Multiple suppression method and device based on data association and long short-term memory network

Abstract

The invention discloses a multiple suppression method and device based on data association and a long short-term memory network, and the method comprises the steps: representing a CMP gather as a two-dimensional array, and enabling the amplitude value position of the CMP gather to meet a time-distance curve relation; performing interpolation operation on the CMP gather to obtain interpolation data; weighting the interpolation data through the interpolation coefficient to obtain a velocity spectrum; according to the two-dimensional array, the interpolation data and the velocity spectrum, recording and storing position index information corresponding to a primary wave and a multiple wave in the CMP trace gather, and extracting a multiple wave two-dimensional data set; training a long-short-term memory network according to the multiple two-dimensional data set, and inputting the original multiple into the long-short-term memory network for prediction to obtain multiple signals of the cross region; and matching the multiple signal with the original multiple according to the position index information. According to the method, data correlation and a long-short-term memory neural network algorithm are combined, accurate suppression of multiple waves is achieved, and primary wave signals are effectively protected.

Description

Technical Field

[0001] The present invention relates to the field of seismic exploration technology, and in particular to a multiple wave suppression method and device based on data association and long short-term memory network, a computing device and a computer storage medium. Background Art

[0002] In seismic data processing, the presence of multiple waves can significantly affect the accuracy and reliability of seismic imaging. Multiple waves are generated by different underground interfaces (such as low-velocity zones, coal seams, and igneous rocks), which often interfere with the reflection signal of the primary wave, thus bringing challenges to the interpretation of seismic data and may even lead to misleading results. Traditional multiple wave suppression methods mainly include filtering methods and predictive subtraction techniques, among which the Radon transform method is the most mature and widely used one. However, these methods face problems such as frequency aliasing and primary wave signal damage when dealing with complex geological structures, and it is difficult to meet the needs of high-precision exploration.

[0003] To solve the above problems, the present invention proposes a multiple wave suppression method based on data correlation and LSTM algorithm, which has better signal fidelity and amplitude retention characteristics compared with the traditional Radon filtering method. Summary of the invention

[0004] In view of the above problems, the present invention is proposed to provide a multiple wave suppression method and apparatus based on data association and long short-term memory network, a computing device and a computer storage medium, so as to achieve accurate suppression of multiple waves and effectively protect the primary wave signal.

[0005] According to one aspect of the present invention, a multiple wave suppression method based on data association and long short-term memory network is provided, comprising: The CMP gather is represented as a two-dimensional array, wherein the amplitude value position of the CMP gather satisfies the time-distance curve relationship; Performing interpolation operation on the CMP gather to obtain interpolation data; weighting the interpolation data by an interpolation coefficient to obtain a velocity spectrum; According to the two-dimensional array, the interpolation data and the velocity spectrum, the position index information corresponding to the primary wave and the multiple waves in the CMP gather is recorded and stored; and the two-dimensional data set of the multiple waves is extracted through the position index information; Training a long short-term memory network according to the two-dimensional multiple wave data set, inputting the original multiple waves into the long short-term memory network for prediction, and obtaining multiple wave signals in the intersection area including the primary wave and the multiple waves; The multiple wave signals are matched with the original multiple waves according to the position index information to achieve multiple wave suppression of the original multiple waves.

[0006] In an optional manner, the time-distance curve relationship is:

[0007]

[0008] Where t is the propagation time from the source to each observation point on the survey line; t o is the self-excitation and self-reception time; x is the offset distance of the observation point relative to the excitation point; v is the propagation speed.

[0009] In an optional manner, the calculation formula of the velocity spectrum is:

[0010]

[0011] in, is the interpolated data value, located in the ig+L+1th row and ihth column of the two-dimensional array; L is the smoothing window size parameter; it is the time index; iv is the spatial index; nb is the number of seismic traces.

[0012] In an optional manner, the interpolated data value is:

[0013]

[0014] Among them, ɑ is the interpolation coefficient; , is the interpolation point; , Under index ih , The data value at the interpolation point.

[0015] In an optional manner, the long short-term memory network includes an input gate, a forget gate, and an output gate;

[0016] Wherein, the input gate is used to control the importance of current information and save it in the cell state;

[0017] The forget gate is used to control the importance of historical information and the number of previous cell states;

[0018] The output gate is used to output the current cell state.

[0019] In an optional manner, the expression of the input gate is:

[0020] C t ~ = tanh( W c .[ h t − 1 , x t ] + b c )

[0021] in, is the weight matrix of the candidate state; is the bias term of the candidate state; is the output status at the previous moment; is the input at the current moment;

[0022] The expression of the forget gate is:

[0023] f t = σ ( W f .[ h t − 1 , x t ] + b f )

[0024] in, is the weight matrix of the forget gate; is the bias term of the forget gate; is the sigmoid activation function;

[0025] The expression of the output gate is:

[0026] o t = σ ( W o .[ h t − 1 , x t ] + b o ) h t = O t × tanh( C t )

[0027] in, is the weight matrix of the output gate; is the bias term of the output gate; is the output status at the current moment; is the output of the output gate; is the unit state, , i t = σ ( W i .[ h t − 1 , x t ] + b i ) , is the weight matrix of the input gate, is the bias term of the input gate.

[0028] In an optional manner, extracting the multiple wave two-dimensional data set includes a fixed time window method and an adaptive time window method.

[0029] According to another aspect of the present invention, there is provided a multiple wave suppression device based on data association and long short-term memory network, comprising:

[0030] A two-dimensional array representation module, used for representing the CMP gather as a two-dimensional array, wherein the amplitude value position of the CMP gather satisfies the time-distance curve relationship;

[0031] A velocity spectrum calculation module is used to perform interpolation operation on the CMP gather to obtain interpolation data; and to weight the interpolation data by an interpolation coefficient to obtain a velocity spectrum;

[0032] A data set generation module is used to record and store position index information corresponding to the primary wave and the multiple waves in the CMP gather according to the two-dimensional array, the interpolation data and the velocity spectrum; and extract the two-dimensional data set of the multiple waves through the position index information;

[0033] A multiple wave prediction module, used for training a long short-term memory network according to the two-dimensional multiple wave data set, inputting the original multiple waves into the long short-term memory network for prediction, and obtaining a multiple wave signal of the intersection area containing the primary wave and the multiple waves;

[0034] A matching and suppression module is used to match the multiple wave signal with the original multiple wave according to the position index information to achieve multiple wave suppression of the original multiple wave.

[0035] According to another aspect of the present invention, there is provided a computing device, comprising: a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus;

[0036] The memory is used to store at least one executable instruction, and the executable instruction enables the processor to execute operations corresponding to the above-mentioned multiple wave suppression method based on data association and long short-term memory network.

[0037] According to another aspect of the present invention, a computer storage medium is provided, wherein the storage medium stores at least one executable instruction, wherein the executable instruction enables a processor to perform operations corresponding to the above-mentioned multiple wave suppression method based on data association and long short-term memory network.

[0038] According to the scheme provided by the present invention, the CMP gather is represented as a two-dimensional array, wherein the amplitude value position of the CMP gather satisfies the time-distance curve relationship; the CMP gather is interpolated to obtain interpolation data; the interpolation data is weighted by the interpolation coefficient to obtain a velocity spectrum; according to the two-dimensional array, the interpolation data and the velocity spectrum, the position index information corresponding to the primary wave and the multiple waves in the CMP gather is recorded and stored; the multiple wave two-dimensional data set is extracted through the position index information; the long short-term memory network is trained according to the multiple wave two-dimensional data set, and the original multiple waves are input into the long short-term memory network for prediction to obtain the multiple wave signal of the intersection area containing the primary wave and the multiple waves; the multiple wave signal is matched with the original multiple waves according to the position index information to achieve the multiple wave suppression of the original multiple waves. The present invention records and stores the position index of the multiple waves in the CMP gather, and uses the LSTM neural network to train and predict the extracted multiple wave data, so as to effectively solve the suppression problem of the intersection area between the primary wave and the multiple waves, realize the accurate suppression of the multiple waves, and protect the primary wave signal at the same time. It performs well in processing complex geological models and has better signal fidelity and amplitude retention characteristics compared to traditional Radon filtering methods.

[0039] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0041] Figure 1 A schematic flow chart of a multiple wave suppression method based on data association and long short-term memory network according to an embodiment of the present invention is shown;

[0042] Figure 2 A schematic diagram showing a flow chart of data processing of a seismic exploration part according to an embodiment of the present invention;

[0043] Figure 3 A schematic diagram of a DC-LSTM algorithm processing framework according to an embodiment of the present invention is shown;

[0044] Figure 4 A schematic diagram showing data correlation between a CMP gather, a velocity spectrum and a stacked profile according to an embodiment of the present invention is shown;

[0045] Figure 5 A schematic diagram showing the DC-LSTM algorithm for extracting and predicting multiple waves according to an embodiment of the present invention is shown;

[0046] Figure 6 A schematic diagram of multiple waves of suppressed synthetic data according to an embodiment of the present invention is shown;

[0047] Figure 7 A schematic diagram of suppressing complex synthetic data multiple waves according to an embodiment of the present invention is shown;

[0048] Figure 8 A schematic diagram of an actual production data experiment of an embodiment of the present invention is shown;

[0049] Fig. 9 A schematic diagram of the structure of a multiple wave suppression device based on data association and long short-term memory network according to an embodiment of the present invention is shown;

[0050] Fig.10 A schematic diagram of the structure of a computing device according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0051] The exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present invention and to enable the scope of the present invention to be fully communicated to those skilled in the art.

[0052] Figure 1 FIG. 1 is a flow chart of a multiple wave suppression method based on data association and long short-term memory network according to an embodiment of the present invention. Specifically, Figure 1 As shown, the following steps are included:

[0053] Step S101, representing the CMP gather as a two-dimensional array, wherein the amplitude value position of the CMP gather satisfies the time-distance curve relationship.

[0054] In this embodiment, the process of processing the shot records collected in the field into seismic data generally requires three steps: preprocessing, pre-stack processing, and post-stack processing. In seismic data processing, the intermediate data of each operation step has a certain correlation. The process of suppressing multiple waves is as follows: Figure 2 As shown in the above processing, the close relationship between CMP (common midpoint) gathers, velocity spectra and stacked profiles, the multiple waves on the CMP gathers (such as Figure 4 a) forms an energy cluster on the velocity spectrum (such as Figure 4 b), the time-velocity is obtained through the energy group, and then the CMP gather is flattened after the NMO operation, and the stacked section is formed by stacking (as shown in Figure 4 c). Based on the data association relationship in the above process, when performing velocity analysis, the position indexes of multiple waves and primary waves in the CMP gather are recorded and stored. Specifically, the CMP gather can be regarded as a two-dimensional array, and its amplitude value position satisfies the time-distance curve relationship as follows:

[0055]

[0056] Where t is the propagation time from the source to each observation point on the survey line; t o is the self-excitation and self-reception time; x is the offset distance of the observation point relative to the excitation point; v is the propagation speed.

[0057] Step S102, performing interpolation operation on the CMP gather to obtain interpolation data; weighting the interpolation data by interpolation coefficients to obtain a velocity spectrum.

[0058] Specifically, interpolation processing is performed on the CMP gathers. Assuming that d(i,j) is the value of the i-th time point and the j-th spatial point in the seismic data, the interpolated data value s in the i+L+1th row and the ihth column can be expressed as:

[0059]

[0060] Among them, ɑ is the interpolation coefficient; , is the interpolation point; , Under index ih , The data value at the interpolation point.

[0061] Then, the interpolation coefficient is used to weight the interpolation data to calculate the velocity spectrum, and the calculation formula is:

[0062]

[0063] in, is the interpolated data value, located in the ig+L+1th row and ihth column of the two-dimensional array; L is the smoothing window size parameter; it is the time index; iv is the spatial index; nb is the number of seismic traces.

[0064] Step S103, recording and storing position index information corresponding to the primary wave and the multiple waves in the CMP gather according to the two-dimensional array, the interpolation data and the velocity spectrum; and extracting the two-dimensional data set of the multiple waves through the position index information.

[0065] In this embodiment, a fixed time window (such as Figure 5 a) and adaptive time window (as shown in Figure 5 The method shown in the figure above can accurately obtain the multiple wave two-dimensional data set. Figure 5 b is the result of multiple wave prediction with fixed time window (SNR = 5), Figure 5 d is the result of adaptive time window prediction of multiple waves.

[0066] Step S104, training a long short-term memory network according to the multiple wave two-dimensional data set, inputting the original multiple waves into the long short-term memory network for prediction, and obtaining multiple wave signals of the intersection area including the primary wave and the multiple waves.

[0067] In this embodiment, there is often an intersection area between the primary wave and the multiple waves in the extracted multiple waves. In order to deal with the problem of the intersection area between the primary wave and the multiple waves, the LSTM neural network algorithm is used to train and predict the multiple wave data. The LSTM neural network algorithm can effectively predict the multiple wave signals in the intersection area through its unique input gate, forget gate and output gate mechanism, thereby realizing the accurate extraction and suppression of the multiple waves. Figure 3 As shown in the figure, first, after performing velocity analysis on the input layer data (Input Data), record the cdp (common center point) and index (index) positioning data, and obtain additional 2D data (extra2Ddata) based on the positioning data. Then, perform dimension processing and feature extraction (Dimension Processing and TrainingExtraction). Then, use Long Short Term Memory Forecast (LSTM prediction) to process the long-term dependencies in the sequence data and obtain the prediction results. Finally, perform data correlation processing (DataCorrelationProcessing), and perform NMO (common center point) and stack (stack) processing.

[0068] In this embodiment, the input gate is used to control the importance of current information and save it in the cell state; the forget gate is used to control the importance of historical information and the number of previous cell states; and the output gate is used to output the current cell state.

[0069] Among them, the expression of the input gate is:

[0070] C t ~ = tanh( W c .[ h t − 1 , x t ] + b c )

[0071] in, is the weight matrix of the candidate state; is the bias term of the candidate state; is the output status at the previous moment; is the input at the current moment;

[0072] The expression of the forget gate is:

[0073] f t = σ ( W f .[ h t − 1 , x t ] + b f )

[0074] in, is the weight matrix of the forget gate; is the bias term of the forget gate; is the sigmoid activation function;

[0075] The expression of the output gate is:

[0076] o t = σ ( W o .[ h t − 1 , x t ] + b o ) h t = O t × tanh( C t )

[0077] in, is the weight matrix of the output gate; is the bias term of the output gate; is the output status at the current moment; is the output of the output gate; is the unit state, , i t = σ ( W i .[ h t − 1 , x t ] + b i ) , is the weight matrix of the input gate, is the bias term of the input gate.

[0078] This embodiment is based on the close correlation in seismic data processing and uses the precise prediction capability of the long short-term memory (LSTM) neural network algorithm. The multiple wave data set is accurately obtained by recording and storing the position index of the multiple waves collected by the CMP, and then the LSTM neural network algorithm is used to train the extracted multiple wave data to achieve accurate extraction and precise suppression of multiple waves.

[0079] Step S105: matching the multiple wave signal with the original multiple wave according to the position index information to achieve multiple wave suppression of the original multiple wave.

[0080] For example, multiple signals are aligned with the original multiples in the time domain to match the time delay with the original multiples. In addition to time delay, the spatial position of the multiple signals can also be adjusted based on the propagation path of the seismic waves to align them spatially with the original multiples. After the multiple signals are accurately matched with the original multiples in both time and space, the multiples are suppressed by methods such as signal subtraction (subtracting the matched multiple signals from the original seismic data to retain the valid reflection waves).

[0081] In order to verify the effectiveness of the method proposed in this embodiment, the following processing experiments of synthetic data and actual data are carried out.

[0082] Figure 6 Demonstrating different methods for suppressing multiples in synthetic data. Figure 6 a Input data containing multiple waves and noise (SNR = 5); Figure 6 b. is the result of Radon transform to remove multiple waves; Figure 6 c is the result of removing multiple waves using the DC-LSTM algorithm; Figure 6d is the difference between (b) and (a) (solid black arrows indicate multiples; dashed black arrows indicate primaries; white arrows indicate aliasing); Figure 6 e is the difference between (c) and (a) (solid black arrows indicate multiple waves).

[0083] Figure 7 Different methods for suppressing multiple waves in complex synthetic data. Figure 7 a is the superimposed section of the original data; Figure 7 b is the result of suppressing multiple waves by Radon transform (black arrows point to multiple waves; white arrows indicate false frequencies); Figure 7 c is the result of suppressing multiple waves using the DC-LSTM algorithm.

[0084] Figure 8 It is the result of actual production data processing; Figure 8 a is the multiple wave input data; Figure 8 b is the result of suppressing multiple waves by Radon transform; Figure 8 c is the result of suppressing multiple waves using the DC-LSTM algorithm (the black solid arrows point to the multiple waves; the black dotted arrows indicate the primary waves; the white arrows point to the false frequencies).

[0085] The above experimental results show that compared with the traditional Radon transform to suppress multiple waves, the DC-LSTM method proposed in this embodiment can effectively protect the primary wave signal and avoid aliasing while effectively suppressing multiple waves, and has better amplitude retention characteristics and higher processing accuracy.

[0086] The solution provided by the above embodiment of the present invention represents the CMP gather as a two-dimensional array, wherein the amplitude value position of the CMP gather satisfies the time-distance curve relationship; interpolation operation is performed on the CMP gather to obtain interpolation data; the interpolation data is weighted by the interpolation coefficient to obtain a velocity spectrum; according to the two-dimensional array, the interpolation data and the velocity spectrum, the position index information corresponding to the primary wave and the multiple waves in the CMP gather is recorded and stored; the multiple wave two-dimensional data set is extracted through the position index information; the long short-term memory network is trained according to the multiple wave two-dimensional data set, and the original multiple waves are input into the long short-term memory network for prediction to obtain the multiple wave signal of the intersection area containing the primary wave and the multiple waves; the multiple wave signal is matched with the original multiple waves according to the position index information to achieve the multiple wave suppression of the original multiple waves. The present invention records and stores the position index of the multiple waves in the CMP gather, and uses the LSTM neural network to train and predict the extracted multiple wave data, so as to effectively solve the suppression problem of the intersection area between the primary wave and the multiple waves, realize the accurate suppression of the multiple waves, and protect the primary wave signal at the same time. It performs well in processing complex geological models and has better signal fidelity and amplitude retention characteristics compared to traditional Radon filtering methods.

[0087] Fig. 9 The schematic diagram of the structure of the multiple wave suppression device based on data association and long short-term memory network according to an embodiment of the present invention is shown, comprising: A two-dimensional array representation module 910 is used to represent the CMP gather as a two-dimensional array, wherein the amplitude value position of the CMP gather satisfies the time-distance curve relationship; The velocity spectrum calculation module 920 is used to perform interpolation operation on the CMP gather to obtain interpolation data; weight the interpolation data by interpolation coefficients to obtain a velocity spectrum; The data set generation module 930 is used to record and store the position index information corresponding to the primary wave and the multiple waves in the CMP gather according to the two-dimensional array, the interpolation data and the velocity spectrum; and extract the two-dimensional data set of the multiple waves through the position index information; A multiple wave prediction module 940 is used to train a long short-term memory network according to the multiple wave two-dimensional data set, input the original multiple waves into the long short-term memory network for prediction, and obtain multiple wave signals of the intersection area containing the primary wave and the multiple waves; The matching and suppression module 950 is used to match the multiple wave signal with the original multiple wave according to the position index information to achieve multiple wave suppression of the original multiple wave.

[0088] The solution provided by the above embodiment of the present invention represents the CMP gather as a two-dimensional array, wherein the amplitude value position of the CMP gather satisfies the time-distance curve relationship; interpolation operation is performed on the CMP gather to obtain interpolation data; the interpolation data is weighted by the interpolation coefficient to obtain a velocity spectrum; according to the two-dimensional array, the interpolation data and the velocity spectrum, the position index information corresponding to the primary wave and the multiple waves in the CMP gather is recorded and stored; the multiple wave two-dimensional data set is extracted through the position index information; the long short-term memory network is trained according to the multiple wave two-dimensional data set, and the original multiple waves are input into the long short-term memory network for prediction to obtain the multiple wave signal of the intersection area containing the primary wave and the multiple waves; the multiple wave signal is matched with the original multiple waves according to the position index information to achieve the multiple wave suppression of the original multiple waves. The present invention records and stores the position index of the multiple waves in the CMP gather, and uses the LSTM neural network to train and predict the extracted multiple wave data, so as to effectively solve the suppression problem of the intersection area between the primary wave and the multiple waves, realize the accurate suppression of the multiple waves, and protect the primary wave signal at the same time. It performs well in processing complex geological models and has better signal fidelity and amplitude retention characteristics compared to traditional Radon filtering methods.

[0089] Fig.10 The schematic diagram of the structure of the computing device embodiment of the present invention is shown. The specific embodiment of the present invention does not limit the specific implementation of the computing device.

[0090] like Fig.10 As shown, the computing device may include: a processor (processor) 1002 , a communication interface (Communications Interface) 1004 , a memory (memory) 1006 , and a communication bus 1008 .

[0091] The processor 1002, the communication interface 1004, and the memory 1006 communicate with each other via a communication bus 1008. The communication interface 1004 is used to communicate with other devices such as a client or other server network elements. The processor 1002 is used to execute a program 1010, which can specifically execute the relevant steps in the above-mentioned embodiment of the multiple wave suppression method based on data association and long short-term memory network.

[0092] Specifically, the program 1010 may include program codes, which include computer operation instructions.

[0093] The processor 1002 may be a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention. The one or more processors included in the computing device may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0094] The memory 1006 is used to store the program 1010. The memory 1006 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0095] An embodiment of the present invention provides a non-volatile computer storage medium, wherein the computer storage medium stores at least one executable instruction, and the computer executable instruction can execute the multiple wave suppression method based on data association and long short-term memory network in any of the above method embodiments.

[0096] The algorithm or display provided herein is not inherently related to any particular computer, virtual system or other equipment. Various general purpose systems can also be used together with the teachings based on this. According to the above description, it is obvious to construct the structure required for this type of system. In addition, the embodiment of the present invention is not directed to any specific programming language yet. It should be understood that various programming languages ​​can be utilized to realize the content of the present invention described herein, and the description made to specific languages ​​above is for disclosing the best mode of the present invention.

[0097] In the description provided herein, a large number of specific details are described. However, it is understood that embodiments of the present invention can be practiced without these specific details. In some instances, well-known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this description.

[0098] Similarly, it should be understood that in order to streamline the present invention and aid in understanding one or more of the various inventive aspects, in the above description of exemplary embodiments of the present invention, the various features of the embodiments of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, this disclosed method should not be interpreted as reflecting the following intention: that the claimed invention requires more features than the features explicitly recited in each claim. More specifically, as reflected in the claims below, inventive aspects lie in less than all the features of the individual embodiments disclosed above. Therefore, the claims that follow the specific embodiment are hereby expressly incorporated into the specific embodiment, with each claim itself serving as a separate embodiment of the present invention.

[0099] Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and further may be divided into a plurality of submodules or subunits or subcomponents. All features disclosed in this specification (including the accompanying claims, abstracts and drawings) and all processes or units of any method or device so disclosed may be combined in any combination, except that at least some of such features and / or processes or units are mutually exclusive. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstracts and drawings) may be replaced by an alternative feature that provides the same, equivalent or similar purpose.

[0100] In addition, those skilled in the art will appreciate that, although some embodiments herein include certain features included in other embodiments but not other features, the combination of features of different embodiments is meant to be within the scope of the present invention and form different embodiments. For example, in the claims below, any one of the claimed embodiments may be used in any combination.

[0101] The various component embodiments of the present invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It should be understood by those skilled in the art that a microprocessor or a digital signal processor (DSP) may be used in practice to implement some or all of the functions of some or all of the components according to the embodiments of the present invention. The present invention may also be implemented as a device or apparatus program (e.g., a computer program and a computer program product) for executing part or all of the methods described herein. Such a program implementing the present invention may be stored on a computer-readable medium, or may be in the form of one or more signals. Such a signal may be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0102] It should be noted that the above embodiments illustrate the present invention rather than limit it, and that those skilled in the art may devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference symbol between brackets shall not be construed as a limitation on the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "one" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention may be implemented by means of hardware comprising a number of different elements and by means of a suitably programmed computer. In a unit claim enumerating a number of devices, several of these devices may be embodied by the same hardware item. The use of the words first, second, and third, etc. does not indicate any order. These words may be interpreted as names. The steps in the above embodiments, unless otherwise specified, should not be understood as limitations on the order of execution.

Claims

1. A multiple wave suppression method based on data association and long short-term memory network, characterized in that: include: The CMP gather is represented as a two-dimensional array, wherein the amplitude value position of the CMP gather satisfies the time-distance curve relationship; Performing interpolation operation on the CMP gather to obtain interpolation data; weighting the interpolation data by an interpolation coefficient to obtain a velocity spectrum; According to the two-dimensional array, the interpolation data and the velocity spectrum, the position index information corresponding to the primary wave and the multiple waves in the CMP gather is recorded and stored; and the two-dimensional data set of the multiple waves is extracted through the position index information; Training a long short-term memory network according to the two-dimensional multiple wave data set, inputting the original multiple waves into the long short-term memory network for prediction, and obtaining multiple wave signals in the intersection area including the primary wave and the multiple waves; The multiple wave signals are matched with the original multiple waves according to the position index information to achieve multiple wave suppression of the original multiple waves.

2. The multiple wave suppression method based on data association and long short-term memory network according to claim 1 is characterized in that: The time-distance curve relationship is: ； Where t is the propagation time from the source to each observation point on the survey line; t o is the self-excitation and self-reception time; x is the offset distance of the observation point relative to the excitation point; v is the propagation speed.

3. The multiple wave suppression method based on data association and long short-term memory network according to claim 1 or 2, characterized in that: The calculation formula of the velocity spectrum is: ； in, is the interpolated data value, located in the ig+L+1th row and ihth column of the two-dimensional array; L is the smoothing window size parameter; it is the time index; iv is the spatial index; nb is the number of seismic traces.

4. The multiple wave suppression method based on data association and long short-term memory network according to claim 3 is characterized in that: The interpolated data value is: ； Among them, ɑ is the interpolation coefficient; , is the interpolation point; , Under index ih , The data value at the interpolation point.

5. The multiple wave suppression method based on data association and long short-term memory network according to claim 1 is characterized in that: The long short-term memory network includes an input gate, a forget gate and an output gate; Wherein, the input gate is used to control the importance of current information and save it in the cell state; The forget gate is used to control the importance of historical information and the number of previous cell states; The output gate is used to output the current cell state.

6. The multiple wave suppression method based on data association and long short-term memory network according to claim 5, characterized in that: The expression of the input gate is: ； in, is the weight matrix of the candidate state; is the bias term of the candidate state; is the output status at the previous moment; is the input at the current moment; The expression of the forget gate is: ； in, is the weight matrix of the forget gate; is the bias term of the forget gate; is the sigmoid activation function; The expression of the output gate is: ； in, is the weight matrix of the output gate; is the bias term of the output gate; is the output status at the current moment; is the output of the output gate; is the unit state, , , is the weight matrix of the input gate, is the bias term of the input gate.

7. The multiple wave suppression method based on data association and long short-term memory network according to claim 1 is characterized in that: Extraction of multiple wave two-dimensional data sets includes fixed time window method and adaptive time window method.

8. A multiple wave suppression device based on data association and long short-term memory network, characterized in that: include: A two-dimensional array representation module, used for representing the CMP gather as a two-dimensional array, wherein the amplitude value position of the CMP gather satisfies the time-distance curve relationship; A velocity spectrum calculation module is used to perform interpolation operation on the CMP gather to obtain interpolation data; and to weight the interpolation data by an interpolation coefficient to obtain a velocity spectrum; A data set generation module is used to record and store position index information corresponding to the primary wave and the multiple waves in the CMP gather according to the two-dimensional array, the interpolation data and the velocity spectrum; and extract the two-dimensional data set of the multiple waves through the position index information; A multiple wave prediction module, used for training a long short-term memory network according to the two-dimensional multiple wave data set, inputting the original multiple waves into the long short-term memory network for prediction, and obtaining a multiple wave signal of the intersection area containing the primary wave and the multiple waves; A matching and suppression module is used to match the multiple wave signal with the original multiple wave according to the position index information to achieve multiple wave suppression of the original multiple wave.

9. A computing device comprising: A processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus; The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform operations corresponding to the multiple wave suppression method based on data association and long short-term memory network as described in any one of claims 1 to 7.

10. A computer storage medium, wherein at least one executable instruction is stored in the storage medium, wherein the executable instruction enables a processor to perform operations corresponding to the multiple wave suppression method based on data association and long short-term memory network as described in any one of claims 1 to 7.

Images