an approximate wave impedance corresponding to the small sample seismic record.
Wherein the Dual WGAN model may also be referred to as the Dual-WGAN-c model. The dual WGAN model consists of two WGANs, the inversion WGANs realize the inversion process from the seismic record to the wave impedance, and the inversion process from the wave impedance to the seismic record is carried out in forward WGAN experiment.
The inversion generator obtains a corresponding wave impedance predicted value according to the seismic record, the forward generator obtains a corresponding seismic record predicted value according to the wave impedance, the inversion discriminator is used for discriminating the performance of the inversion generator, and the forward discriminator is used for discriminating the performance of the forward generator.
Step 203, optimizing the dual WGAN model through the small sample seismic record, the small sample wave impedance, and the large sample unlabeled seismic record.
Optionally, after the dual WGAN model is constructed, an inversion discriminator, an inversion generator, a forward wave discriminator and a forward wave generator in the dual WGAN model are optimized through a small sample seismic record, a small sample wave impedance and a large sample label-free seismic record, so that the dual WGAN model is optimized.
And 204, inputting the large sample unlabeled seismic records into an optimized inversion generator, and generating corresponding large sample wave impedance.
Optionally, after the dual WGAN model is optimized, inputting the large sample unlabeled seismic record into the optimized inversion generator to generate a corresponding large sample wave impedance.
According to the seismic record inversion method provided by the embodiment, a small sample seismic record, a small sample wave impedance corresponding to the small sample seismic record and a large sample label-free seismic record are obtained, a dual WGAN model is constructed based on a neural network model, the dual WGAN model comprises inversion WGAN and forward WGAN, wherein the inversion WGAN comprises an inversion generator and an inversion discriminator, the forward WGAN comprises the forward generator and the forward discriminator, the dual WGAN model is optimized through the small sample seismic record, the small sample wave impedance and the large sample label-free seismic record, the large sample label-free seismic record is input into the optimized inversion generator, the corresponding large sample wave impedance is generated, the dual WGAN model is constructed, high-precision inversion of the small sample seismic record can be realized based on cyclic consistency of dual learning, the problem of gradient disappearance and mode collapse in the traditional GAN inversion process is avoided by using the Wasserstein distance to replace the cross entropy loss function in the traditional GAN inversion process, and the stability of inversion of the seismic record is improved.
On the basis of the technical scheme provided by the embodiment, optionally, the inversion generator is a one-dimensional U-net generator which is input into two channels under the constraint of the condition information, and the inversion discriminator is a one-dimensional Markovian discriminator which is input into two channels under the constraint of the condition information; the forward generator is a one-dimensional U-net generator, and the forward arbiter is a one-dimensional AlexNet arbiter.
Fig. 3 is a schematic diagram of a network structure of an inversion generator according to an embodiment of the present invention. As shown in fig. 3, the inversion generator is a one-dimensional U-net generator that is input as a two-channel (channel=2) under the constraint of the condition information. Specifically, a pair of small sample seismic records and small sample wave impedance, and a pair of large sample unlabeled seismic records and wave impedance low-frequency models can be input, and two pairs of data can be circularly input in the input process. Wherein the rectangle marked with the number 2 in the figure is denoted as a double channel, the rest being similar. The arrow to the right indicates the convolution + BN (Batch Normalization) + activation function, the arrow to the bottom indicates the pooling layer, the arrow to the top indicates the transposed convolution, a in [ a,1] indicates the vector length of the input data, 1 indicates the dimension, a/2 indicates the vector length of the input data is 1/2 of the original, and the rest are similar.
Fig. 4 is a schematic diagram of a network structure of an inversion discriminator according to the embodiment of the invention. As shown in fig. 4, the inversion discriminator is a one-dimensional Markovian discriminator which is input as a two-channel under the constraint of condition information. In the figure, black arrows represent convolution, dashed arrows represent activation function +convolution +BN, a in [ a,1] represents vector length of input data, 1 represents dimension, a/2 represents vector length of input data as original 1/2, and the rest are similar.
Fig. 5 is a schematic diagram of a network structure of a forward generator according to an embodiment of the present invention. As shown in fig. 5, the forward generator is a one-dimensional U-net generator. In particular, the input may be a small sample wave impedance. Wherein the rectangle marked with the number 1 in the figure is shown as a single channel, the remainder being similar. The arrow to the right represents the convolution +bn+ activation function, the arrow to the bottom represents the pooling layer, the arrow to the top represents the transposed convolution, [ a,1] a represents the vector length of the input data, 1 represents the dimension, a/2 represents the vector length of the input data as originally 1/2, and the rest are similar.
Fig. 6 is a schematic diagram of a network structure of a forward arbiter according to an embodiment of the present invention. As shown in fig. 6, the forward arbiter is a one-dimensional AlexNet arbiter. Wherein, the initial dotted arrow represents convolution, the middle solid arrow represents activation function +convolution +BN, the last two arrows represent full connection, a in [ a,1] represents vector length of input data, 1 represents dimension, a/2 represents vector length of input data is 1/2 of original, and the rest are similar.
In the neural network architecture, a convolution layer realizes weight sharing and feature extraction, a pooling layer realizes downsampling, an activation function realizes algorithm nonlinearity, and a BN layer prevents training from being fitted.
The low-frequency information constraint is added through the Conditional GAN in the deep learning inversion process, and the black box type neural network training method is endowed with geophysical meaning, namely, the numerical values obtained in the model training process have certain physical meaning, and are not only simple numerical values.
Optionally, optimizing the dual WGAN model by the small sample seismic record, the small sample wave impedance, and the large sample unlabeled seismic record includes:
determining an inversion discriminant loss function, a forward discriminant loss function, and a generator loss function from the small sample seismic record, the small sample wave impedance, and the large sample unlabeled seismic record; optimizing the inversion discriminator, the forward decision maker and the generator according to the inversion discriminator loss function, the forward decision maker loss function and the generator loss function; wherein the generator comprises the forward generator and the inversion generator.
Optionally, the small sample seismic record, the small sample wave impedance and the large sample label-free seismic record are input into the dual WGAN model, an inversion discrimination loss function, a forward discrimination loss function and a generator loss function can be obtained, when the value of the loss function of the inversion discriminator is minimum, the optimization of the inversion discriminator can be realized, when the value of the loss function of the forward discriminator is minimum, the optimization of the forward discriminator can be realized, and when the value of the loss function of the generator is minimum, the optimization of the generator can be realized.
In this embodiment, the generator and the arbiter in the dual WGAN model are optimized to optimize the network model, so that the network optimization efficiency can be improved.
Fig. 7 is a schematic diagram of a framework of a dual WGAN model according to an embodiment of the present invention, as shown in fig. 7, determining an inversion discriminator loss function, a forward discriminator loss function, and a generator loss function according to the small sample seismic record, the small sample wave impedance, and the large sample unlabeled seismic record, including:
inputting a low-frequency model, the small sample wave impedance and a seismic record into the inversion generator to obtain a wave impedance predicted value, wherein the seismic record comprises a small sample seismic record and a large sample label-free seismic record; inputting the wave impedance predicted value and the small sample wave impedance into the inversion discriminator to obtain an inversion discriminator loss function; inputting the wave impedance predicted value into the forward generator to obtain a reconstructed seismic record; inputting the small sample wave impedance into the forward generator to obtain a seismic record predicted value; inputting the seismic record predicted value and the seismic record into the forward arbiter to obtain a forward arbiter loss function; inputting the seismic record predicted value into the inversion generator to obtain reconstructed wave impedance; and determining the generator loss function according to the wave impedance predicted value, the seismic record predicted value, the reconstructed seismic record and the reconstructed wave impedance.
The wave impedance predicted values comprise a small sample wave impedance predicted value and a large sample wave impedance predicted value, and the reconstructed seismic records comprise a small sample reconstructed seismic record and a large sample reconstructed seismic record.
Optionally, a pair of small sample seismic records and corresponding small sample wave impedances are input into an inversion generator to obtain a small sample wave impedance predicted value, and a pair of large sample seismic records and corresponding wave impedance low-frequency models are input into the inversion generator to obtain a large sample wave impedance predicted value. And inputting the wavelet impedance predicted value and the wave impedance of the small sample into the inversion discriminator to obtain the loss function of the inversion discriminator.
And inputting the small sample wave impedance predicted value into a forward generator to obtain a small sample reconstruction seismic record, and inputting the large sample wave impedance predicted value into the forward generator to obtain a large sample reconstruction seismic record.
And inputting the small sample wave impedance into a forward generator to obtain a small sample seismic record predicted value. And inputting the earthquake record predicted value and the small sample earthquake record into a forward discriminator to obtain a forward discriminator loss function. And inputting the seismic record predicted value into an inversion generator to obtain the reconstructed wave impedance.
From the wave impedance predictions, seismic record predictions, reconstructed seismic records, and reconstructed wave impedances obtained above, a generator loss function can be determined.
In the embodiment, based on the cyclic consistency of dual learning, the problem of low inversion precision caused by insufficient sample data in seismic exploration is avoided, and the inversion precision of small sample seismic records is realized.
Optionally, the inversion discriminator loss function is:
wherein D is inver Inversion discriminant for inverting WGAN, G inver For inversion generator of inversion WGAN, AI is wavelet impedance of small sample, S is seismic record of small sample corresponding to AI, (A|B) represents A input network under constraint of B condition, gp 1 And lambda is a gradient penalty term coefficient, and m is the number of small samples for training.
Where m is specifically expressed as the number of small sample seismic records that are trained. D (D) inver (AI|AI) is expressed as AI input network under AI condition constraint, and the real wavelet impedance AI, D is obtained by inversion of the action of the discriminator inver (G inver (S|AI) |AI) is represented as an S input network under AI condition constraints, generating the corresponding G by the action of an inversion generator inver (S|AI) and then putting G under AI condition constraint inver (S|AI) input network, the wavelet impedance predictive value is generated through inversion discriminator action.
Optionally, a minimum value between the real wavelet impedance and the generated wavelet impedance predicted value is calculated in the inversion discriminator, so that the loss function of the inversion discriminator is optimized, and the inversion discriminator is optimized.
The forward arbiter loss function is:
wherein D is forward Inversion discriminant for inverting WGAN, G forward Inversion generator for inverting WGAN, gp 2 Is a gradient penalty term.
Wherein S is a real small sample seismic record, and AI is a label corresponding to the small sample seismic record.
Optionally, a minimum value between the real small sample seismic record and the generated small sample seismic record predicted value is calculated in the inversion discriminator, so that the loss function of the forward discriminator is optimized, and the forward discriminator is optimized.
In this embodiment, the minimum value between the actual small sample wave impedance and the generated small sample wave impedance predicted value is calculated, and the minimum value between the actual small sample seismic record and the generated small sample seismic record predicted value is calculated, so that the inversion discriminator and the forward discriminator are optimized, and the accuracy of the generated wave impedance can be improved.
Optionally, the generator loss function includes generating a loss function for the challenge portion and a loss function for the loop consistency portion; the generating the loss function of the countermeasure section includes generating the loss function of the countermeasure inversion section and generating the loss function of the countermeasure forward section; the loss function of the loop consistency part comprises the loss function of the loop consistency open-loop part and the loss function of the loop consistency closed-loop part;
The generating a loss function against the inversion portion is:
wherein S is * AI for large sample unlabeled seismic recording low For waves obtained by logging dataAn impedance low frequency model;
the generating a loss function against the forward part is:
the loss function of the loop-consistent open-loop part is:
the loss function of the cyclic consistency closed loop portion is:
the generator loss function is:
g_loss＝g_inver_loss+g_forward_loss+α×(loss_3+loss_4+loss_5)+β×(loss_1+loss_2)
wherein α and β are constraint coefficients respectively controlling contributions of the open loop portion and the closed loop portion, M is the number of large samples to be trained, and β > α > 1, in the embodiment α=10, β=1000, and M is specifically indicated as the number of large sample unlabeled seismic records to be trained.
Alternatively, the generator loss function may be determined from the wave impedance predictions, seismic record predictions, reconstructed seismic records, and reconstructed wave impedance obtained in FIG. 7.
In this embodiment, the minimum value is obtained through the generator loss function, so that the generator can be optimized under the condition of the embodiment, and thus the network model can be optimized.
FIG. 8 is a flow chart of another seismic record inversion method according to an embodiment of the invention. Based on the foregoing embodiment, the dual WGAN model is optimized by specifically combining a hyper-parameter and an optimization method, as shown in fig. 8, where the method includes:
The seismic records comprise small sample seismic records and large sample unlabeled seismic records.
Optionally, labeling the small sample seismic record and the small sample wave impedance to obtain a small sample seismic record and a small sample wave impedance label data pair, and labeling the large sample non-label seismic record without label data.
Alternatively, a dual WGAN model may be constructed based on a deep neural network construction method.
Optionally, after the network model is built, a proper loss function is selected to calculate the error, and when the error reaches the minimum value, the obtained network model is the optimal network model.
Optionally, the loss function is subjected to super-parameter adjustment, so that an optimal loss function value can be obtained through calculation. The adjusted super-parameters are iteration number (epoch) =300, learning rate (learning rate) =0.001, and batch size (batch size) =33.
Optionally, an optimization algorithm (Adam) is selected through the determined super parameters, the loss function is optimized, and the minimum value of the loss function is determined, so that the aim of training the dual WGAN model is fulfilled.
And 805, inverting through the optimized network.
Optionally, the large sample label-free seismic record is input into a trained dual WGAN model, and the corresponding large sample wave impedance is obtained through seismic record inversion.
In the embodiment, the dual WGAN model is optimized by combining the super-parameters and the optimization method, so that the optimization of the network model can be improved, and the inversion accuracy of the network model can be improved.
Fig. 9 is a schematic diagram of a Marmousi2 wave impedance model according to an embodiment of the invention. As shown in FIG. 9, the abscissa is x, the left ordinate is time in ms, the right ordinate is impedance value in g/cm 3 *m/s。
FIG. 10 is a schematic representation of wavelets for use in synthesizing a seismic record in accordance with an embodiment of the present invention. As shown in fig. 10, the abscissa is time in ms and the ordinate is amplitude.
FIG. 11 is a cross-sectional view of a synthetic seismic record according to an embodiment of the invention. As shown in fig. 11, the abscissa is x, the ordinate is time, and the unit is ms.
FIG. 12 is a graph comparing the results of a single pass inversion with the actual results provided by the practice of the present invention. As shown in fig. 12, the abscissa is time in ms and the ordinate is impedance value.
Fig. 13A is a two-dimensional cross-sectional view of a real impedance value according to an embodiment of the present invention. As shown in FIG. 13A, the abscissa is x, the left ordinate is time in ms, the right ordinate is impedance value in g/cm 3 *m/s。
Fig. 13B is a two-dimensional cross-sectional view of a low-frequency model according to an embodiment of the present invention. As shown in FIG. 13B, the abscissa is x, the left ordinate is time in ms, the right ordinate is impedance value in g/cm 3 *m/s。
Fig. 13C is a two-dimensional cross-sectional view of a dual WGAN inversion result according to an embodiment of the present invention. As shown in fig. 13C, the abscissa is x, the left ordinate is time in ms, the right ordinate is impedance value, and the left ordinate is singleThe position is g/cm 3 *m/s。
In an actual numerical simulation experiment, the invention can improve inversion accuracy. The Marmousi2 wave impedance model shown in FIG. 9 and the 20Hz Rake wavelet shown in FIG. 10 are used to generate the seismic record shown in FIG. 11 by a convolution formula, and 33 traces are extracted as small sample seismic records and the remaining traces are taken as large sample unlabeled seismic records. Inversion is carried out through the seismic record inversion method, and the inversion result is compared with one-dimensional single channels of a real model and a low-frequency model, so that the inversion result shown in fig. 12 is very close to the real result. Comparing the inversion result of FIG. 13C with the two-dimensional sections of the real model of FIG. 13A and the low-frequency model of FIG. 13B, the inversion result is very close to the real value, and the invention can be seen to have very high precision.
Fig. 14 is a schematic structural diagram of an apparatus for seismic record inversion according to an embodiment of the present invention. As shown in fig. 14, the seismic record inversion apparatus provided in this embodiment may include:
an acquiring module 1401, configured to acquire a small sample seismic record, a small sample wave impedance corresponding to the small sample seismic record, and a large sample label-free seismic record;
a building module 1402, configured to build a dual WGAN model based on a neural network model, the dual WGAN model including an inversion WGAN and a forward WGAN, wherein the inversion WGAN includes an inversion generator and an inversion arbiter, and the forward WGAN includes a forward generator and a forward arbiter;
an optimization module 1403 for optimizing the dual WGAN model with the small sample seismic record, the small sample wave impedance, and the large sample unlabeled seismic record;
a generating module 1404, configured to input the large sample unlabeled seismic record into an optimized inversion generator, and generate a corresponding large sample wave impedance.
Optionally, the inversion generator is a one-dimensional U-net generator which is input into two channels under the constraint of the condition information, and the inversion discriminator is a one-dimensional Markovian discriminator which is input into two channels under the constraint of the condition information;
The forward generator is a one-dimensional U-net generator, and the forward arbiter is a one-dimensional AlexNet arbiter.
Optionally, the optimizing module 1403 is specifically configured to:
determining an inversion discriminant loss function, a forward discriminant loss function, and a generator loss function from the small sample seismic record, the small sample wave impedance, and the large sample unlabeled seismic record;
optimizing the inversion discriminator, the forward decision maker and the generator according to the inversion discriminator loss function, the forward decision maker loss function and the generator loss function;
wherein the generator comprises the forward generator and the inversion generator.
Optionally, the optimizing module 1403 is specifically configured to, when determining an inversion discriminator loss function, a forward discriminator loss function, and a generator loss function from the small sample seismic record, the small sample wave impedance, and the large sample unlabeled seismic record:
inputting a low-frequency model, the small sample wave impedance and a seismic record into the inversion generator to obtain a wave impedance predicted value, wherein the seismic record comprises a small sample seismic record and a large sample label-free seismic record;
Inputting the wave impedance predicted value and the small sample wave impedance into the inversion discriminator to obtain the inversion discriminator loss function;
inputting the wave impedance predicted value into the forward generator to obtain a reconstructed seismic record;
inputting the small sample wave impedance into the forward generator to obtain a seismic record predicted value;
inputting the seismic record predicted value and the seismic record into the forward model arbiter to obtain the forward model arbiter loss function;
inputting the seismic record predicted value into the inversion generator to obtain reconstructed wave impedance;
and determining the generator loss function according to the wave impedance predicted value, the seismic record predicted value, the reconstructed seismic record and the reconstructed wave impedance.
Optionally, the inversion discriminator loss function is:
wherein D is inver Inversion discriminant for inverting WGAN, G inver For inversion generator of inversion WGAN, AI is wavelet impedance of small sample, S is small sample seismic record corresponding to AI, (A|B) is expressed as A input network under B condition constraint, gp 1 Lambda is a gradient penalty term coefficient, and m is the number of small samples for training;
The forward arbiter loss function is:
wherein D is forward Inversion discriminant for inverting WGAN, G forward Inversion generator for inverting WGAN, gp 2 Is a gradient penalty term.
Optionally, the generator loss function includes generating a loss function for the challenge portion and a loss function for the loop consistency portion; the generating the loss function of the countermeasure section includes generating the loss function of the countermeasure inversion section and generating the loss function of the countermeasure forward section; the loss function of the loop consistency part comprises the loss function of the loop consistency open-loop part and the loss function of the loop consistency closed-loop part;
the generating a loss function against the inversion portion is:
wherein S is * For large sample no-label seismic record, AI low Is a wave impedance low-frequency model;
the generating a loss function against the forward part is:
the loss function of the loop-consistent open-loop part is:
the loss function of the cyclic consistency closed loop portion is:
the generator loss function is:
g_loss＝g_inver_loss+g_forward_loss+α×(loss_3+loss_4+loss_5)+β×(loss_1+loss_2)
wherein, alpha and beta are constraint coefficients respectively controlling the contribution of an open loop part and a closed loop part, and M is the number of large samples for training.
The device provided in this embodiment may implement the technical solution of the method embodiment shown in fig. 1 to 13C, and its implementation principle and technical effects are similar, and are not described here again.
Fig. 15 is a schematic structural diagram of an earthquake record inversion apparatus according to an embodiment of the present invention. As shown in fig. 15, the apparatus provided in this embodiment may include: a memory 152 and at least one processor 151;
the memory 152 stores computer-executable instructions;
the at least one processor 151 executes computer-executable instructions stored in the memory 152, such that the at least one processor 151 performs the method described in any of the embodiments above.
Wherein the memory 152 and the processor 151 may be connected by a bus 153.
The specific implementation principle and effect of the device provided in this embodiment may refer to the relevant descriptions and effects corresponding to the embodiments shown in fig. 1 to 13C, which are not repeated here.
The embodiment of the invention also provides a computer readable storage medium, on which a computer program is stored, the computer program being executed by a processor to implement the seismic record inversion method provided by any of the embodiments of the invention.
The embodiment of the invention also provides a computer program product, which comprises a computer program, and the computer program realizes the seismic record inversion method according to any embodiment of the invention when being executed by a processor.
In the several embodiments provided by the present invention, it should be understood that the disclosed apparatus and method may be implemented in other manners. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of modules is merely a logical function division, and there may be additional divisions of actual implementation, e.g., multiple modules or components may be combined or integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or modules, which may be in electrical, mechanical, or other forms.
The modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical units, may be located in one place, or may be distributed over multiple network units. Some or all of the modules may be selected according to actual needs to implement the solution of this embodiment.
In addition, each functional module in the embodiments of the present invention may be integrated in one processing unit, or each module may exist alone physically, or two or more modules may be integrated in one unit. The units formed by the modules can be realized in a form of hardware or a form of hardware and software functional units.
The integrated modules, which are implemented in the form of software functional modules, may be stored in a computer readable storage medium. The software functional modules described above are stored in a storage medium and include instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) or processor to perform some of the steps of the methods described in the various embodiments of the invention.
It should be appreciated that the processor may be a central processing unit (Central Processing Unit, CPU for short), other general purpose processors, digital signal processor (Digital Signal Processor, DSP for short), application specific integrated circuit (Application Specific Integrated Circuit, ASIC for short), etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of a method disclosed in connection with the present invention may be embodied directly in a hardware processor for execution, or in a combination of hardware and software modules in a processor for execution.
The memory may comprise a high-speed RAM memory, and may further comprise a non-volatile memory NVM, such as at least one magnetic disk memory, and may also be a U-disk, a removable hard disk, a read-only memory, a magnetic disk or optical disk, etc.
The bus may be an industry standard architecture (Industry Standard Architecture, ISA) bus, an external device interconnect (Peripheral Component Interconnect, PCI) bus, or an extended industry standard architecture (Extended Industry Standard Architecture, EISA) bus, among others. The buses may be divided into address buses, data buses, control buses, etc. For ease of illustration, the buses in the drawings of the present invention are not limited to only one bus or to one type of bus.
The storage medium may be implemented by any type or combination of volatile or nonvolatile memory devices such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disk. A storage media may be any available media that can be accessed by a general purpose or special purpose computer.
An exemplary storage medium is coupled to the processor such the processor can read information from, and write information to, the storage medium. In the alternative, the storage medium may be integral to the processor. The processor and the storage medium may reside in an application specific integrated circuit (Application Specific Integrated Circuits, ASIC for short). It is also possible that the processor and the storage medium reside as discrete components in an electronic device or a master device.
Those of ordinary skill in the art will appreciate that: all or part of the steps for implementing the method embodiments described above may be performed by hardware associated with program instructions. The foregoing program may be stored in a computer readable storage medium. The program, when executed, performs steps including the method embodiments described above; and the aforementioned storage medium includes: various media that can store program code, such as ROM, RAM, magnetic or optical disks.
Other embodiments of the invention will be apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention following, in general, the principles of the invention and including such departures from the present disclosure as come within known or customary practice within the art to which the invention pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the invention being indicated by the following claims.
It is to be understood that the invention is not limited to the precise arrangements and instrumentalities shown in the drawings, which have been described above, and that various modifications and changes may be effected without departing from the scope thereof. The scope of the invention is limited only by the appended claims.
Claims (9)
1. A method of seismic record inversion comprising:
acquiring a small sample seismic record, small sample wave impedance corresponding to the small sample seismic record and a large sample label-free seismic record;
constructing a dual WGAN model based on a neural network model, wherein the dual WGAN model comprises an inversion WGAN and a forward WGAN, the inversion WGAN comprises an inversion generator and an inversion discriminator, and the forward WGAN comprises a forward generator and a forward discriminator;
Optimizing the dual WGAN model by the small sample seismic record, the small sample wave impedance, and the large sample unlabeled seismic record;
and inputting the large sample unlabeled seismic records into an optimized inversion generator, and generating corresponding large sample wave impedance.
2. The method of claim 1, wherein the inversion generator is a one-dimensional U-net generator that is input as two channels under the constraint of the condition information, and the inversion arbiter is a one-dimensional Markovian arbiter that is input as two channels under the constraint of the condition information;
the forward generator is a one-dimensional U-net generator, and the forward arbiter is a one-dimensional AlexNet arbiter.
3. The method of claim 1, wherein optimizing the dual WGAN model by the small sample seismic record, the small sample wave impedance, and the large sample unlabeled seismic record comprises:
determining an inversion discriminant loss function, a forward discriminant loss function, and a generator loss function from the small sample seismic record, the small sample wave impedance, and the large sample unlabeled seismic record;
optimizing the inversion discriminator, the forward decision maker and the generator according to the inversion discriminator loss function, the forward decision maker loss function and the generator loss function;
Wherein the generator comprises the forward generator and the inversion generator.
4. The method of claim 3, wherein determining an inversion discriminator loss function, a forward discriminator loss function, and a generator loss function from the small sample seismic record, the small sample wave impedance, and the large sample unlabeled seismic record comprises:
inputting a low-frequency model, the small sample wave impedance and a seismic record into the inversion generator to obtain a wave impedance predicted value, wherein the seismic record comprises a small sample seismic record and a large sample label-free seismic record;
inputting the wave impedance predicted value and the small sample wave impedance into the inversion discriminator to obtain the inversion discriminator loss function;
inputting the wave impedance predicted value into the forward generator to obtain a reconstructed seismic record;
inputting the small sample wave impedance into the forward generator to obtain a seismic record predicted value;
inputting the seismic record predicted value and the seismic record into the forward model arbiter to obtain the forward model arbiter loss function;
inputting the seismic record predicted value into the inversion generator to obtain reconstructed wave impedance;
And determining the generator loss function according to the wave impedance predicted value, the seismic record predicted value, the reconstructed seismic record and the reconstructed wave impedance.
5. The method of claim 3 or 4, wherein the inversion arbiter loss function is:
wherein D is inver Inversion discriminant for inverting WGAN, G inver For inversion generator of inversion WGAN, AI is wavelet impedance of small sample, S is small sample seismic record corresponding to AI, (A|B) is expressed as A input network under B condition constraint, gp 1 Lambda is a gradient penalty term coefficient, and m is the number of small samples for training;
the forward arbiter loss function is:
wherein D is forward Inversion discriminant for inverting WGAN, G forward Inversion generator for inverting WGAN, gp 2 Is a gradient penalty term.
6. The method of claim 5, wherein the generating the loss function comprises generating a loss function for the countermeasure portion and a loss function for the cyclical compliance portion; the generating the loss function of the countermeasure section includes generating the loss function of the countermeasure inversion section and generating the loss function of the countermeasure forward section; the loss function of the loop consistency part comprises the loss function of the loop consistency open-loop part and the loss function of the loop consistency closed-loop part;
The generating a loss function against the inversion portion is:
wherein S is * AI for large sample unlabeled seismic recording low Is a wave impedance low-frequency model;
the generating a loss function against the forward part is:
the loss function of the loop-consistent open-loop part is:
the loss function of the cyclic consistency closed loop portion is:
the generator loss function is:
g_loss＝g_inver_loss+g_forward_loss
+α×(loss_3+loss_4+loss_5)
+β×(loss_1+loss_2)
wherein, alpha and beta are constraint coefficients respectively controlling the contribution of an open loop part and a closed loop part, and M is the number of large samples for training.
7. A seismic record inversion apparatus, the apparatus comprising:
the acquisition module is used for acquiring a small sample seismic record, small sample wave impedance corresponding to the small sample seismic record and a large sample label-free seismic record;
the construction module is used for constructing a dual WGAN model based on the neural network model, wherein the dual WGAN model comprises an inversion WGAN and a forward WGAN, the inversion WGAN comprises an inversion generator and an inversion discriminator, and the forward WGAN comprises a forward generator and a forward discriminator;
an optimization module for optimizing the dual WGAN model by the small sample seismic record, the small sample wave impedance, and the large sample unlabeled seismic record;
The generation module is used for inputting the large sample unlabeled seismic records into an optimized inversion generator and generating corresponding large sample wave impedance.
8. A seismic record inversion apparatus, comprising: a memory and at least one processor;
the memory stores computer-executable instructions;
the at least one processor executing computer-executable instructions stored in the memory causes the at least one processor to perform the seismic record inversion method of any of claims 1-6.
9. A computer readable storage medium having stored therein computer executable instructions which when executed by a processor are for implementing the seismic record inversion method of any of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110925389.0A CN113722893B (en) | 2021-08-12 | 2021-08-12 | Seismic record inversion method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110925389.0A CN113722893B (en) | 2021-08-12 | 2021-08-12 | Seismic record inversion method, device, equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113722893A CN113722893A (en) | 2021-11-30 |
| CN113722893B true CN113722893B (en) | 2023-06-09 |
Family
ID=78675762
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110925389.0A Active CN113722893B (en) | 2021-08-12 | 2021-08-12 | Seismic record inversion method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113722893B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117950025A (en) * | 2022-10-26 | 2024-04-30 | 中国石油天然气股份有限公司 | Reservoir physical property parameter prediction method and device |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111983681A (en) * | 2020-08-31 | 2020-11-24 | 电子科技大学 | Seismic wave impedance inversion method based on countermeasure learning |
| CN112213771A (en) * | 2019-07-10 | 2021-01-12 | 中国石油天然气股份有限公司 | Seismic wave impedance inversion method and device |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11105942B2 (en) * | 2018-03-27 | 2021-08-31 | Schlumberger Technology Corporation | Generative adversarial network seismic data processor |
-
2021
- 2021-08-12 CN CN202110925389.0A patent/CN113722893B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112213771A (en) * | 2019-07-10 | 2021-01-12 | 中国石油天然气股份有限公司 | Seismic wave impedance inversion method and device |
| CN111983681A (en) * | 2020-08-31 | 2020-11-24 | 电子科技大学 | Seismic wave impedance inversion method based on countermeasure learning |
Non-Patent Citations (2)
| Title |
|---|
| 地震波阻抗反演实验分析;李国发;王艳仓;熊金良;马彦彦;李皓;;石油地球物理勘探(第06期);第868-872+936+788页 * |
| 基于生成式对抗网络的无监督图像生成研究;谭宏卫;《中国优秀硕士论文集 社会科学II辑》(第4期);第1-126页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113722893A (en) | 2021-11-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Mandelli et al. | Interpolation and denoising of seismic data using convolutional neural networks | |
| CN107589448B (en) | A kind of multitrace seismogram reflection coefficient sequence Simultaneous Inversion method | |
| Sun et al. | ML-descent: An optimization algorithm for full-waveform inversion using machine learning | |
| CN105842732B (en) | The inversion method and system of the sparse reflectance factor of multiple tracks | |
| CN110361778B (en) | Seismic data reconstruction method based on generation countermeasure network | |
| Yang et al. | FWIGAN: Full‐Waveform Inversion via a Physics‐Informed Generative Adversarial Network | |
| CN112213771B (en) | Seismic wave impedance inversion method and device | |
| CN111596366B (en) | Wave impedance inversion method based on seismic signal optimization processing | |
| Mustafa et al. | Joint learning for spatial context-based seismic inversion of multiple data sets for improved generalizability and robustness | |
| CN110895348B (en) | Method, system and storage medium for extracting low-frequency information of seismic elastic impedance | |
| CN112415583A (en) | Seismic data reconstruction method and device, electronic equipment and readable storage medium | |
| CN113722893B (en) | Seismic record inversion method, device, equipment and storage medium | |
| CN110146923B (en) | High-efficiency high-precision depth domain seismic wavelet extraction method | |
| CN115598697A (en) | Thin-layer structure high-resolution seismic inversion method, device, medium and equipment | |
| CN118295029A (en) | Seismic data denoising method integrating self-attention and Mamba architecture | |
| CN117076858B (en) | Deep learning-based low-frequency geomagnetic strong interference suppression method and system | |
| Zhu et al. | Adaptive Gaussian mixture model and convolution autoencoder clustering for unsupervised seismic waveform analysis | |
| CN114545494A (en) | Non-supervision seismic data reconstruction method and device based on sparse constraint | |
| Sun et al. | Full-waveform inversion using a learned regularization | |
| Turek et al. | On MAP and MMSE estimators for the co-sparse analysis model | |
| Tetali et al. | Wave physics informed dictionary learning in one dimension | |
| Qian et al. | Unsupervised intense VSP coupling noise suppression with iterative robust deep learning | |
| CN112305586A (en) | Unsteady-state seismic data time-frequency analysis method, computer storage medium and system | |
| CN116090352A (en) | Full waveform inversion method based on gate cycle unit and attention mechanism | |
| CN115859012A (en) | Underwater DOA and SAP estimation method based on rapid multi-snapshot inverse-free sparse Bayes |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |