Seismic data fault enhancement method based on optimal local space-time vector

Abstract

The invention provides a seismic data fault enhancement method based on an optimal local space-time vector, and relates to the technical field of seismic data processing, and the method comprises the steps: calculating and correcting a seismic data event inclination angle, and obtaining a corrected event inclination angle; based on the corrected event inclination angle, obtaining a fault enhancement local space-time vector, and calculating a fault enhancement coefficient and a fault enhancement distance truncation threshold; calculating the distance between the center reference waveform vector and the local space-time vector; extracting an optimal local space-time vector according to a distance truncation threshold value; and finally, carrying out weighted stacking on the seismic data by utilizing the optimal local space-time vector so as to enhance seismic fault features. The method effectively overcomes the defects that an existing reflection seismic data extraction and enhancement method is too strong in dependence on seismic fault information and damages seismic faults, and can remarkably enhance discontinuous characteristics of stratums, so that fault planes are clearer, and oil-gas exploration and development requirements are better met.

Description

Technical Field

[0001] The invention relates to the technical field of seismic data processing, and in particular to a seismic data fault enhancement method based on an optimal local space-time vector. Background Art

[0002] The underground fault geological structure affects the storage, sealing and migration of oil and gas resources, and is one of the key information for oil and gas exploration and development. Seismic data uses reflected seismic waves to image underground geological structures, including strata, anticlines, synclines and faults. It is currently the most effective method for studying deep underground fault information. Obtaining seismic data containing high-quality fault information is important for the interpretation of fault geological structures and oil and gas exploration and development.

[0003] Actual seismic data are inevitably disturbed during field acquisition, so reflection data extraction methods are needed to obtain high-quality seismic data during seismic data processing. Existing reflection seismic data extraction methods mainly include linear and nonlinear methods. Linear methods include FX prediction filtering, singular value spectrum analysis, and Radon transform. These methods usually extract reflection seismic signals based on transformation functions with linear signal expression capabilities. Since information such as faults is spatially discontinuous information, this type of linear method will destroy fault information during practical applications. Nonlinear methods include median filtering and anisotropic diffusion filtering. Median filtering can be implemented in two ways: scalar and vector. It has certain edge preservation, but because the median is used in the signal extraction process, the reflection data extraction results have local anomalies. Anisotropic diffusion filtering has the ability to enhance faults, but it needs to provide fault information, which makes the enhancement results heavily dependent on the provided fault information, limiting the application of this type of method. Therefore, in view of the above existing problems, it is important to develop fault-enhanced seismic data processing methods to serve the needs of oil and gas exploration and development. Summary of the invention

[0004] In view of this, the present invention provides a seismic data fault enhancement method based on the optimal local space-time vector, which can enhance the fault information in the seismic data and overcome the defects of the existing reflection seismic data extraction and enhancement methods that are too dependent on seismic fault information and damage to seismic faults.

[0005] To achieve the above objectives, the present invention adopts the following technical solutions:

[0006] A method for fault enhancement of seismic data based on an optimal local space-time vector, the method comprising:

[0007] Calculating the dip correction weight of the event dip of the seismic data, and correcting the event dip to obtain a corrected event dip;

[0008] Based on the corrected event dip, the local space-time vector of fault enhancement is obtained, and the fault enhancement coefficient and the fault enhancement distance cutoff threshold are calculated;

[0009] Calculating the distance between the central reference waveform vector and the local space-time vector;

[0010] According to the distance cutoff threshold, the best local spatiotemporal vector matching the central reference waveform vector is extracted;

[0011] Weighted stacking of seismic data using optimal local space-time vectors to enhance earthquake faults.

[0012] Preferably,

[0013] Calculate the dip correction weight of the event dip of the seismic data, and correct the event dip to obtain the corrected event dip, including:

[0014] Construct a structural tensor matrix based on seismic data and calculate the initial event dip of seismic data;

[0015] Extracting a local space-time vector along the initial event dip direction in the seismic data to obtain a local space-time vector before correction; the local space-time vector before correction has a time direction length and a space direction length;

[0016] At each spatial position of the local space-time vector before correction, the local space-time vector before correction is divided into two parts, and the two parts are superimposed along the spatial direction respectively to generate two waveform vectors corresponding to each spatial position;

[0017] Calculating the Pearson correlation coefficient and the cosine similarity between the two waveform vectors corresponding to each spatial position of the local space-time vector before correction, and calculating the inclination correction weight of the event inclination according to the combination of the two;

[0018] Weighting the Gaussian smoothing function according to the inclination correction weight to obtain a weighted Gaussian smoothing function;

[0019] The weighted Gaussian smoothing function is used to smooth the derivatives of the seismic data in the time direction and the space direction, a new structural tensor is constructed based on the smoothed derivatives, and a preliminary corrected event dip is calculated by performing eigenvalue decomposition on the new structural tensor;

[0020] The weighted Gaussian smoothing function is used to smooth the preliminary corrected event dip to obtain a corrected event dip.

[0021] Preferably,

[0022] Based on the corrected event dip, the local space-time vector of fault enhancement is obtained, and the fault enhancement coefficient and fault enhancement distance cutoff threshold are calculated, including:

[0023] Extracting a local space-time vector along the corrected event dip direction in the seismic data to obtain a fault-enhanced local space-time vector; the fault-enhanced local space-time vector has a time direction length and a space direction length;

[0024] At each spatial position of the fault-enhanced local space-time vector, the fault-enhanced local space-time vector is divided into two parts, and the two parts are superimposed along the spatial direction to generate two waveform vectors corresponding to each spatial position;

[0025] Calculating the Pearson correlation coefficient and the cosine similarity between two waveform vectors corresponding to each spatial position of the fault enhancement local space-time vector, and determining the fault enhancement coefficient according to the calculation results;

[0026] Setting a fault enhancement distance coefficient, and calculating an adaptive fault enhancement distance cutoff threshold according to the cosine similarity between two waveform vectors corresponding to each spatial position of the fault enhancement local space-time vector and the fault enhancement distance coefficient;

[0027] Preferably,

[0028] Calculate the distance between the central reference waveform vector and the local space-time vector, including:

[0029] Preset a regional threshold, and for the space-time position where the fault enhancement coefficient is less than the regional threshold, extract a local space-time vector from the seismic data along the direction of the corrected event dip to obtain a signal enhancement local space-time vector; the signal enhancement local space-time vector has a time direction length and a space direction length;

[0030] Calculating the distance between the signal enhancement local space-time vector and the signal enhancement reference waveform vector to obtain the signal enhancement distance; calculating the distance between the fault enhancement local space-time vector and the fault enhancement reference waveform vector to obtain the fault enhancement distance;

[0031] The signal enhancement reference waveform vector is a vector corresponding to the spatial center position in the signal enhancement local space-time vector; the fault enhancement reference waveform vector is a vector corresponding to the spatial center position in the fault enhancement local space-time vector; the waveform vector refers to a vector corresponding to a specific spatial position in the local space-time vector, and has only a length in the time direction;

[0032] The signal enhancement distance and the fault enhancement distance are normalized to the maximum and minimum values ​​to generate a normalized signal enhancement distance and a normalized fault enhancement distance.

[0033] Preferably,

[0034] According to the distance cutoff threshold, the best local spatiotemporal vector matching the central reference waveform vector is extracted, including:

[0035] For the normalized signal enhancement distance, setting a signal enhancement distance cutoff threshold, and determining a signal enhancement weight according to the signal enhancement distance cutoff threshold;

[0036] For the normalized fault enhancement distance, determining a fault enhancement weight according to the fault enhancement distance cutoff threshold;

[0037] The signal enhancement local space-time vector and the fault enhancement local space-time vector are weighted according to the signal enhancement weight and the fault enhancement weight, and the best fault enhancement local space-time vector and the best signal enhancement local space-time vector matching the central reference waveform vector are extracted.

[0038] Preferably,

[0039] Weighted stacking of seismic data using optimal local space-time vectors to enhance earthquake faults, including:

[0040] At the spatiotemporal position where the fault enhancement coefficient is greater than the regional threshold, a one-dimensional Gaussian smoothing function is used to weightedly superimpose the optimal fault enhancement local spatiotemporal vector along the spatial direction to obtain a fault enhancement waveform vector;

[0041] At the spatiotemporal position where the fault enhancement coefficient is less than the regional threshold, a one-dimensional Gaussian smoothing function is used to weightedly superimpose the optimal signal enhancement local spatiotemporal vector along the spatial direction to obtain a signal enhancement waveform vector;

[0042] The fault enhancement waveform vector and the signal enhancement waveform vector are combined to achieve fault enhancement of seismic data. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0044] Figure 1 A schematic flow chart of a method for fault enhancement of seismic data based on an optimal local space-time vector provided by an embodiment of the present invention;

[0045] Figure 2 Seismic data provided by the embodiment of the present invention;

[0046] Figure 3 For the corresponding Figure 2The seismic data fault enhancement method based on the optimal local space-time vector of the present invention is used to enhance the seismic data fault results. DETAILED DESCRIPTION

[0047] In order to enable those skilled in the art to better understand the technical solution of the present invention, the technical solution of the present invention will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other implementation methods obtained by ordinary technicians in the field without making creative work belong to the scope of protection of the present invention.

[0048] Example

[0049] Figure 1 FIG. 1 is a flow chart of an embodiment of a seismic data fault enhancement method based on an optimal local space-time vector according to the present invention. Figure 1 As shown, the method comprises the following steps:

[0050] Step S1, calculating the dip correction weight of the event dip of the seismic data, and correcting the event dip to obtain a corrected event dip;

[0051] Step S2, based on the corrected event dip, obtaining the fault enhancement local space-time vector, and calculating the fault enhancement coefficient and the fault enhancement distance cutoff threshold;

[0052] Step S3, calculating the distance between the central reference waveform vector and the local space-time vector;

[0053] Step S4, extracting the best local space-time vector matching the central reference waveform vector according to the distance cutoff threshold;

[0054] Step S5: weighted stacking of seismic data using the best local space-time vector to enhance seismic faults.

[0055] Step S1 is specifically as follows:

[0056] The structural tensor matrix is ​​constructed based on the seismic data, and the initial event axis inclination of the seismic data is calculated. Specifically, the seismic data has a sampling time dimension and a spatial distance dimension, such as Figure 2 As shown, the coordinate axis t represents the sampling time in seconds (s), the coordinate axis x represents the spatial distance in meters (m), and the seismic data can be expressed as: d(t,x); a structural tensor matrix is ​​constructed for the seismic data, and eigenvalue decomposition is performed to obtain the initial event axis dip angle θ1(t,x) of the seismic data.

[0057] Set the half length of the local window in the time direction to t h1 (t h1 >0), the half length in the space direction is xh1 (x h1 >0), extract local window data along the initial event dip angle θ1(t,x) direction from all time-space position points (t,x) in the seismic data d(t,x) to obtain the local time-space vector before correction. The formula for obtaining the local time-space vector before correction is as follows:

[0058] d1(t1,x1)=d(t+tan(θ1)x1,x+x1) (1)

[0059] Among them, -t h1 ≤t1≤t h1 , -x h1 ≤x1≤x h1 The length of the local space-time vector before correction in the time direction is 2t h1 , the length in the spatial direction is 2x1.

[0060] At each spatial position x′1 of the local space-time vector before correction, the local space-time vector before correction is divided into two parts, and the two parts are superimposed along the spatial direction to generate two waveform vectors R corresponding to each spatial position. a1 (x′1) and R b1 (x′1). The two waveform vectors R corresponding to each spatial position a1 (x′1) and R b1 (x′1) can be expressed as follows:

[0061]

[0062] Among them, -x h1 <x′1<x h1 .

[0063] Calculate the Pearson correlation coefficient P between the two waveform vectors corresponding to each spatial position of the local space-time vector before correction ab1 (x′1) and cosine similarity S ab1 (x′1), and the tilt correction weight of the event tilt is calculated according to the combination of the two. The tilt correction weight of the event tilt is expressed as:

[0064]

[0065] Among them, t and x represent the position of the center of the local space-time vector in the seismic data before correction, and x h1 is half the spatial length of the local space-time vector before correction, x ang1 (0<x ang1 <x h1 ) is the preset space length, S1 is the similarity threshold, x′1(0<x′1<2xh1 ) is the spatial position dividing point where the local space-time vector before correction is divided into two parts, x′ 1min =argmin(P ab1 ).

[0066] The Gaussian smoothing function is weighted according to the inclination correction weight to obtain a weighted Gaussian smoothing function I s (m,n). The weighted Gaussian smoothing function expression is:

[0067] I s (m,n)=Angw(t+m,x+n)I(m,n) (5)

[0068] Among them, I(m,n) is a two-dimensional Gaussian smoothing function, and the formula is as follows:

[0069]

[0070] The weighted Gaussian smoothing function is used to smooth the derivatives of the seismic data in the time direction and the space direction, a new structural tensor is constructed based on the smoothed derivatives, and a preliminary corrected event axis inclination is calculated by performing eigenvalue decomposition on the new structural tensor.

[0071] Specifically, the derivative of seismic data in the time direction is d t , the derivative in the spatial direction is d x , using the modified Gaussian smoothing function to t d x d t d t d x d x After smoothing, the following results are obtained:

[0072]

[0073] Among them, m a (m a >0) and n a (n a >0) is the range of the weighted Gaussian smoothing function selected for the derivative of smoothed seismic data, C2 is the integral normalization coefficient of the smoothing weight,

[0074] pass <d t d t > s , <d t d x > s , <d x d x > sConstruct a new structure tensor J1(t,x). The new structure tensor can be expressed as:

[0075]

[0076] Perform eigenvalue decomposition on the new structure tensor, extract the maximum eigenvalue λ1 and the minimum eigenvalue λ2, and finally calculate the preliminary revised event axis inclination θ2(t,x). The formula for calculating the preliminary revised event axis inclination is as follows:

[0077]

[0078] The weighted Gaussian smoothing function is used to smooth the preliminary corrected event dip to obtain a corrected event dip. The corrected event dip formula is as follows:

[0079]

[0080] Among them, m b (m b >0) and n b (n b >0) is the range of the weighted Gaussian smoothing function used to obtain the corrected event inclination, C3 is the integral normalization coefficient of the smoothing weight,

[0081] Step S2 is specifically as follows:

[0082] Set the half length of the local window in the time direction to t h2 (t h2 >0), the half length in the space direction is x h2 (x h2 >0), extract the local space-time vector along the corrected event dip angle θ3(t,x) direction from all the space-time position points (t,x) in the seismic data d(t,x) to obtain the fault-enhanced local space-time vector. The fault-enhanced local space-time vector can be expressed as:

[0083] d2(t2,x2)=d(t+tan(θ3)x2,x+x2) (11)

[0084] Among them, -t h2 ≤t2≤t h2 , -x h2 ≤x2≤x h2 The length of the fault-enhanced local space-time vector in the time direction is 2t h2 , with a length of 2x2 in the spatial direction.

[0085] At each spatial position x′2 of the fault-enhanced local space-time vector, the fault-enhanced local space-time vector is divided into two parts, and the two parts are superimposed along the spatial direction to generate two waveform vectors R corresponding to each spatial position. a2 (x′2) and R b2 (x′2). Among them, the two waveform vectors R corresponding to each spatial position a2 (x′2) and R b2 The calculation formula of (x′2) is as follows:

[0086]

[0087] Among them, -x h2 <x′2<x h2 .

[0088] Calculate the two waveform vectors R corresponding to each spatial position of the local space-time vector of the fault enhancement a2 (x′2) and R b2 Pearson correlation coefficient P between (x′2) ab2 (x′2) and cosine similarity S ab2 (x′2), and determine the fault enhancement coefficient based on the calculation results. The formula used to determine the fault enhancement coefficient is as follows:

[0089]

[0090] Where t and x represent the position of the center of the fault enhancement local space-time vector in the seismic data, x h2 is half the spatial length of the fault enhancement local space-time vector, x ang2 (0<x ang2 <x h2 ) is the preset space length, S2 is the similarity threshold, x′2(0<x′2<2x h2 ) is the spatial position dividing point that divides the fault enhancement local space-time vector into two parts, x′ 2min =argmin(P ab2 ).

[0091] The fault enhancement distance coefficient σ1 is set, which affects the selection of the best local space-time vector. The cosine similarity S between the two waveform vectors corresponding to each spatial position of the fault enhancement local space-time vector is calculated according to the ab2 (x′2) and the fault enhancement distance coefficient σ1, calculate the adaptive fault enhancement distance cutoff threshold σ2(t,x), the expression is as follows:

[0092]

[0093] Where L0 = σ1·(max(1-Sab2 )) 0.5 , 0≤λ≤1.

[0094] Step S3 is specifically as follows:

[0095] The preset regional threshold is 0.5. For the spatiotemporal position where the fault enhancement coefficient is less than the regional threshold, the half length of the local window in the time direction is set to t h3 (t h3 >0), the half length in the space direction is x h3 (x h3 >0), extract the local space-time vector from the seismic data along the corrected event dip direction to obtain the signal enhancement local space-time vector d3(t3,x3). The signal enhancement local space-time vector expression is:

[0096] d3(t3,x3)=d(t+tan(θ3)x3,x+x3) (16)

[0097] Among them, -t h3 ≤t3≤t h3 , -x h3 ≤x3≤x h3 The length of the signal enhancement local space-time vector in the time direction is 2t h3 , with a length of 2x3 in the spatial direction.

[0098] The distance between the signal enhancement local space-time vector d3(t3,x3) and the signal enhancement reference waveform vector d3(t3,0) is calculated to obtain the signal enhancement distance D3(x3); the distance between the fault enhancement local space-time vector d2(t2,x2) and the fault enhancement reference waveform vector d2(t2,0) is calculated to obtain the fault enhancement distance D2(x2). The two distance expressions are as follows:

[0099]

[0100] It should be noted that the signal enhancement reference waveform vector is a vector corresponding to the spatial center position within the signal enhancement local space-time vector, and the fault enhancement reference waveform vector is a vector corresponding to the spatial center position within the fault enhancement local space-time vector. The waveform vector refers to a vector in the local space-time vector corresponding to a specific spatial position and has only a length in the time direction.

[0101] The signal enhancement distance D3(x3) and the fault enhancement distance D2(x2) are normalized to the maximum and minimum values ​​to generate a normalized signal enhancement distance D n3 (x3) and normalized fault enhancement distance D n2 (x2). The two normalized distance expressions are as follows:

[0102]

[0103] Among them, ε is a small positive number to prevent the denominator from being 0, D n2 and D n3 The value is between 0 and 1.

[0104] Step S4 is specifically as follows:

[0105] For the normalized signal enhancement distance, a signal enhancement distance cutoff threshold σ3 is set, and a signal enhancement weight is determined according to the signal enhancement distance cutoff threshold. The signal enhancement weight W3(x3) can be expressed as:

[0106]

[0107] For the normalized fault enhancement distance, the fault enhancement weight is determined according to the fault enhancement distance cutoff threshold σ2(t,x). The expression for determining the fault enhancement weight is:

[0108]

[0109] Among them, the weight sum of the signal enhancement local space-time vector d3(t3,x3) is Set the minimum weight and z. If I3 < z, then normalize the distance D n3 Arrange in ascending order, select the signal enhancement local space-time vector corresponding to the previous local space [0, z] as the best local space-time vector, and set the weight to 1.

[0110] It should be noted that W2(x2) and W3(x3) only have the values ​​of 1 and 0. According to the signal enhancement weight W3(x3) and the fault enhancement weight W2(x2), the signal enhancement local space-time vector and the fault enhancement local space-time vector are multiplied respectively to extract the best fault enhancement local space-time vector and the best signal enhancement local space-time vector that match their respective central reference waveform vectors.

[0111] Step S5 is specifically as follows:

[0112] At the spatiotemporal position where the fault enhancement coefficient FP(t,x) is greater than the regional threshold, a one-dimensional Gaussian smoothing function is used to weightedly superimpose the optimal fault enhancement local spatiotemporal vector along the spatial direction to obtain a fault enhancement waveform vector. The calculation method of the fault enhancement waveform vector can be expressed as:

[0113]

[0114] Among them, G(x) is a one-dimensional Gaussian function, W2(x) is the fault enhancement weight, d2(t2,x2) is the fault enhancement local space-time vector, W2(x2)d2(t2,x2) is the optimal fault enhancement local space-time vector, and the one-dimensional Gaussian function formula is as follows:

[0115]

[0116] Here, ρ is the scale of the one-dimensional Gaussian function.

[0117] At the spatiotemporal position where the fault enhancement coefficient FP(t,x) is less than the regional threshold, the optimal signal enhancement local spatiotemporal vector is weightedly superimposed along the spatial direction using a one-dimensional Gaussian smoothing function to obtain a signal enhancement waveform vector. The calculation method of the signal enhancement waveform vector can be expressed as:

[0118]

[0119] Among them, G(x) is a one-dimensional Gaussian function, W3(x) is the signal enhancement weight, d3(t3,x3) is the signal enhancement local space-time vector, and W3(x3)d3(t3,x3) is the optimal signal enhancement local space-time vector.

[0120] Finally, the fault enhancement waveform vectors obtained at all time and space positions where the fault enhancement coefficient FP(t,x) is greater than the regional threshold and the signal enhancement waveform vectors obtained at all time and space positions where the fault enhancement coefficient FP(t,x) is less than the regional threshold are combined to achieve fault enhancement of the seismic data.

[0121] The present invention adopts the above technical scheme, a method for seismic data fault enhancement based on the best local space-time vector, including: calculating the dip correction weight of the event dip of the seismic data, and correcting the event dip to obtain the corrected event dip; based on the corrected event dip, obtaining the fault enhancement local space-time vector, and calculating the fault enhancement coefficient and the fault enhancement distance cutoff threshold; calculating the distance between the central reference waveform vector and the local space-time vector; according to the distance cutoff threshold, extracting the best local space-time vector that matches the central reference waveform vector; using the best local space-time vector to weightedly superimpose the seismic data to enhance the seismic fault. Based on this, by introducing the seismic data fault enhancement method based on the best local space-time vector, the defects of the existing reflection seismic data extraction and enhancement methods that are too dependent on seismic fault information and destroy seismic faults are effectively overcome, and the discontinuity characteristics of the strata can be significantly enhanced, making the fault plane clearer, and meeting the needs of oil and gas exploration and development.

[0122] In a specific implementation example, Figure 2 is the seismic data that has not been processed by the method of the present invention, Figure 3 To correspond to Figure 2The seismic data fault enhancement method based on the optimal local space-time vector of the present invention is used to enhance the seismic data fault. Figure 2 and Figure 3 It can be found that Figure 2 The seismic fault characteristics in the image are relatively vague; after applying the method of the present invention, the seismic fault characteristics are significantly enhanced, the fault plane is clearer, and it is presented as an intuitive and crisp linear structure, which is in sharp contrast with the surrounding continuous strata.

Claims

1. A method for fault enhancement of seismic data based on optimal local space-time vector, characterized in that: include: Calculating the dip correction weight of the event dip of the seismic data, and correcting the event dip to obtain a corrected event dip; Based on the corrected event dip, the local space-time vector of fault enhancement is obtained, and the fault enhancement coefficient and the fault enhancement distance cutoff threshold are calculated; Calculating the distance between the central reference waveform vector and the local space-time vector; According to the distance cutoff threshold, the best local spatiotemporal vector matching the central reference waveform vector is extracted; Weighted stacking of seismic data using optimal local space-time vectors to enhance earthquake faults.

2. The method according to claim 1, characterized in that Calculate the dip correction weight of the event dip of the seismic data, and correct the event dip to obtain the corrected event dip, including: Construct a structural tensor matrix based on seismic data and calculate the initial event dip of seismic data; Extracting a local space-time vector along the initial event dip direction in the seismic data to obtain a local space-time vector before correction; the local space-time vector before correction has a time direction length and a space direction length; At each spatial position of the local space-time vector before correction, the local space-time vector before correction is divided into two parts, and the two parts are superimposed along the spatial direction respectively to generate two waveform vectors corresponding to each spatial position; Calculating the Pearson correlation coefficient and the cosine similarity between two waveform vectors corresponding to each spatial position of the local space-time vector before correction, and calculating the inclination correction weight of the event inclination according to the combination of the two; Weighting the Gaussian smoothing function according to the inclination correction weight to obtain a weighted Gaussian smoothing function; The weighted Gaussian smoothing function is used to smooth the derivatives of the seismic data in the time direction and the space direction, a new structural tensor is constructed based on the smoothed derivatives, and a preliminary corrected event dip is calculated by performing eigenvalue decomposition on the new structural tensor; The weighted Gaussian smoothing function is used to smooth the preliminary corrected event dip to obtain a corrected event dip.

3. The method according to claim 1, characterized in that Based on the corrected event dip, the local space-time vector of fault enhancement is obtained, and the fault enhancement coefficient and fault enhancement distance cutoff threshold are calculated, including: Extracting a local space-time vector along the corrected event dip direction in the seismic data to obtain a fault-enhanced local space-time vector; the fault-enhanced local space-time vector has a time direction length and a space direction length; At each spatial position of the fault-enhanced local space-time vector, the fault-enhanced local space-time vector is divided into two parts, and the two parts are superimposed along the spatial direction to generate two waveform vectors corresponding to each spatial position; Calculating the Pearson correlation coefficient and the cosine similarity between two waveform vectors corresponding to each spatial position of the fault enhancement local space-time vector, and determining the fault enhancement coefficient according to the calculation results; A fault enhancement distance coefficient is set, and an adaptive fault enhancement distance cutoff threshold is calculated according to the cosine similarity between two waveform vectors corresponding to each spatial position of the fault enhancement local space-time vector and the fault enhancement distance coefficient.

4. The method according to claim 1, characterized in that: Calculate the distance between the central reference waveform vector and the local space-time vector, including: Preset a regional threshold, and for the space-time position where the fault enhancement coefficient is less than the regional threshold, extract a local space-time vector from the seismic data along the direction of the corrected event dip to obtain a signal enhancement local space-time vector; the signal enhancement local space-time vector has a time direction length and a space direction length; Calculating the distance between the signal enhancement local space-time vector and the signal enhancement reference waveform vector to obtain the signal enhancement distance; calculating the distance between the fault enhancement local space-time vector and the fault enhancement reference waveform vector to obtain the fault enhancement distance; The signal enhancement reference waveform vector is a vector corresponding to the spatial center position in the signal enhancement local space-time vector; the fault enhancement reference waveform vector is a vector corresponding to the spatial center position in the fault enhancement local space-time vector; the waveform vector refers to a vector corresponding to a specific spatial position in the local space-time vector, and has only a length in the time direction; The signal enhancement distance and the fault enhancement distance are normalized to the maximum and minimum values ​​to generate a normalized signal enhancement distance and a normalized fault enhancement distance.

5. The method according to claim 1, characterized in that According to the distance cutoff threshold, the best local spatiotemporal vector matching the central reference waveform vector is extracted, including: For the normalized signal enhancement distance, setting a signal enhancement distance cutoff threshold, and determining a signal enhancement weight according to the signal enhancement distance cutoff threshold; For the normalized fault enhancement distance, determining a fault enhancement weight according to the fault enhancement distance cutoff threshold; The signal enhancement local space-time vector and the fault enhancement local space-time vector are weighted according to the signal enhancement weight and the fault enhancement weight, and the best fault enhancement local space-time vector and the best signal enhancement local space-time vector matching the central reference waveform vector are extracted.

6. The method according to claim 1, characterized in that Weighted stacking of seismic data using optimal local space-time vectors to enhance earthquake faults, including: At the spatiotemporal position where the fault enhancement coefficient is greater than the regional threshold, a one-dimensional Gaussian smoothing function is used to weightedly superimpose the optimal fault enhancement local spatiotemporal vector along the spatial direction to obtain a fault enhancement waveform vector; At the spatiotemporal position where the fault enhancement coefficient is less than the regional threshold, a one-dimensional Gaussian smoothing function is used to perform weighted superposition along the spatial direction on the optimal signal enhancement local spatiotemporal vector to obtain a signal enhancement waveform vector; The fault enhancement waveform vector and the signal enhancement waveform vector are combined to achieve fault enhancement of seismic data.

Images