Data-driven direction signal deconvolution method and apparatus, and readable storage medium

Abstract

The present invention belongs to the technical field of seismic data processing. Provided are a data-driven direction signal deconvolution method and apparatus, and a readable storage medium. The method comprises: acquiring a pre-stack seismic trace gather and a preset desired wavelet; performing forward Fourier transform, three-dimensional τ-ρ transform and inverse Fourier transform on the pre-stack seismic trace gather, so as to obtain first seismic data; performing time difference correction on the first seismic data, so as to obtain second seismic data; projecting each of the first seismic data and the second seismic data onto a preset bin grid, and forming first directional seismic data and second directional seismic data; obtaining a direction matching operator on the basis of the first directional seismic data and the preset desired wavelet; and using the direction matching operator to correct the second directional seismic data, and performing inverse three-dimensional τ-ρ transform on corrected data, so as to obtain final seismic trace gather data. The present invention has the advantages of achieving high calculation precision and efficiency, greatly reducing the occupancy rate of a memory and the calculation cost, and increasing the signal-to-noise ratio of seismic data.

Description

Data-driven directional signal deconvolution method, device and readable storage medium Technical Field

[0001] The present invention relates to the technical field of seismic data processing, and in particular to a data-driven directional signal deconvolution method, a data-driven directional signal deconvolution device and a readable storage medium. Background Art

[0002] During marine oil and gas exploration, airguns are often used as excitation sources due to limited sea surface excitation conditions to minimize damage to the marine environment and provide stable excitation energy. However, as the depth of the seabed and the depth of the target underground layer increase, the seismic signal emitted by a single airgun loses its energy. As a result, the signal reflected from the underground, received by the geophone, is easily obscured by the bubble response formed by the transient pulse, making it difficult to obtain satisfactory seismic data. To increase the excitation source energy and improve the energy proportion of the effective signal, an airgun array is often used as the excitation source in actual marine seismic acquisition.

[0003] An airgun array arrangement involves arranging multiple single airguns in a specific spatial configuration (including both horizontal and vertical arrangements). The signals emitted by each airgun are then synthesized into a primary excitation wavelet signal based on the airgun's spatial position. Although this array arrangement can effectively enhance excitation energy, because the spatial combination center no longer satisfies the assumption of point source excitation, the superposition response of the signals emitted by each airgun varies at different angles. Ultimately, the synthesized far-field wavelet varies with the superposition angle, exhibiting significant directional dependence. In marine seismic data processing, to improve data quality, it is important to extract seismic wavelets with a high signal-to-noise ratio and stable morphology. However, directional effects, which disrupt the lateral consistency of the wavelet, must be eliminated. Historically, marine seismic exploration has primarily relied on streamer acquisition, which only yields narrow-azimuth seismic data. The directional effect generated by the superposition of far-field wavelets at different angles is less pronounced. One-dimensional signal deconvolution can effectively address this issue. In addition, in recent years, with the development of ocean OBN (Ocean Bottom Nodes) acquisition technology, it has become popular because it can provide all-round, high-density, long-offset data information, which is more helpful to improve the quality of seismic imaging of complex oil and gas reservoirs. However, it also poses more severe challenges to the processing of OBN data. Among them, the directional effect of far-field wavelets has become extremely prominent. In order to solve this problem, three technical difficulties are faced: (1) How to characterize the directionality of far-field wavelets; (2) How to extract far-field wavelets with direction; (3) How to eliminate the directional effect of far-field wavelets. In view of the above three technical difficulties, a method and device that can effectively eliminate the directional effect of far-field wavelets is urgently needed.

[0004] Summary of the Invention

[0005] The purpose of the embodiments of the present invention is to provide a data-driven directional signal deconvolution method, device and readable storage medium to at least solve the problems in the above-mentioned prior art, that is, the directional effect of the far-field wavelet is extremely prominent, the directionality of the far-field wavelet cannot be characterized, the far-field wavelet with direction cannot be extracted; and the directional effect of the far-field wavelet cannot be eliminated.

[0006] In order to achieve the above object, the present invention provides a data-driven direction signal deconvolution method in a first aspect, the method comprising:

[0007] Obtain pre-stack seismic gather data and preset expected wavelets during seismic data acquisition;

[0008] performing a Fourier transform, a three-dimensional τ-p transform, and an inverse Fourier transform on the pre-stack seismic gather data in sequence to obtain first seismic data;

[0009] Performing time difference correction on each seismic trace in the first seismic data in the same plane to obtain second seismic data;

[0010] Projecting the azimuth of each seismic trace in the second seismic data onto a preset bin grid, determining a seismic trace in each bin, and forming first directional seismic data based on the seismic traces of all bins;

[0011] Projecting the azimuth of each seismic trace in the first seismic data onto a preset bin grid and determining a seismic trace in each bin, and forming second directional seismic data based on the seismic traces of all bins;

[0012] Obtaining a direction matching operator based on the first directional seismic data and the preset expected wavelet;

[0013] The second directional seismic data is corrected using the direction matching operator, and a three-dimensional τ-p inverse transformation is performed on the corrected data to obtain final seismic gather data.

[0014] Optionally, performing a Fourier forward transform, a three-dimensional τ-p transform, and an inverse Fourier transform on the pre-stack seismic gather data in sequence to obtain first seismic data includes:

[0015] Based on the shot point coordinates and receiver point coordinates of pre-stack seismic gather data, the offset of each seismic trace is calculated;

[0016] Performing a Fourier forward transform on the pre-stack seismic gather data along a time direction to obtain first seismic sub-data;

[0017] Based on the offset of each seismic trace, the conjugate operator of the horizontal transformation operator and the vertical transformation operator is obtained;

[0018] Performing a three-dimensional τ-p forward transformation on the seismic sub-data based on the conjugate operator to obtain second seismic sub-data;

[0019] Perform an inverse Fourier transform on the second seismic sub-data along the time direction to obtain the first seismic data.

[0020] Optionally, the expression of the conjugate operator of the transverse transformation operator is:

[0021] in, is the conjugate operator of the lateral transformation operator; ω is the circular frequency of the seismic data; offsetx is the offset in the x direction; m is the number of lateral seismic traces; p k is the transverse ray parameter; k is the number of transverse channels in the τ-p domain; i is the imaginary unit;

[0022] The expression of the conjugate operator of the longitudinal transformation operator is:

[0023] in, is the conjugate operator of the lateral transformation operator; ω is the circular frequency of the seismic data; offsety is the offset in the y direction; n is the number of lateral seismic traces; p j is the transverse ray parameter; j is the number of longitudinal channels in the τ-p domain; i is the imaginary unit.

[0024] Optionally, performing time difference correction on each seismic trace in the first seismic data in the same plane to obtain second seismic data includes:

[0025] The following formula is used for time difference correction:

[0026] Among them, M cor (p x ,p y ,t) is the second earthquake data; M(p x ,p y ,t) is the first seismic data; τ0 is the intercept time at zero offset, v is the first interval velocity of the medium; p x and p y is the ray parameter of each seismic trace in the first seismic data; h is the distance from the correction plane to the water surface.

[0027] Optionally, the azimuth angle includes a horizontal azimuth angle and a vertical emission angle, and the method further includes:

[0028] The horizontal azimuth angle and vertical emission angle are calculated using the following formula:

[0029] Among them, α is the horizontal azimuth angle; β is the vertical emission angle; p x and p y is the ray parameter of each seismic trace; v is the layer velocity of the first medium.

[0030] Optionally, obtaining a direction matching operator based on the first directional seismic data and the preset expected wavelet includes:

[0031] Performing wavelet matching on the first directional seismic data and the preset expected wavelet to obtain a matching operator;

[0032] Based on the matching operator, a direction matching operator is obtained using a least squares criterion.

[0033] Optionally, obtaining a direction matching operator based on the first directional seismic data and the preset expected wavelet further includes:

[0034] The direction matching operator is obtained using the following formula: f(α,β,t)*w(t)=w(α,β,t);

[0035] Among them, f(α, β, t) is the direction matching operator; w(α, β, t) is the first directional seismic data; w(t) is the preset expected wavelet.

[0036] Optionally, using the direction matching operator to correct the second directional seismic data, and performing a three-dimensional τ-p inverse transform on the corrected data to obtain final seismic gather data, including:

[0037] The final seismic gather data is obtained using the following formula:

[0038] in, is the final seismic gather data; L x is the transverse operator; M(p xα ,p yβ ,t) is the second directional seismic data; f(α,β,t) is the direction matching operator; L y is the vertical operator.

[0039] A second aspect of the present invention provides a data-driven directional signal deconvolution device, the device comprising:

[0040] A data acquisition module, used to obtain pre-stack seismic gather data and preset expected wavelets during seismic data acquisition;

[0041] a data transformation module, configured to sequentially perform Fourier forward transform, three-dimensional τ-p transform, and inverse Fourier transform on the pre-stack seismic gather data to obtain first seismic data;

[0042] a time difference correction module, configured to perform time difference correction on each seismic trace in the first seismic data in the same plane to obtain second seismic data;

[0043] a first determining module, configured to project the azimuth of each seismic trace in the second seismic data onto a preset bin grid, determine a seismic trace in each bin, and form first directional seismic data based on the seismic traces of all bins;

[0044] A second determining module is configured to project the azimuth of each seismic trace in the first seismic data onto a preset bin grid and determine a seismic trace in each bin, and form second directional seismic data based on the seismic traces of all bins;

[0045] An operator determination module, configured to obtain a direction matching operator based on the first directional seismic data and the preset expected wavelet;

[0046] The data output module is used to modify the second directional seismic data using the direction matching operator and perform a three-dimensional τ-p inverse transform on the modified data to obtain the final seismic gather data.

[0047] On the other hand, the present invention further provides a readable storage medium having instructions stored thereon, wherein the instructions are used to enable a machine to execute the above-mentioned data-driven directional signal deconvolution method.

[0048] This technical solution does not require known prior information or interference from human factors during the seismic data processing process, thus avoiding the problem of low-frequency signals in the time and space domain of seismic data not converging as the offset distance increases. It has high calculation accuracy and is easy to implement, improves the signal-to-noise ratio of seismic data, significantly reduces memory usage and computing costs, and is suitable for processing large amounts of seismic data.

[0049] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following detailed description, they are used to explain the embodiments of the present invention, but do not constitute a limitation of the embodiments of the present invention. In the accompanying drawings:

[0051] FIG1 is a flow chart of a data-driven direction signal deconvolution method provided by the present invention;

[0052] FIG2 is a schematic diagram of a common receiver gather of marine OBN seismic data provided by the present invention;

[0053] FIG3 is a schematic diagram of a preset expected wavelet of a source acquisition design provided by the present invention;

[0054] FIG4 is a schematic diagram of transforming common detection point gathers provided by the present invention into three-dimensional τ-p domain seismic data;

[0055] FIG5 is a schematic diagram of first-arrival correction of three-dimensional τ-p domain seismic data provided by the present invention;

[0056] FIG6 is a schematic diagram of horizontal azimuth information extracted from a three-dimensional τ-p domain provided by the present invention;

[0057] FIG7 is a schematic diagram of vertical emission angle information extracted from a three-dimensional τ-p domain provided by the present invention;

[0058] FIG8 is a schematic diagram showing the display result of projecting the three-dimensional τ-p domain first arrival corrected data onto the angle bins provided by the present invention;

[0059] FIG9 is a schematic diagram of extracting directional far-field wavelets in a three-dimensional τ-p domain provided by the present invention;

[0060] FIG10 is a schematic diagram of an operator after matching a far-field wavelet with a desired wavelet provided by the present invention;

[0061] FIG11 is a schematic diagram of the application of the three-dimensional τ-p domain matching operator provided by the present invention;

[0062] FIG12 is a schematic diagram of data after three-dimensional τ-p inverse transformation after application of the operator provided by the present invention;

[0063] FIG13 is a schematic diagram of seismic wavelet spectrum data for eliminating directional effects provided by the present invention;

[0064] FIG14 is a schematic structural diagram of a data-driven direction signal deconvolution device provided by the present invention.

[0065] Description of Reference Numerals

[0066] 10-data acquisition module; 20-data conversion module; 30-time difference correction module;

[0067] 40-first determination module; 50-second determination module; 60-operator determination module;

[0068] 70-Data output module. DETAILED DESCRIPTION

[0069] The following describes the specific embodiments of the present invention in detail with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only used to illustrate and explain the present invention and are not intended to limit the present invention.

[0070] Figure 1 is a flow chart of the data-driven directional signal deconvolution method provided by the present invention; Figure 2 is a schematic diagram of the common detection point gather of marine OBN seismic data provided by the present invention; Figure 3 is a schematic diagram of the preset expected wavelet of the source acquisition design provided by the present invention; Figure 4 is a schematic diagram of the common detection point gather transformed into three-dimensional τ-p domain seismic data provided by the present invention; Figure 5 is a schematic diagram of the first arrival correction of the three-dimensional τ-p domain seismic data provided by the present invention; Figure 6 is a schematic diagram of the horizontal azimuth information extracted from the three-dimensional τ-p domain provided by the present invention; Figure 7 is a schematic diagram of the vertical take-off angle information extracted from the three-dimensional τ-p domain provided by the present invention; Figure 8 is a schematic diagram of the present invention FIG10 is a schematic diagram of the operator after matching the far-field wavelet with the expected wavelet; FIG11 is a schematic diagram of the application of the three-dimensional τ-p domain matching operator provided by the present invention; FIG12 is a schematic diagram of the three-dimensional τ-p inverse transformed data after the application of the operator provided by the present invention; FIG13 is a schematic diagram of the seismic wavelet spectrum data with the directional effect eliminated provided by the present invention; and FIG14 is a schematic diagram of the structure of the data-driven directional signal deconvolution device provided by the present invention.

[0071] As shown in FIG1 , an embodiment of the present invention provides a data-driven directional signal deconvolution method, the method comprising:

[0072] Step 1: Obtain pre-stack seismic gather data and preset expected wavelets during seismic data acquisition;

[0073] Specifically, in this example, the known pre-stack seismic gathers d(x, y, t) and the preset expected wavelet w(t) from the source acquisition design are used as input data, where x and y are the horizontal and vertical coordinates within the seismic acquisition area, respectively, and t is the traveltime coordinate of the seismic wave propagation. Specifically, Figure 2 shows the input pre-stack common-receiver gathers and Figure 3 shows the expected wavelet from the acquisition design.

[0074] Step 2: performing Fourier forward transform, three-dimensional τ-p transform and inverse Fourier transform on the pre-stack seismic gather data in sequence to obtain first seismic data;

[0075] Specifically, according to the shot point coordinates (s x ,s y ) and the coordinates of the detection point (r x ,r y), where the prestack seismic gather d(x,y,t) is a common receiver gather (the receiver coordinates of each gather are the same, but the shot coordinates are different). The distance from the receiver to each shot is calculated, called the shot offset, and is represented by the vectors offsetx and offsety.

[0076] Perform a fast Fourier transform on each trace in the seismic data set d(x, y, t) along the time direction, and transform the time-space domain data into the frequency-space domain to obtain the seismic data D(x, y, ω), which is used as the first seismic sub-data, where ω is the circular frequency of the seismic data, which is calculated by ω=2πf, and f is the acquisition frequency of the seismic data, which is a known parameter.

[0077] Perform a three-dimensional τ-p transform on the seismic data D(x, y, ω). First, use the vector offsets offsetx and offsety of each trace in step 2) to construct the lateral and longitudinal transform operators of the three-dimensional τ-p transform, which can be expressed as:

[0078] In formula (1), L x is the lateral transformation operator; ω is the circular frequency of the seismic data, m is the number of lateral seismic traces, and p k is the transverse ray parameter (slope), k is the number of transverse channels in the τ-p domain, similarly, in formula (2), L y is the longitudinal transformation operator; ω is the circular frequency of the seismic data, n is the number of longitudinal seismic traces, and p j is the longitudinal ray parameter, j is the number of longitudinal channels in the τ-p domain, and i is the imaginary unit.

[0079] The conjugate operator of the three-dimensional τ-p transform operator of each channel can be calculated from formulas (1) and (2): The calculation formulas are as follows (3) and (4);

[0080] in, is the conjugate operator of the lateral transformation operator; ω is the circular frequency of the seismic data; offsetx is the offset in the x direction; m is the number of lateral seismic traces; p k is the transverse ray parameter; k is the number of transverse channels in the τ-p domain; is the conjugate operator of the lateral transformation operator; ω is the circular frequency of the seismic data; offsety is the offset in the y direction; n is the number of lateral seismic traces; p j is the transverse ray parameter; j is the number of longitudinal channels in the τ-p domain; i is the imaginary unit.

[0081] Using the conjugate operator of each seismic trace obtained above, the seismic data D(x, y, ω) is subjected to a three-dimensional τ-p forward transformation, transforming it from the frequency-space domain to the frequency-slowness domain, i.e., the second seismic sub-data, which can be expressed by the following formula:

[0082] Among them, in formula (5), M(p x ,p y ,ω) is the seismic data volume (the second seismic sub-data) after the three-dimensional τ-p forward transformation, and the seismic data M(p x ,p y ,ω), perform fast inverse Fourier transform along the time direction to obtain the three-dimensional τ-p domain spectrum data M(p x ,p y ,t), namely the first seismic data; as shown in FIG4, it is the spectrum data display of the input pre-stack gather (FIG. 2) transformed into the three-dimensional τ-p domain.

[0083] Step 3: performing time difference correction on each seismic trace in the first seismic data in the same plane to obtain second seismic data;

[0084] Specifically, the three-dimensional τ-p domain seismic data M(p x ,p y ,t) Perform the initial arrival coaxial correction. The expansion process of the bubble energy can be identified by using the flattened coaxial axis. The correction time difference formula for each data is:

[0085] Where, in formula (6), Δτ is the correction time difference of each trace, τ0 is the intercept time at the zero offset position, v is the first layer velocity of the medium, which is generally 1500 m / s water velocity, and h is the distance from the correction plane to the water surface. The correction plane is an artificially defined horizontal plane and is a constant. x ,p y All traces in (t) are corrected to this plane, and the corrected seismic data M is obtained. cor (p x ,p y ,t), which means that the first arrival isotropic axis correction process is achieved. Figure 5 shows the result of the spectral data of the three-dimensional τ-p domain (Figure 4) after the first arrival correction to the horizontal plane. It can be seen that the reverse arc axis is effectively flattened, and it is easy to identify the low-frequency bubble expansion process (the arrow in the figure indicates the position). cor (p x ,p y ,t)=M(p x ,p y ,t)+Δτ (7)

[0086] The final transformation is:

[0087] Among them, M cor (p x ,p y ,t) is the second earthquake data; M(p x ,p y ,t) is the first seismic data; τ0 is the intercept time at the zero offset position, v is the first layer velocity of the medium, generally the water velocity is 1500m / s; p x and p y is the ray parameter of each seismic trace in the first seismic data; h is the distance from the correction plane to the water surface;

[0088] Step 4: Projecting the azimuth of each seismic trace in the second seismic data onto a preset bin grid, determining a seismic trace in each bin, and forming first directional seismic data based on the seismic traces of all bins;

[0089] The first arrival corrected seismic data M in the above steps cor (p x ,p y ,t), that is, the ray parameter p of each seismic trace in the second seismic data x and p y The horizontal azimuth angle α is calculated. Figure 6 shows the azimuth information extracted from the seismic data after first arrival correction (same departure angle, different azimuth angles are displayed).

[0090] Similarly, according to the ray parameter p of each seismic trace x 、p y , and the first layer velocity v of the medium, generally the water velocity is 1500m / s, and the vertical exit angle β of each trace is calculated. Figure 7 shows the exit angle information extracted from the seismic data after first arrival correction (the same azimuth angle, different exit angles are displayed).

[0091] The range of the azimuth angle and the emission angle in formula (8) and formula (9) is [0°~90°], and the azimuth angle α and the emission angle β are divided into surface element grids in the manner of 1°×1°.

[0092] The earthquake data M cor (p x ,p y The azimuth and exit angle of each trace in t) are projected onto the corresponding bin grid defined in the above steps, and the seismic data are rearranged to obtain a new seismic data M with angle information. cor(α, β, t); as shown in Figure 8, it is the display result of projecting seismic data onto angle bins. The phase of the first arrival information in each bin is the same.

[0093] Since the azimuth and take-off angles of the seismic traces in each bin are very small and the first arrival signals have been leveled, the purpose of in-phase superposition and enhancement of effective signals can be achieved. All traces in each bin are superimposed to obtain a unique seismic trace in each bin. In this way, the seismic data in all bins are superimposed to form a directional seismic data, namely the directional far-field wavelet w(α, β, t), which is the first directional seismic data. Figure 9 shows the directional far-field wavelet data in each bin obtained by superimposing the seismic traces in the bin.

[0094] Step 5: Projecting the azimuth of each seismic trace in the first seismic data onto a preset bin grid and determining a seismic trace in each bin, and forming second directional seismic data based on the seismic traces of all bins;

[0095] Specifically, for the seismic data M(p x ,p y ,t), that is, the first seismic data is also divided into the same surface element grid using the same method to obtain the new seismic data M(p xα ,p yβ ,t), i.e., the second directional seismic data. The specific method is the same as the steps of projecting the azimuth of each seismic trace in the second seismic data onto a preset bin grid, determining a seismic trace within each bin, and forming the first directional seismic data based on the seismic traces of all bins, and will not be repeated here.

[0096] Step 6: obtaining a direction matching operator based on the first directional seismic data and the preset expected wavelet;

[0097] Specifically, the obtained preset expected wavelet w(t) and the obtained directional far-field wavelet w(α, β, t), i.e., the first directional seismic data, are used to perform wavelet matching to obtain a matching operator. FIG10 shows the matching operator obtained by wavelet matching. The specific calculation formula is as follows: f(α, β, t)*w(t)=w(α, β, t) (10)

[0098] Then, using the least squares criterion, the direction matching operator f(α, β, t) is obtained from formula (10). The calculation formula is as follows:

[0099] Among them, f(α, β, t) is the direction matching operator; w(α, β, t) is the first directional seismic data; w(t) is the preset expected wavelet.

[0100] Step seven: Use the direction matching operator to correct the second directional seismic data, and perform a three-dimensional τ-p inverse transform on the corrected data to obtain the final seismic gather data.

[0101] The direction matching operator f(α, β, t) determined in formula (11) is applied to the seismic data M(p xα ,p yβ ,t), that is, the second directional seismic data, as follows:

[0102] Seismic data The transverse operator L calculated above x and the longitudinal operator L y Interaction, get three-dimensional τ-p inverse transformation of seismic data That is, the time-space domain seismic data after the directional signal deconvolution is applied, which is the final seismic gather data. The inverse transformation formula is as follows;

[0103] After fusion, we get the final formula:

[0104] in, is the final seismic gather data; L x is the transverse operator; M(p xα ,p yβ ,t) is the second directional seismic data; f(α,β,t) is the direction matching operator; L y is the vertical operator.

[0105] As shown in Figure 12, the comparison results of seismic data before and after the application of directional signal deconvolution are shown. At the positions marked by arrows in the figure, low-frequency interference is effectively eliminated, and the lateral consistency of the seismic wavelet is more continuous and reliable. In Figure 12, the left image is a schematic diagram before the application of directional signal deconvolution technology, and the right image is a schematic diagram after the application of directional signal deconvolution technology; as shown in Figure 13, Data1 (black line) is a schematic diagram of the spectrum analysis of seismic data before the application of directional signal deconvolution, and Data2 (gray line) is a schematic diagram of the spectrum analysis of seismic data after the application of directional signal deconvolution. After the far-field wavelet directionality is eliminated, the spectrum of the seismic data is effectively expanded, and the resolution is significantly improved.

[0106] Through the above-mentioned processing steps, the data-driven directional signal deconvolution process is completed, eliminating the problem of seismic wavelet directionality introduced by the combination of excitation sources. And for the processing of marine OBN wide-azimuth seismic data, this solution proposes a method for eliminating the directional effect of far-field wavelets. Due to the different spatial arrangement positions of the guns in the air gun array, the signal excited by each gun, in the process of superposition to form far-field wavelets, obviously shows a directional effect, which easily causes the lateral consistency of the wavelets to deteriorate, seriously affecting the quality of seismic data. The present application can effectively eliminate the directionality of far-field wavelets, which is of great significance to the processing of OBN data. A completely data-driven method is used, which only requires the input of original seismic data and expected wavelets to complete the directional elimination process of far-field wavelets. It does not require known prior information and human factors to interfere, and has high calculation accuracy and is easy to implement. The azimuth and emission angle information obtained in the three-dimensional τ-p domain avoids the low-frequency signal of seismic data in the time and space domain as the deviation increases. The problem of non-convergence when the shift distance increases is solved. Since the three-dimensional τ-p domain can highlight the low-frequency information more, it is easy to identify the propagation period of low-frequency bubbles and ghost reflections at the shot point end, suppress the two, and thus improve the signal-to-noise ratio of the data; the "first arrival correction" + "small facet grid" method is adopted, among which the first arrival correction achieves the purpose of effective signal in-phase enhancement, and the small facet grid can more accurately calculate the azimuth and exit angle information, which provides high-precision data guarantee for the subsequent synthesis of directional far-field wavelets; for pre-stack gather processing, multi-core parallel computing is realized, and the computing efficiency is fast; at the same time, taking advantage of the fact that pre-stack gathers are not very sensitive to underground structural changes, the matching operator calculated from one gather can be applied to several adjacent gathers, which greatly reduces the memory occupancy and computing cost. This technology is suitable for processing large amounts of seismic data.

[0107] As shown in FIG14 , an embodiment of the present invention further provides a data-driven direction signal deconvolution device, the device comprising:

[0108] The data acquisition module 10 is used to acquire pre-stack seismic gather data and preset expected wavelets during seismic data acquisition;

[0109] A data transformation module 20 is configured to sequentially perform Fourier forward transform, three-dimensional τ-p transform, and inverse Fourier transform on the pre-stack seismic gather data to obtain first seismic data;

[0110] A time difference correction module 30 is used to perform time difference correction on each seismic trace in the first seismic data in the same plane to obtain second seismic data;

[0111] a first determining module 40 for projecting the azimuth of each seismic trace in the second seismic data onto a preset bin grid, determining a seismic trace in each bin, and forming first directional seismic data based on the seismic traces of all bins;

[0112] A second determining module 50 is configured to project the azimuth of each seismic trace in the first seismic data onto a preset bin grid and determine a seismic trace in each bin, and form second directional seismic data based on the seismic traces of all bins;

[0113] An operator determination module 60 is configured to obtain a direction matching operator based on the first directional seismic data and the preset expected wavelet;

[0114] The data output module 70 is used to modify the second directional seismic data using the direction matching operator, and perform a three-dimensional τ-p inverse transform on the modified data to obtain final seismic gather data.

[0115] An embodiment of the present invention further provides a readable storage medium having instructions stored thereon, wherein the instructions are used to enable a machine to execute the above-mentioned data-driven directional signal deconvolution method.

[0116] Those skilled in the art will appreciate that all or part of the steps in the methods of the aforementioned embodiments can be accomplished by instructing the relevant hardware through a program, which is stored in a storage medium and includes a number of instructions for causing a single-chip microcomputer, chip, or processor to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0117] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other and are not used to limit the scope of protection of this application.

[0118] The above describes in detail the optional embodiments of the present invention in conjunction with the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details in the above embodiments. Within the technical concept of the embodiments of the present invention, a variety of simple modifications can be made to the technical solutions of the embodiments of the present invention, and these simple modifications all fall within the scope of protection of the embodiments of the present invention. It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner unless there is any contradiction. In order to avoid unnecessary repetition, the embodiments of the present invention will no longer describe the various possible combinations separately.

[0119] In addition, the various embodiments of the present invention may be arbitrarily combined, and as long as they do not violate the concept of the embodiments of the present invention, they should also be regarded as the contents disclosed in the embodiments of the present invention.

Claims

1. A data-driven directional signal deconvolution method, characterized in that: The method comprises: Obtain pre-stack seismic gather data and preset expected wavelets during seismic data acquisition; Performing Fourier forward transform, three-dimensional τ-p transform and inverse Fourier transform on the pre-stack seismic gather data in sequence to obtain first seismic data; Performing time difference correction on each seismic trace in the first seismic data in the same plane to obtain second seismic data; Projecting the azimuth of each seismic trace in the second seismic data onto a preset bin grid, determining a seismic trace in each bin, and forming first directional seismic data based on the seismic traces of all bins; Projecting the azimuth of each seismic trace in the first seismic data onto a preset bin grid, determining a seismic trace in each bin, and forming second directional seismic data based on the seismic traces of all bins; Obtaining a direction matching operator based on the first directional seismic data and the preset expected wavelet; The second directional seismic data is corrected by using the direction matching operator, and a three-dimensional τ-p inverse transformation is performed on the corrected data to obtain the final seismic gather data.

2. The data-driven directional signal deconvolution method according to claim 1, characterized in that: The pre-stack seismic gather data is sequentially subjected to Fourier forward transform, three-dimensional τ-p transform and inverse Fourier transform to obtain first seismic data, including: Based on the shot point coordinates and receiver point coordinates of pre-stack seismic gather data, the shot offset of each seismic trace is calculated; Performing Fourier forward transform on the pre-stack seismic gather data along the time direction to obtain first seismic sub-data; Based on the offset of each seismic trace, the conjugate operator of the lateral transformation operator and the longitudinal transformation operator is obtained; Performing a three-dimensional τ-p forward transformation on the seismic sub-data based on the conjugate operator to obtain second seismic sub-data; Perform an inverse Fourier transform on the second seismic sub-data along the time direction to obtain the first seismic data.

3. The data-driven directional signal deconvolution method according to claim 2, characterized in that: The expression of the conjugate operator of the transverse transformation operator is: in, is the conjugate operator of the lateral transformation operator; ω is the circular frequency of the seismic data; offsetx is the offset in the x direction; m is the number of lateral seismic traces; p k is the transverse ray parameter; k is the number of transverse channels in the τ-p domain; i is the imaginary unit; The expression of the conjugate operator of the longitudinal transformation operator is: in, is the conjugate operator of the lateral transformation operator; ω is the circular frequency of the seismic data; offsety is the offset in the y direction; n is the number of lateral seismic traces; p j is the transverse ray parameter; j is the number of longitudinal channels in the τ-p domain; i is the imaginary unit.

4. The data-driven directional signal deconvolution method according to claim 1, characterized in that: Performing time difference correction on each seismic trace in the first seismic data in the same plane to obtain second seismic data includes: The following formula is used for time difference correction: Among them, M cor (p x ,p y ,t) is the second earthquake data; M(p x ,p y ,t) is the first seismic data; τ0 is the intercept time at zero offset, v is the first layer velocity of the medium; p x and p y is the ray parameter of each seismic trace in the first seismic data; h is the distance from the correction plane to the water surface.

5. The data-driven directional signal deconvolution method according to claim 1, characterized in that: The azimuth angle includes a horizontal azimuth angle and a vertical emission angle, and the method further includes: The horizontal azimuth angle and vertical emission angle are calculated using the following formula: Among them, α is the horizontal azimuth angle; β is the vertical incidence angle; p x and p y is the ray parameter of each seismic trace; v is the layer velocity of the first medium.

6. The data-driven directional signal deconvolution method according to claim 1, characterized in that: Based on the first directional seismic data and the preset expected wavelet, a direction matching operator is obtained, including: Performing wavelet matching on the first directional seismic data and the preset expected wavelet to obtain a matching operator; Based on the matching operator, a direction matching operator is obtained by using the least squares criterion.

7. The data-driven directional signal deconvolution method according to claim 1, characterized in that: Based on the first directional seismic data and the preset expected wavelet, a direction matching operator is obtained, further comprising: The direction matching operator is obtained using the following formula: f(α,β,t)*w(t)=w(α,β,t); Among them, f(α, β, t) is the direction matching operator; w(α, β, t) is the first directional seismic data; w(t) is the preset expected wavelet.

8. The data-driven directional signal deconvolution method according to claim 1, characterized in that: The second directional seismic data is corrected by using the direction matching operator, and the corrected data is subjected to a three-dimensional τ-p inverse transformation to obtain final seismic gather data, including: The final seismic gather data is obtained using the following formula: in, is the final seismic gather data; L x is the transverse operator; M(p xα ,p yβ ,t) is the second directional seismic data; f(α,β,t) is the direction matching operator; L y is the vertical operator.

9. A data-driven directional signal deconvolution device, characterized in that: The device comprises: A data acquisition module, used to acquire pre-stack seismic gather data and preset expected wavelets during seismic data acquisition; A data transformation module, used for sequentially performing Fourier forward transformation, three-dimensional τ-p transformation and inverse Fourier transformation on the pre-stack seismic gather data to obtain first seismic data; A time difference correction module, used for performing time difference correction on each seismic trace in the first seismic data in the same plane to obtain second seismic data; A first determination module is used to project the azimuth of each seismic trace in the second seismic data onto a preset bin grid, and determine a seismic trace in each bin, and form first directional seismic data based on the seismic traces of all bins; A second determination module is used to project the azimuth of each seismic trace in the first seismic data onto a preset bin grid and determine a seismic trace in each bin, and form second directional seismic data based on the seismic traces of all bins; An operator determination module, used for obtaining a direction matching operator based on the first directional seismic data and the preset expected wavelet; The data output module is used to modify the second directional seismic data using the direction matching operator, and perform a three-dimensional τ-p inverse transformation on the modified data to obtain the final seismic gather data.

10. A readable storage medium having instructions stored thereon, the instructions being used to enable a machine to execute the data-driven directional signal deconvolution method according to any one of claims 1 to 8.

Images