Method and device for removing sedimentary background based on stack auto-encoder

Abstract

The invention discloses a deposition background removing method and device based on a stack auto-encoder. The method comprises the following steps: performing well control phase conversion on an original seismic body, and performing Wheeler domain conversion to obtain a Wheeler domain seismic body after well control phase conversion; training a pre-constructed initial sedimentary background prediction model by using the Wheeler domain seismic body after well control phase conversion to obtain a sedimentary background prediction model; wherein the initial sedimentary background prediction model is formed based on a stack auto-encoder; inputting the Wheeler domain seismic body after well control phase conversion into the sedimentary background prediction model to obtain sedimentary background single-channel data; expanding sedimentary background single-channel data based on the original seismic body to obtain a sedimentary background body; and carrying out body operation on the Wheeler domain seismic body after well control phase conversion and the sedimentary background body to obtain a lithologic body after the sedimentary background is removed. According to the method, the stack auto-encoder is used for extracting the sedimentary background body, and the accuracy of effective reservoir seismic analysis is improved.

Description

Technical Field

[0001] The invention relates to the technical field of petroleum geophysical exploration, and in particular to a method and a device for removing sedimentation background based on a stacked autoencoder. Background Art

[0002] In recent years, strong reflection background seismic profiles are often encountered in the exploration of major oil and gas fields at home and abroad. The reflection coefficient of the reservoir is small, which often affects the adjacent upper and lower reservoirs, that is, it is difficult to carry out corresponding reservoir prediction research for the target layer affected by strong reflection. At the same time, due to the influence of structural and sedimentary factors, it is difficult to predict small river channels with existing seismic exploration technology. In view of this situation, it is urgent to develop methods to remove sedimentary background (including removing the influence of strong reflection formations).

[0003] In order to remove the influence of sedimentary background, in-depth research has been conducted at home and abroad. Matching pursuit and various derivative algorithms are mainly used to extract sedimentary background from seismic data, and then the sedimentary background is subtracted from the seismic record to eliminate the influence of sedimentary background (and strong reflection background) and improve the accuracy of seismic interpretation and reservoir prediction. At the same time, domestic scholars have used linear PCA method and its deformation algorithm to study the method of removing sedimentary background. Summary of the invention

[0004] In order to better remove the deposition background, an embodiment of the present invention provides a method and device for removing the deposition background based on a stacked autoencoder.

[0005] In a first aspect, an embodiment of the present invention provides a method for removing a deposition background based on a stacked autoencoder, the method comprising:

[0006] Perform well-controlled phase conversion on the original seismic volume, and perform Wheeler domain conversion to obtain the Wheeler domain seismic volume after well-controlled phase conversion;

[0007] Using the Wheeler domain seismic volume after the well-controlled phase conversion to train the pre-constructed initial sedimentary background prediction model, a sedimentary background prediction model is obtained; wherein the initial sedimentary background prediction model is based on a stacked autoencoder, including multiple encoders and multiple decoders;

[0008] Inputting the Wheeler domain seismic volume after the well-controlled phase conversion into the sedimentary background prediction model to obtain sedimentary background single-channel data;

[0009] Expanding the sedimentary background single-channel data based on the original seismic volume to obtain a sedimentary background volume;

[0010] Volume operations are performed on the Wheeler domain seismic volume after the well-controlled phase conversion and the sedimentary background volume to obtain a lithology volume after removing the sedimentary background.

[0011] In one or some optional implementations of the embodiments of the present application, the initial deposition background prediction model includes a plurality of encoders connected in sequence and a plurality of decoders having the same number as the encoders;

[0012] The method of using the Wheeler domain seismic volume after the well-controlled phase conversion to train the pre-built initial sedimentary background prediction model to obtain the sedimentary background prediction model includes:

[0013] Using the Wheeler domain seismic volume after the well-controlled phase conversion, training an autoencoder composed of a first encoder and a last decoder to obtain a trained first encoder;

[0014] Using the trained first encoder to encode the Wheeler domain seismic volume after the well control phase conversion, to obtain a first hidden vector data set;

[0015] Using the first hidden vector data set to train an autoencoder consisting of a second encoder and a penultimate decoder to obtain a trained second encoder;

[0016] Encode the first hidden vector data set using the trained second encoder to obtain a second hidden vector data set;

[0017] The above process of training the autoencoder and using the corresponding encoder to encode the input hidden vector data set is repeated until the deposition background prediction model consisting of a plurality of trained encoders connected in sequence is obtained through training.

[0018] In one or some optional implementations of the embodiment of the present application, the step of expanding the sedimentary background single-channel data based on the original seismic volume to obtain the sedimentary background volume includes:

[0019] Obtaining the total number of dimensions of the original seismic volume;

[0020] The single-channel sedimentary background data is copied and expanded based on the total number of dimensions to obtain a sedimentary background volume with the same total number of dimensions as the original seismic volume.

[0021] In one or some optional implementations of the embodiment of the present application, the volume operation is performed on the Wheeler domain seismic volume after the well-controlled phase conversion and the sedimentary background volume to obtain a lithological volume after removing the sedimentary background, including:

[0022] Based on the following formula 1, volume operation is performed on the Wheeler domain seismic volume after the well-controlled phase conversion and the sedimentary background volume, and the sedimentary background volume in the Wheeler domain seismic volume after the well-controlled phase conversion is removed to obtain a lithological volume after the sedimentary background is removed:

[0023] S l =SS b Formula 1;

[0024] Where S is the Wheeler domain seismic volume after well-controlled phase conversion, S b is the sedimentary background, S l Lithology data.

[0025] In one or some optional implementations of the embodiment of the present application, performing well-controlled phase conversion on the original seismic volume and performing Wheeler domain conversion to obtain the Wheeler domain seismic volume after well-controlled phase conversion includes:

[0026] Perform well-controlled phase conversion on the original seismic volume to obtain a 90-degree phased seismic volume;

[0027] The 90-degree phased seismic volume is transformed in the Wheeler domain to obtain a well-controlled phase-converted Wheeler domain seismic volume.

[0028] In one or some optional implementations of the embodiment of the present application, before the pre-constructed initial sedimentary background prediction model is trained using the Wheeler domain seismic volume after the well-controlled phase conversion to obtain the sedimentary background prediction model, the method further includes:

[0029] Based on the following formula 2, the Wheeler domain seismic volume after the well-controlled phase conversion is normalized:

[0030] S=-1+(seisdata w -seisdata w min ) / (seisdata w max -seisdata w min )×(1-(-1))Formula 2;

[0031] Where, seisdata w is the Wheeler domain seismic volume after well-controlled phase conversion, seisdata w min is the minimum value of the Wheeler domain seismic volume after well-controlled phase conversion, seisdata w max It is the maximum value of the Wheeler domain seismic volume after well-controlled phase conversion.

[0032] In a second aspect, an embodiment of the present invention provides a device for removing deposition background based on a stacked autoencoder, the device comprising:

[0033] The first conversion module is used to perform well-controlled phase conversion on the original seismic volume and perform Wheeler domain conversion to obtain a Wheeler domain seismic volume after well-controlled phase conversion;

[0034] A first training module is used to train a pre-constructed initial sedimentation background prediction model using the Wheeler domain seismic volume after the well-controlled phase conversion to obtain a sedimentation background prediction model; wherein the initial sedimentation background prediction model is based on a stacked autoencoder and includes multiple encoders and multiple decoders;

[0035] A first prediction module is used to input the Wheeler domain seismic volume after the well-controlled phase conversion into the sedimentary background prediction model to obtain sedimentary background single-channel data;

[0036] A first operation module is used to expand the sedimentary background single-channel data based on the original seismic volume to obtain a sedimentary background volume;

[0037] The second operation module is used to perform volume operation on the Wheeler domain seismic volume after the well-controlled phase conversion and the sedimentary background volume to obtain a lithology volume after removing the sedimentary background.

[0038] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned method for removing deposition background based on a stacked autoencoder.

[0039] In a fourth aspect, an embodiment of the present invention provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for removing deposition background based on the stacked autoencoder as described above is implemented.

[0040] In a fifth aspect, an embodiment of the present invention provides a computer program product comprising instructions, which, when executed on a computer device, enables the computer device to execute the above-mentioned method for removing deposition background based on a stacked autoencoder.

[0041] In a sixth aspect, an embodiment of the present invention provides a chip, the chip including a processor and a communication interface, the communication interface and the processor are coupled, and the processor is used to run a computer program or instruction to implement the above-mentioned method for removing deposition background based on the stack autoencoder.

[0042] The beneficial effects of the above technical solution provided by the embodiment of the present invention include at least:

[0043] The method for removing sedimentary background based on stacked autoencoders provided in an embodiment of the present invention uses phase-converted Wheeler domain seismic volume data as a training data set, and performs training and learning through stacked autoencoders to obtain sedimentary background bodies, and finally obtain more accurate lithological bodies. Using stacked autoencoders to extract features by dimensionality reduction can more effectively and stably extract features, overcome the problem of strong reflection layers and sedimentary background affecting reservoirs, significantly improve the precision and accuracy of using seismic data for effective reservoir seismic analysis, and can more finely characterize sedimentary reservoirs.

[0044] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings.

[0045] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0047] Figure 1 A schematic diagram of a flow chart of a method for removing deposition background based on a stacked autoencoder provided in an embodiment of the present invention;

[0048] Figure 2 It is a schematic diagram of a seismic volume section in the Wheeler domain after well-controlled phase conversion provided in an embodiment of the present application;

[0049] Figure 3 is a schematic diagram of an initial deposition background prediction model provided in an embodiment of the present application;

[0050] Figure 4 is a schematic diagram of a specific initial deposition background prediction model network structure provided in an embodiment of the present application;

[0051] Figure 5 is a schematic diagram of the extracted sedimentation background provided in an embodiment of the present application;

[0052] Figure 6 It is a schematic diagram of the lithology body obtained after removing the sedimentary background provided in the embodiment of the present application;

[0053] Figure 7 A schematic diagram of the structure of a device for removing deposition background based on a stacked autoencoder provided in an embodiment of the present application. DETAILED DESCRIPTION

[0054] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0055] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0056] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0057] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.

[0058] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0059] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0060] It should be understood that the size of the serial numbers of the steps in the following embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0061] In order to illustrate the technical solution of the present application, a specific embodiment is provided below for illustration.

[0062] The inventor found that in the prior art, matching pursuit and various derivative algorithms are mainly used to extract strong reflection waveforms from seismic data, and then the strong reflection waveform is subtracted from the seismic record to eliminate the influence of the strong reflection waveform and improve the accuracy of seismic interpretation and reservoir prediction. However, the matching pursuit algorithm also has shortcomings, such as the poor effect of removing intermittent strong axes and multiple strong axes, and it is basically unsolvable, and the wavelet library is easily redundant, resulting in a large amount of calculation and other problems. Due to the influence of tectonic and sedimentary background, the difficulty of seismic reservoir prediction analysis is increased, and it is necessary to remove the sedimentary background (including strong reflection waveforms). At the same time, domestic scholars have tried using linear PCA method and its deformation algorithm, but the effect is not stable.

[0063] In summary, the existing methods cannot completely and effectively remove the sedimentary background (including removing the strong reflection strata), and the increasing amount of seismic data also puts forward higher requirements for the method of removing the sedimentary background. Based on this, the inventors have made the present invention after further research and development, providing a method and device for removing the sedimentary background based on a stacked autoencoder.

[0064] Embodiment 1

[0065] The embodiment of the present invention provides a method for removing deposition background based on a stacked autoencoder, referring to Figure 1 As shown, the method includes:

[0066] S101: Perform well-controlled phase conversion on the original seismic volume, and perform Wheeler domain conversion to obtain a Wheeler domain seismic volume after well-controlled phase conversion.

[0067] In the embodiment of the present application, in the above step S101, the original seismic volume is subjected to well-controlled phase conversion and Wheeler domain conversion to obtain a Wheeler domain seismic volume after well-controlled phase conversion, including:

[0068] The original seismic volume is subjected to well-controlled phase conversion to obtain a 90-degree phased seismic volume; the 90-degree phased seismic volume is subjected to Wheeler domain transformation to obtain a Wheeler domain seismic volume after well-controlled phase conversion.

[0069] In an embodiment of the present application, seismic wave data collected by a multi-channel scanning method is obtained, a phase angle step is defined, the seismic body is corrected with different phase angle shifts, the waveform similarity or maximum energy criterion is used to determine the optimal phase shift, and finally the original seismic body is converted into a 90-degree phased seismic body, which is beneficial to the interpretation of the sand body. Among them, the steps of performing well-controlled phase conversion on the original seismic body to obtain a 90-degree phased seismic body include: the first step is to calculate the Hilbert transform, for the original seismic data X(t), perform the Hilbert transform to convert it into a complex signal H(t), through the Hilbert transform, the original seismic data is divided into a real part and an imaginary part to form a complex signal, wherein the real part contains the amplitude information of the original data, and the imaginary part contains the 90-degree phase data; the second step is to calculate the amplitude and phase, from the complex signal H(t), the amplitude A(t) and the phase Φ(t) can be calculated, wherein the amplitude A(t) represents the amplitude information of the data, and the phase Φ(t) represents the phase information of the data; the third step is to create a 90-degree phased seismic body, by combining the amplitude A(t) and the phase Φ(t), that is, increasing the phase by 90 degrees, to obtain a 90-degree phased seismic body.

[0070] After obtaining the 90-degree phased seismic volume, the 90-degree phased seismic volume is subjected to a three-dimensional Wheeler transformation, and the steps of obtaining the Wheeler domain seismic volume after well-controlled phase transformation include: the first step, using the layer tracking method to determine the sequence boundary of the underground strata, the goal of this step is to determine the layered structure of the underground strata; the second step, selecting the control sequence boundary, after determining the sequence boundary, it is necessary to select the control point, and select the continuous strong reflection axis above and below the target layer as the control sequence boundary; the third step, proportional division, the strata between the control sequence boundaries are equally divided, this step will divide the underground strata into several parts, and assign depth values ​​to each stratum; the fourth step, the layer is flattened, the depth value of the underground strata will be adjusted according to the control sequence boundary, and a uniform sequence model will be established in the underground strata. After the above steps, the Wheeler transformation is completed, and finally the 90-degree phased seismic volume is converted into the Wheeler domain seismic volume after well-controlled phase transformation.

[0071] In the embodiment of the present application, the Wheeler domain seismic volume after the well-controlled phase conversion is normalized to between [-1, 1] based on the following formula 2:

[0072] S=-1+(seisdata w -seisdata w min ) / (seisdata w max -seisdata wmin )×(1-(-1))Formula 2;

[0073] Where, seisdata w is the Wheeler domain seismic volume after well-controlled phase conversion, seisdata w min is the minimum value of the Wheeler domain seismic volume after well-controlled phase conversion, seisdata w max It is the maximum value of the Wheeler domain seismic volume after well-controlled phase conversion.

[0074] S102: using the Wheeler domain seismic volume after the well-controlled phase conversion to train a pre-constructed initial sedimentation background prediction model to obtain a sedimentation background prediction model; wherein the initial sedimentation background prediction model is based on a stacked autoencoder and includes multiple encoders and multiple decoders.

[0075] In the embodiment of the present application, in the above step S102, the normalized Wheeler domain seismic volume after well-controlled phase conversion is input as a training data set into the pre-constructed initial sedimentation background prediction model for training, and the Wheeler domain seismic volume after well-controlled phase conversion is also used as a sample label data set. The Wheeler domain seismic volume after well-controlled phase conversion has n seismic traces, each seismic trace has m sampling points, and the dimension of n in the seismic volume is very large. In order to obtain compressed encoded feature data, an initial sedimentation background prediction model is constructed based on a stacked autoencoder. The initial sedimentation background prediction model includes multiple encoders and multiple decoders, and the number of encoders is the same as the number of decoders.

[0076] The initial sedimentary background prediction model is greedily trained layer by layer: first, the first encoder and the last encoder are combined into the first shallow autoencoder, and the first shallow autoencoder is trained using the Wheeler domain seismic volume after well-controlled phase conversion to obtain the trained first encoder; next, the Wheeler domain seismic volume after well-controlled phase conversion is encoded using the trained first encoder to obtain the first hidden vector data set; then, the second encoder and the penultimate encoder are combined into the second autoencoder, and the second autoencoder is trained using the first hidden vector data set to obtain the trained second encoder; next, the first hidden vector data set is encoded using the trained second encoder to obtain the second hidden vector data set; repeat the above process of forming a new deep autoencoder and training using the hidden vector data obtained by the upper layer encoding until the training of the autoencoder composed of the last encoder and the first decoder in the initial sedimentary background prediction model is completed. The multiple trained encoders obtained in the above initial sedimentary background prediction model training process are sequentially connected to form a sedimentary background prediction model.

[0077] S103: Inputting the Wheeler domain seismic volume after the well-controlled phase conversion into the sedimentary background prediction model to obtain sedimentary background single-channel data.

[0078] S104: Expanding the sedimentary background single-channel data based on the original seismic volume to obtain a sedimentary background volume.

[0079] S105: Performing volume operation on the Wheeler domain seismic volume after the well-controlled phase conversion and the sedimentary background volume to obtain a lithology volume after removing the sedimentary background.

[0080] In the embodiment of the present application, in the above step S103, the Wheeler domain seismic volume after well control phase conversion is input into the sedimentation background prediction model to obtain sedimentation background single-channel data. The total number of dimensions of the original seismic volume data is obtained, and the sedimentation background single-channel data is copied and expanded based on the total number of dimensions to obtain a sedimentation background volume with the same dimension as the original seismic volume.

[0081] Based on the following formula 1, volume operations are performed on the Wheeler domain seismic volume and the sedimentary background volume after the well-controlled phase conversion, and the sedimentary background volume in the Wheeler domain seismic volume after the well-controlled phase conversion is removed to obtain the lithology volume after the sedimentary background is removed:

[0082] S l =SS b Formula 1;

[0083] Where S is the Wheeler domain seismic volume after well-controlled phase conversion, S b is the sedimentary background, S l Lithology data.

[0084] In order to explain the embodiment of the present invention more clearly, in a specific example, see Figure 2-Figure 6 As shown, the method for removing the deposition background based on the stacked autoencoder according to the embodiment of the present invention is described in detail:

[0085] The specific example data comes from the Sulige area of ​​the Ordos Basin, which is a typical low-porosity, low-permeability tight gas field. The main gas layers are the He 8, Shan 1, and Shan 2 sandstone layers of the Upper Paleozoic Permian, which are thin reservoirs with strong heterogeneity. The sedimentary facies characteristics of Shan 2 and Shan 1 affected by the Taiyuan Formation coal seam are as follows: Shan 2 is a meandering river delta deposit, mainly developed delta plain distributary channels, and floodplain mudstone deposits developed between channels; Shan 1 is a delta front subfacies, with relatively developed underwater distributary channel microfacies, and distributary bay mudstone deposits developed between channels. However, coal seams are commonly developed in the Shan 2 and Taiyuan Formations in this area, which show strong amplitude characteristics in seismic, and have a strong impact on the Shan 1 and Shan 2 sandstone reservoirs. At the same time, the background sedimentation also conceals some characteristics of the reservoir, and the sedimentary facies characteristics characterized by conventional seismic data are very unclear. Therefore, the present invention is applied to this area, with 201 data channels and 101 sampling points in each channel. The method of the present invention is used to remove the sedimentary background and finely characterize the lithology. The steps are as follows:

[0086] The first step is to perform a 90-degree phase conversion on the original seismic volume and perform a Wheeler domain transformation, such as Figure 2 The figure shows the cross section of a line in the work area after 90-degree phase conversion and Wheeler domain. Figure 2 It can be seen that due to the strong reflection of the Taiyuan Formation coal seam, the Shan 2 reservoir is located in the trough and its reservoir characteristics are obscured. At the same time, due to the influence of background sedimentation, the Shan 1 reservoir is too continuous and inconsistent with the actual sedimentary characteristics. This Wheeler domain profile cannot be directly used for subsequent reservoir prediction and analysis, and the influence of the background sedimentary body needs to be removed.

[0087] The second step is to extract the sedimentary background using a stacked autoencoder. First, the Wheeler domain seismic volume after 90-degree phase processing is normalized, and then the stacked autoencoder is used to train and extract the final sedimentary background. A stacked autoencoder with a 7-layer symmetrical structure is designed as the initial sedimentary background prediction model, such as Figure 3 As shown, it includes 3 encoders and 3 decoders. In the designed symmetrical structure stacked autoencoder deep network, there are a total of n features as the training sample parameters of the network (because the Wheeler domain seismic data is used as the training feature set input of n features, generally speaking, the dimension of n is large, hundreds or even tens of thousands are relatively normal). In order to obtain the compressed coded feature data body, a 7-layer deep network model with a network structure of 201-25-10-1-10-25-201 is constructed. The number of neurons gradually decreases from the 1st layer to the 4th layer, and the number of neurons in the 4th layer is reduced to one, which plays a role of compression coding in the network; the number of neurons in the remaining 3 layers is the same as the number of neurons in the first 3 layers. The number of nodes in the entire network is in a trumpet-like stacked shape, and a total of 3 decoders are set. The specific network structure of the initial sedimentary background prediction model is as follows Figure 4As shown in Figure 2. The activation function of the network training is set to the ReLU function. After the training is completed, the output result of the third encoder is extracted as the single-channel data of the sedimentary background, and expanded to the original seismic volume dimension to obtain the final sedimentary background volume. Figure 5 is the sedimentary background body obtained. Figure 2 By comparison, the coal seam at sampling point 40 was extracted, and some other sedimentary backgrounds were also extracted.

[0088] The third step is to perform volume operation on the Wheeler domain seismic volume after the well-controlled phase conversion and the sedimentary background volume to obtain a lithology volume after removing the sedimentary background.

[0089] Perform volume operations on the Wheeler domain seismic volume after well-controlled phase conversion and the obtained sedimentary background volume to obtain the lithology volume. Figure 6 The lithology obtained after removing the sedimentary background is shown. Figure 2 It can be seen that the sand body characteristics of Shan 2 and Shan 1 are prominent. This lithology can be used for detailed reservoir prediction and other work in the future.

[0090] The method for removing sedimentary background based on stacked autoencoders provided in the embodiment of the present scheme uses phase-converted Wheeler domain seismic volume data as a training data set and a sample data set, performs training and learning through a stacked autoencoder, performs dimensionality reduction extraction of ultra-high-dimensional features, obtains a sedimentary background body, and extracts features more effectively and stably, overcoming the problem of the influence of strong reflection layers and sedimentary background on the reservoir, and ultimately obtains a more accurate lithology body. Based on the more accurate lithology body, seismic reservoir prediction research can be carried out, which can improve the precision and accuracy of using seismic data for effective reservoir seismic analysis, and can more finely characterize sedimentary reservoirs.

[0091] Embodiment 2

[0092] Based on the same inventive concept, the embodiment of the present invention also provides a device for removing deposition background based on a stacked autoencoder, referring to Figure 7 As shown, the device comprises:

[0093] The first conversion module 101 is used to perform well-controlled phase conversion on the original seismic volume and perform Wheeler domain conversion to obtain a Wheeler domain seismic volume after well-controlled phase conversion;

[0094] The first training module 102 is used to train a pre-constructed initial sedimentation background prediction model using the Wheeler domain seismic volume after the well control phase conversion to obtain a sedimentation background prediction model; wherein the initial sedimentation background prediction model is based on a stacked autoencoder and includes multiple encoders and multiple decoders;

[0095] The first prediction module 103 is used to input the Wheeler domain seismic volume after the well-controlled phase conversion into the sedimentary background prediction model to obtain sedimentary background single-channel data;

[0096] A first operation module 104 is used to expand the sedimentary background single-channel data based on the original seismic volume to obtain a sedimentary background volume;

[0097] The second operation module 105 is used to perform volume operation on the Wheeler domain seismic volume after the well-controlled phase conversion and the sedimentary background volume to obtain a lithology volume after removing the sedimentary background.

[0098] Embodiment 3

[0099] Based on the same inventive concept, an embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the method for removing deposition background based on a stacked autoencoder as described in the above-mentioned embodiment 1 is implemented.

[0100] Embodiment 4

[0101] Based on the same inventive concept, an embodiment of the present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for removing deposition background based on the stacked autoencoder as described in the above-mentioned embodiment 1 is implemented.

[0102] Embodiment 5

[0103] Based on the same inventive concept, an embodiment of the present invention further provides a computer program product comprising instructions. When the computer program product is executed on a computer device, the computer device executes the method for removing deposition background based on a stacked autoencoder as described in the first embodiment above.

[0104] Embodiment 6

[0105] Based on the same inventive concept, an embodiment of the present invention also provides a chip, the chip includes a processor and a communication interface, the communication interface and the processor are coupled, and the processor is used to run a computer program or instruction to implement the method for removing deposition background based on the stack autoencoder as described in the above embodiment one.

[0106] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) containing computer-usable program code.

[0107] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0108] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0109] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0110] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for removing sedimentation background based on stacked autoencoders, characterized in that: include: Perform well-controlled phase conversion on the original seismic volume, and perform Wheeler domain conversion to obtain the Wheeler domain seismic volume after well-controlled phase conversion; Using the Wheeler domain seismic volume after the well-controlled phase conversion to train the pre-constructed initial sedimentary background prediction model, a sedimentary background prediction model is obtained; wherein the initial sedimentary background prediction model is based on a stacked autoencoder, including multiple encoders and multiple decoders; Inputting the Wheeler domain seismic volume after the well-controlled phase conversion into the sedimentary background prediction model to obtain sedimentary background single-channel data; Expanding the sedimentary background single-channel data based on the original seismic volume to obtain a sedimentary background volume; Volume operations are performed on the Wheeler domain seismic volume after the well-controlled phase conversion and the sedimentary background volume to obtain a lithology volume after removing the sedimentary background.

2. The method according to claim 1, characterized in that The initial deposition background prediction model comprises a plurality of encoders connected in sequence and a plurality of decoders having the same number as the encoders; The method of using the Wheeler domain seismic volume after the well-controlled phase conversion to train the pre-built initial sedimentary background prediction model to obtain the sedimentary background prediction model includes: Using the Wheeler domain seismic volume after the well-controlled phase conversion, training an autoencoder composed of a first encoder and a last decoder to obtain a trained first encoder; Using the trained first encoder to encode the Wheeler domain seismic volume after the well control phase conversion, to obtain a first hidden vector data set; Using the first hidden vector data set to train an autoencoder consisting of a second encoder and a penultimate decoder to obtain a trained second encoder; Encode the first hidden vector data set using the trained second encoder to obtain a second hidden vector data set; The above process of training the autoencoder and using the corresponding encoder to encode the input hidden vector data set is repeated until the deposition background prediction model consisting of a plurality of trained encoders connected in sequence is obtained through training.

3. The method according to claim 1, characterized in that The step of expanding the sedimentary background single-channel data based on the original seismic volume to obtain a sedimentary background volume includes: Obtaining the total number of dimensions of the original seismic volume; The single-channel sedimentary background data is copied and expanded based on the total number of dimensions to obtain a sedimentary background volume with the same total number of dimensions as the original seismic volume.

4. The method according to claim 1, characterized in that The performing volume operation on the Wheeler domain seismic volume after the well-controlled phase conversion and the sedimentary background volume to obtain a lithological volume after removing the sedimentary background includes: Based on the following formula 1, volume operation is performed on the Wheeler domain seismic volume after the well-controlled phase conversion and the sedimentary background volume, and the sedimentary background volume in the Wheeler domain seismic volume after the well-controlled phase conversion is removed to obtain a lithological volume after the sedimentary background is removed: S l =SS b Formula 1; Where S is the Wheeler domain seismic volume after well-controlled phase conversion, S b is the sedimentary background body, S l Lithology data.

5. The method according to claim 1, characterized in that The method of performing well-controlled phase conversion on the original seismic volume and performing Wheeler domain conversion to obtain a Wheeler domain seismic volume after well-controlled phase conversion includes: Perform well-controlled phase conversion on the original seismic volume to obtain a 90-degree phased seismic volume; The 90-degree phased seismic volume is transformed in the Wheeler domain to obtain a well-controlled phase-converted Wheeler domain seismic volume.

6. The method according to claim 1, characterized in that Before the pre-constructed initial sedimentary background prediction model is trained using the Wheeler domain seismic volume after the well-controlled phase conversion to obtain the sedimentary background prediction model, the method further includes: Based on the following formula 2, the Wheeler domain seismic volume after the well-controlled phase conversion is normalized: S=-1+(seisdata w -seisdata w min ) / (seisdata w max -seisdata w min )×(1-(-1)) Formula 2; Where, seisdata w is the Wheeler domain seismic volume after well-controlled phase conversion, seisdata w min is the minimum value of the Wheeler domain seismic volume after well-controlled phase conversion, seisdata w max It is the maximum value of the Wheeler domain seismic volume after well-controlled phase conversion.

7. A device for removing deposition background based on a stacked autoencoder, characterized in that: include: The first conversion module is used to perform well-controlled phase conversion on the original seismic volume and perform Wheeler domain conversion to obtain a Wheeler domain seismic volume after well-controlled phase conversion; A first training module is used to train a pre-constructed initial sedimentation background prediction model using the Wheeler domain seismic volume after the well-controlled phase conversion to obtain a sedimentation background prediction model; wherein the initial sedimentation background prediction model is based on a stacked autoencoder and includes multiple encoders and multiple decoders; A first prediction module is used to input the Wheeler domain seismic volume after the well-controlled phase conversion into the sedimentary background prediction model to obtain sedimentary background single-channel data; A first operation module is used to expand the sedimentary background single-channel data based on the original seismic volume to obtain a sedimentary background volume; The second operation module is used to perform volume operation on the Wheeler domain seismic volume after the well-controlled phase conversion and the sedimentary background volume to obtain a lithology volume after removing the sedimentary background.

8. A computer-readable storage medium storing instructions, which, when executed on a terminal, enable the terminal to execute the method for removing deposition background based on a stacked autoencoder as described in any one of claims 1 to 6.

9. A computer device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for removing deposition background based on a stacked autoencoder as described in any one of claims 1 to 6 is implemented.

10. A computer program product comprising instructions, which, when executed on a computer device, enables the computer device to execute the method for removing deposition background based on a stacked autoencoder according to any one of claims 1 to 6.

11. A chip, comprising a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run a computer program or instruction to implement the method for removing deposition background based on a stacked autoencoder as described in any one of claims 1 to 6.

Images