Seismic data structure guided low-frequency background model construction method, device and medium

Abstract

The invention provides a seismic data structure guided low-frequency background model construction method, equipment and a medium. The low-frequency background model construction method comprises the following steps: calculating a relative geological age body according to seismic data; performing spatial interpolation on the parameters of the logging data based on the relative geologic age body to obtain an interpolation result; and properly smoothing the interpolation result of the parameter to obtain a low-frequency background model for the parameter. The invention provides a novel seismic data structure-guided low-frequency background model construction method, which comprises the following steps of: firstly, calculating a relative geological age body of seismic data, and then guiding logging data to carry out spatial interpolation by using the relative geological age body so as to obtain a high-resolution interpolation result consistent with a geological structure. And finally, properly smoothing the interpolation result to obtain a low-frequency background model required by inversion. The method can be applied to any parameter on the well, such as longitudinal wave velocity, transverse wave velocity, density, longitudinal and transverse wave velocity ratio, longitudinal wave impedance, transverse wave impedance and the like.

Description

Technical Field

[0001] The present invention relates to the field of oil and gas geophysical exploration, and is a method for constructing a low-frequency background model, an electronic device, and a storage medium, which are based on well logging data and seismic data and use the seismic data structure to guide the spatial interpolation of well logging data. Background Art

[0002] In the process of seismic inversion, the main function of the low-frequency background model is to supplement low-frequency information to more accurately reflect the geological structure underground and the lithological characteristics of the reservoir. Since seismic data lacks low-frequency components, this will affect the interpretation of reservoir lithology and physical properties, resulting in inaccurate reservoir prediction. By constructing a low-frequency background model, reliable low-frequency components can be supplemented to improve the accuracy of seismic inversion results.

[0003] The low-frequency background model plays a very important role in the process of seismic inversion. However, there are still great challenges in constructing a reasonable low-frequency background model based on seismic data and well logging data. In the prior art, the low-frequency background model is usually obtained by spatially interpolating well logging data according to a certain algorithm. Among them, the most commonly used is the Kriging interpolation algorithm, which requires a large number of data samples to derive a suitable covariance model, so the cost is relatively high. In addition, the Kriging interpolation assumes that the data correlation only depends on the distance and direction in physical space. Therefore, when the structure of seismic data changes greatly, it is difficult for this interpolation algorithm to obtain a good low-frequency background model. Another type of method is to extract structure tensor information from seismic data to represent the structure of seismic data, so as to guide the spatial interpolation of well logging data. However, when the well logging data is scarce, the seismic data has noise, or the seismic structure changes greatly, the structure tensor directly extracted from the seismic data may have the risk of misleading the low-frequency background model. Summary of the Invention

[0004] In view of the problems existing in the prior art, the present invention proposes a method for constructing a low-frequency model guided by the seismic data structure. Compared with the prior methods, the method proposed by the present invention can construct a low-frequency background model that better conforms to the spatial variation of the geological structure, including some complex folds and faults, and has better stability and noise resistance.

[0005] To achieve the above object, the present invention provides a method for constructing a low-frequency background model guided by the seismic data structure, including:

[0006] Calculating a relative geological age body according to the seismic data;

[0007] Performing spatial interpolation on the parameters of the well logging data based on the relative geological age body to obtain an interpolation result;

[0008] Appropriately smoothing the interpolation result of the parameters to obtain a low-frequency background model for the parameters.

[0009] Further, calculating the relative geological age body based on seismic data includes:

[0010] Calculating the structure tensor of the seismic data;

[0011] Estimating the reflection dip field based on the structure tensor;

[0012] Calculating all stratigraphic isochrones in the seismic data through the structure tensor and seismic formation dip information to obtain the relative geological age body.

[0013] Further, the spatial interpolation of the parameters of the logging data based on the relative geological age body to obtain the interpolation result includes:

[0014] Projecting the logging data into the interpolation grid space;

[0015] Projecting the relative geological age body into the interpolation grid space and converting the logging data into the relative geological age domain;

[0016] Using the relative geological age body to guide the spatial interpolation of the logging data.

[0017] Further, the structure tensor of the seismic data is used to estimate the seismic structure and formation direction. For three-dimensional seismic data, its structure tensor T(x) is expressed as:

[0018]

[0019] g 1 (x), g 2 (x) and g 3 (x) respectively represent the gradients of the intercepted seismic data in the z direction, x direction, and y direction, and <·> represents applying three-dimensional smoothing filtering to each component of the structure tensor.

[0020] Further, estimating the reflection dip field based on the structure tensor includes calculating the reflection dip fields in the x direction and y direction, respectively:

[0021]

[0022]

[0023] Among them, p(x) represents the dip angle in the x direction, q(x) represents the dip angle in the y direction, and u 1 (x), u 2 (x) and u 3 (x) are respectively the components of the seismic reflection eigenvector μ(x) in the z direction, x direction, and y direction.

[0024] Further, projecting the logging data into the interpolation grid space includes:

[0025] Suppose there are N l wells in a certain work area, and the logging curve L j (j ∈ N l ) of one of the wells consists of N(j) logging data sampling points:

[0026] L j = {(x 1,j , f 1,j ), (x 2,j , f 2,j ),..., (x i,j , f i,j ),..., (x N(j),j , f N(j),j )} (3)

[0027] where, (x i,j , f i,j ) represents the i-th (i ∈ N(j)) logging observation value f j in the logging curve L i,j at x i,j ;

[0028] According to the pre-given interpolation resolution, each logging curve is resampled into the corresponding adjacent regular interpolation grid, so that the vertical resolution of the interpolation model along the logging direction is equal to or less than the original logging observation data:

[0029]

[0030] where represents the resampled logging parameter curve, represents the i-th (i ∈ N(j)) logging observation value in the logging curve at the corresponding adjacent regular interpolation grid ;

[0031] Further, a corresponding relative geological age u(x i,j ) = τ i,j is found at each gridded logging sampling point, and the definition of the logging curve in the relative geological age domain is expressed in the following form:

[0032]

[0033] Further, the relative geological age body is used to guide the spatial interpolation of logging data, and the parameter at any interpolation grid point x is expressed in the following form:

[0034]

[0035] where, r(x) represents the interpolation result of logging data, τi,j = u(z i,j ) represents the relative geological age at the logging sampling point z i,j where N j represents the number of parameter sampling points of the j-th well logging, and g(τ i,j ) = f i,j represents the mapping relationship between the relative geological age and the well logging observation value, and h(τ i,j ) is a one-dimensional weight function used to represent the influence range of the local interpolation field at each logging sampling point in the relative geological age domain. is a two-dimensional weight function related to the spatial distance, and the distance between the interpolation grid point x and the logging sampling point is used as the weight for weighted average. ∈ is a radial parameter factor.

[0036] According to another aspect of the present invention, there is provided an electronic device, which includes:

[0037] a memory storing executable instructions;

[0038] a processor that runs the executable instructions in the memory to implement the method for constructing a low-frequency background model guided by the seismic data structure.

[0039] According to another aspect of the present invention, there is provided a non-transitory computer-readable storage medium, on which a computer program is stored, characterized in that the computer program, when executed by a processor, implements the method for constructing a low-frequency background model guided by the seismic data structure.

[0040] Compared with the existing methods, the method proposed by the present invention can construct a low-frequency background model that better conforms to the spatial variation of geological structures, including some complex folds and faults, and has better stability and anti-noise ability. The method for constructing a low-frequency background model guided by the seismic data structure of the present invention has the following characteristics:

[0041] The present invention uses the method based on seismic dip angle and flattening of the multi-scale seismic trace cross-correlation volume to calculate the flattening offset of the seismic data volume, and then estimates the corresponding three-dimensional relative geological age volume of the seismic data volume according to the offset. The stratigraphic characteristics in the relative geological age volume can well match the seismic reflection axis information.

[0042] The present invention uses the relative geological age volume to guide the spatial interpolation of well logging data. The interpolation result can be quickly obtained by inverse distance interpolation between wells along the equal geological time, and the low-frequency background model required for seismic inversion can be obtained by appropriately smoothing the interpolation result.

[0043] The method of the present invention can be applied to any parameter on the well, such as P-wave velocity, S-wave velocity, density, P-wave to S-wave velocity ratio, P-wave impedance, S-wave impedance, etc. Description of the Drawings

[0044] The above and other objects, features, and advantages of the present invention will become more apparent by describing the exemplary embodiments of the present invention in more detail in conjunction with the accompanying drawings, wherein, in the exemplary embodiments of the present invention, the same reference numerals generally represent the same components.

[0045] Figure 1 It is a flowchart of a method for constructing a low-frequency background model guided by a seismic data structure according to the present invention.

[0046] Figure 2 It is a flowchart of a method for constructing a low-frequency background model of longitudinal wave impedance according to an embodiment of the present invention.

[0047] Figure 3 It is a schematic diagram of 3D actual seismic data and well logging data of longitudinal wave impedance according to an embodiment of the present invention.

[0048] Figure 4 It is a schematic diagram of a relative geological body according to an embodiment of the present invention.

[0049] Figure 5 It is a schematic diagram of an isogeochron extracted based on a relative geological body according to an embodiment of the present invention.

[0050] Figure 6 It is a schematic diagram of longitudinal wave impedance obtained by well-to-well interpolation guided by a seismic data structure according to an embodiment of the present invention. Detailed Description of the Embodiments

[0051] The preferred embodiments of the present invention will be described in more detail below. Although the preferred embodiments of the present invention are described below, 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 to make the present invention more thorough and complete, and to convey the scope of the present invention fully to those skilled in the art.

[0052] The present invention discloses a method for constructing a low-frequency background model. Mainly aiming at the problem that when the traditional method has large structural changes or low signal-to-noise ratio in seismic data, the low-frequency background model obtained by spatially interpolating well logging data has large errors and has many artifacts and noise features. The present invention proposes a new method for constructing a low-frequency background model guided by a seismic data structure. This method first needs to calculate the relative geological body of the seismic data, and then use the relative geological body to guide the spatial interpolation of the well logging data, so as to obtain a high-resolution interpolation result consistent with the geological structure. Finally, the interpolation result is appropriately smoothed to obtain the low-frequency background model required for inversion.

[0053] Embodiment 1

[0054] AsFigure 1 As shown in Figure 1 , the present invention provides a method for constructing a low-frequency background model guided by a seismic data structure, including:

[0055] Calculating a relative geological age body based on seismic data;

[0056] Performing spatial interpolation on the parameters of well logging data based on the relative geological age body to obtain an interpolation result;

[0057] Appropriately smoothing the interpolation result of the parameters to obtain a low-frequency background model for the parameters.

[0058] Further, calculating the relative geological age body based on seismic data includes:

[0059] Calculating the structure tensor of seismic data;

[0060] Estimating a reflection dip field based on the structure tensor;

[0061] Calculating all stratigraphic isochrones in the seismic data through the structure tensor and seismic formation dip information to obtain a relative geological age body.

[0062] Further, performing spatial interpolation on the parameters of well logging data based on the relative geological age body to obtain an interpolation result includes:

[0063] Projecting well logging data into the interpolation grid space;

[0064] Projecting the relative geological age body into the interpolation grid space and converting the well logging data into the relative geological age domain;

[0065] Guiding the spatial interpolation of well logging data using the relative geological age body.

[0066] Further, the structure tensor of the seismic data is used to estimate the seismic structure and formation direction. For three-dimensional seismic data, its structure tensor T(x) is expressed as:

[0067]

[0068] g 1 (x), g 2 (x), and g 3 (x) respectively represent the gradients of the intercepted seismic data in the z direction, x direction, and y direction, and <•> represents applying three-dimensional smoothing filtering to each component of the structure tensor.

[0069] Further, estimating the reflection dip field based on the structure tensor includes calculating the reflection dip fields in the x direction and y direction, which are respectively:

[0070]

[0071]

[0072] Among them, p(x) represents the dip angle in the x direction, q(x) represents the dip angle in the y direction, and u 1 (x), u 2 (x), and u 3 (x) are the components of the seismic reflection feature vector μ(x) along the z direction, x direction, and y direction, respectively.

[0073] Furthermore, projecting the logging data into the interpolation grid space includes:

[0074] Assume that there are N l wells in a certain work area, and the parameter logging curve L j (j ∈ N l ) of one of the wells consists of N(j) logging data sampling points:

[0075] L j = {(x 1,j , f 1,j ), (x 2,j , f 2,j ),..., (x i,j , f i,j ),..., (x N(j),j , f N(j),j )} (3)

[0076] Among them, (x i,j , f i,j ) represents the i-th (i ∈ N(j)) logging observation value at x j in the logging curve L i,j ;

[0077] According to the pre-given interpolation resolution, each logging curve is resampled into the corresponding adjacent regular interpolation grid so that the vertical resolution of the interpolation model along the logging direction is equal to or less than the original logging observation data:

[0078]

[0079] Among them represents the resampled logging parameter curve, represents the i-th (i ∈ N(j)) logging observation value at the corresponding adjacent regular interpolation grid ;

[0080] Furthermore, find a corresponding relative geological age u(x i,j ) = τ i,j at each gridded logging sampling point, and the definition of the logging curve in the relative geological age domain is expressed in the following form:

[0081]

[0082] Furthermore, the relative geological age body is used to guide the spatial interpolation of logging data, and the parameter at any interpolation grid point x is expressed in the following form:

[0083]

[0084] where r(x) represents the interpolation result of logging data, and τ i,j = u(z i,j ) represents the relative geological age at the logging sampling point z i,j , N j represents the number of parameter sampling points of the j-th logging, and g(τ i,j ) = f i,j represents the mapping relationship between the relative geological age and the logging observation value. h(τ i,j ) is a one-dimensional weight function used to represent the influence range of the local interpolation field at each logging sampling point in the relative geological age domain. is a two-dimensional weight function related to the spatial distance, and weighted averaging is performed with the distance between the interpolation grid point x and the logging sampling point as the weight, and ∈ is a radial parameter factor.

[0085] Example Two

[0086] This embodiment provides a method for constructing a low-frequency background model guided by seismic data structures. This method can be applied to the construction of low-frequency background models for any parameter (such as P-wave velocity, S-wave velocity, density, P-to-S wave velocity ratio, P-wave impedance, S-wave impedance...) on the well. Hereinafter, the construction of the low-frequency background model of P-wave impedance will be taken as an example to illustrate the content of the present invention.

[0087] The method provided in this embodiment includes three steps in total:

[0088] (1) Calculate the relative geological age body according to the seismic data;

[0089] (2) Perform spatial interpolation on the logging data based on the relative geological age body to obtain the interpolation result of the P-wave impedance;

[0090] (3) Appropriately smooth the interpolation result of the P-wave impedance to obtain the low-frequency background model of the P-wave impedance.

[0091] The detailed process of step (1) is as follows:

[0092] ① Calculate the structure tensor of the seismic data, and this structure tensor can be used to estimate the seismic structure and formation direction. For three-dimensional seismic data, its structure tensor x(x) is expressed as:

[0093]

[0094] g 1 (x), g 2 (x), and g 3 (x) represent the gradients of the intercepted seismic data in the z - direction, x - direction, and y - direction respectively. This structure tensor has 6 independent components, and each component is a three - dimensional seismic image with the same size as the intercepted seismic data. <·> represents applying a three - dimensional smoothing filter to each component of the structure tensor.

[0095] ② Estimate the reflection dip field based on the structure tensor. Assuming that the eigenvector μ(x) of the seismic reflection is always downward, we can calculate the reflection dip fields in the x - direction and y - direction respectively as follows:

[0096]

[0097]

[0098] where, u 1 (x), u 2 (x), and u 3 (x) are the components of the seismic reflection eigenvector μ(x) along the z - direction (vertical direction), x - direction (inline direction), and y - direction (crossline direction) respectively.

[0099] ③ Calculate all the stratigraphic isochrones in the seismic data through the structure tensor and the seismic formation dip information, and then obtain an implicit scalar field, that is, the relative geological age body.

[0100] The detailed process of step (2) is as follows:

[0101] ① Project the logging data into the interpolation grid space: Assume that there are N l wells in a certain work area, and the P - wave impedance logging curve L j (j ∈ N l ) of one of the wells consists of N(j) logging data sampling points:

[0102] L j = {(x 1,j , f 1,j ), (x 2,j , f 2,j ),..., (x i,j , f i,j ),..., (x N(j),j , f N(j),j )} (3)

[0103] where, (x i,j , f i,j ) represents a point on the logging curve Lj The i-th (i ∈ N(j)) logging observation value f at x i,j in it. According to the pre-given interpolation resolution, we resample each logging curve into the corresponding adjacent regular interpolation grid, so that the vertical resolution of the interpolation model along the logging direction is equal to or less than the original logging observation data: i,j In this gridding process, since the spatial sampling rate of the logging data is greater than or equal to the interpolation resolution, there may be multiple logging sampling points within a single interpolation grid. We only retain the sampling points corresponding to the median value of the logging observations.

[0104]

[0105] ② Project the relative geological age body obtained in step (1) into the interpolation grid space and transform the logging data into the relative geological age domain: After calculating a relative geological age body from the seismic data volume, we project this relative geological age body into the interpolation grid space to make its spatial resolution consistent with that of the gridded logging data. We find a corresponding relative geological age u(x

[0106] ) = τ i,j at each gridded logging sampling point. Therefore, the definition of the logging curve in the relative geological age domain can be expressed in the following form: i,j Therefore, formula 5 strictly constrains the interpolation process to be carried out within the same seismic stratigraphic layer with equal relative geological ages. Since the relative geological age body contains all the structural information in the seismic data, the relative geological age body can be used to realize well-to-well interpolation guided by the seismic data structure.

[0107]

[0108] ③ Use the relative geological age body to guide the spatial interpolation of the logging data: For each logging curve

[0109] This invention uses the linear interpolation method to find the mapping relationship g(τ ) = f i,j between the relative geological age and the logging observation value, and estimates the P-wave impedance in the area without logging data coverage along the equal relative geological ages. In addition, we define a one-dimensional weight function along the logging curve to highlight the role of the original logging data sampling points in the interpolation process, set it to 1 at its corresponding spatial coordinates, and set it to 0 at the local absence of the logging observation values. After being gridded, this one-dimensional weight function can be used to balance the influence range h(τ i,j of the local interpolation field at each logging sampling point in the relative geological age domain i,j)。To further control the interpolation process within relatively equivalent geological age horizons, we introduce a two-dimensional weight function related to spatial distance, using a radial basis function to adjust the importance of different logging sampling points, which depends on the distance between the sampling point and the interpolation grid point along the horizontal section or tangent. In terms of the selection of the radial basis function, the present invention uses an inverse square distance weighting function to perform weighted averaging with the distance between the interpolation grid point x and the logging sampling point as the weight. Sampling points closer to the interpolation point are given greater weights:

[0110]

[0111] where ∈ is a radial parameter factor used to balance the influence degree of this distance weight term on the interpolation process. According to the results of multiple preliminary experiments, we set it to 1. The present invention adds a constant value 1 to the denominator of formula 6 to avoid numerical calculation instability caused by the complete coincidence of the interpolation point and the sampling point. This weight function is based on a basic assumption that within the same formation, the longitudinal wave impedance of the reservoir changes continuously, and the longitudinal wave impedance at any grid point always has the greatest similarity with the logging sampling point closest to it, and this similarity decreases as the distance between the two increases. Therefore, by laterally spreading the logging observation data along the equivalent geological age horizons, the longitudinal wave impedance at any interpolation grid point x can be expressed in the following form:

[0112]

[0113] where τ i,j = u(z i,j ) represents the relative geological age at the logging sampling point z i,j , and N j represents the number of longitudinal wave impedance sampling points of the j-th well logging.

[0114] Example 3

[0115] Referring to Figures 3 - 6 as shown, this example takes the data of a certain actual work area as an example, and uses the method provided by the present invention to construct a low-frequency background model of longitudinal wave impedance. It should be noted that the method provided in this example can not only be applied to the construction of the low-frequency background model of longitudinal wave impedance, but also can be applied to the construction of the low-frequency background model of any parameter (such as longitudinal wave velocity, shear wave velocity, density, longitudinal-to-shear wave velocity ratio, shear wave impedance...) on the well.

[0116] Figure 3 shows the three-dimensional actual seismic data and the logging data of longitudinal wave impedance. The longitudinal wave impedance values are represented in color. There are a total of 6 wells in this work area, among which only 1 well is a vertical well, and the remaining 5 wells are all deviated wells.

[0117] Figure 4 The relative geological age body is calculated based on seismic data. First, the flattened offset of the seismic data volume is calculated using the method based on seismic dip and cross-correlation flattening of multi-scale seismic traces. Then, the relative geological age body of the seismic data is estimated based on the offset.

[0118] Figure 5 The isochrones extracted from the relative geological age body can also be regarded as isogeochronal lines. It can be seen that these isochrones can well fit the seismic horizons, which reflects that the relative geological age body can accurately reflect all the stratigraphic and structural information in the seismic data volume, and thus can provide a fine structural guidance for the next cross-well interpolation of P-wave impedance.

[0119] Figure 6 The P-wave impedance obtained by cross-well interpolation guided by the seismic data structure. Guided by the seismic data structure information reflected by the relative geological age body, the P-wave impedance of 6 wells is spatially interpolated. The interpolation result of the P-wave impedance can be quickly obtained by inverse distance weighting along the isogeochronal lines. This result shows a high vertical resolution and fits well with the seismic structure horizontally. By appropriately smoothing the interpolation result, the low-frequency background model of the P-wave impedance required for inversion can be obtained.

[0120] Example 4

[0121] This example provides an electronic device, which includes:

[0122] A memory storing executable instructions;

[0123] A processor that runs the executable instructions in the memory to implement the above-mentioned method for constructing a low-frequency background model guided by seismic data structure. The method includes:

[0124] Calculating a relative geological age body according to seismic data;

[0125] Performing spatial interpolation on the parameters of well logging data based on the relative geological age body to obtain an interpolation result;

[0126] Appropriately smoothing the interpolation result of the parameters to obtain a low-frequency background model for the parameters.

[0127] Example 5

[0128] This example provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the above-mentioned method for constructing a low-frequency background model guided by seismic data structure. The method includes:

[0129] Calculating a relative geological age body according to seismic data;

[0130] Spatial interpolation is performed on the parameters of the logging data based on the relative geological age body to obtain an interpolation result;

[0131] The interpolation result of the parameter is appropriately smoothed to obtain a low-frequency background model for the parameter.

[0132] The above computer-readable storage medium includes but is not limited to: optical storage media (e.g., CD-ROM and DVD), magneto-optical storage media (e.g., MO), magnetic storage media (e.g., magnetic tape or external hard drive), media with built-in rewritable non-volatile memory (e.g., memory card), and media with built-in ROM (e.g., ROM cartridge).

[0133] In summary, the present invention proposes a new method for constructing a low-frequency background model guided by a seismic data structure. This method first needs to calculate the relative geological age body of the seismic data, and then uses the relative geological age body to guide the spatial interpolation of the logging data, so as to obtain a high-resolution interpolation result consistent with the geological structure. Finally, the interpolation result is appropriately smoothed to obtain the low-frequency background model required for inversion.

[0134] The embodiments of the present invention have been described above. The above description is exemplary and not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations are obvious to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments.

Claims

1. A method for constructing a low-frequency background model guided by seismic data structure, characterized in that, it includes: Calculating a relative geological age body from seismic data; Performing spatial interpolation on the parameters of well logging data based on the relative geological age body to obtain an interpolation result; Appropriately smoothing the interpolation result of the parameters to obtain a low-frequency background model for the parameters.

2. The method for constructing a low-frequency background model guided by seismic data structure according to claim 1, characterized in that, Calculating a relative geological age body from seismic data includes: Calculating the structure tensor of seismic data; Estimating the reflection dip field based on the structure tensor; Calculating all stratigraphic isochrones in the seismic data through the structure tensor and seismic formation dip information to obtain a relative geological age body.

3. The method for constructing a low-frequency background model guided by seismic data structure according to claim 1, characterized in that, Performing spatial interpolation on the parameters of well logging data based on the relative geological age body to obtain an interpolation result includes: Projecting well logging data into the interpolation grid space; Projecting the relative geological age body into the interpolation grid space and converting the well logging data into the relative geological age domain; Using the relative geological age body to guide the spatial interpolation of well logging data.

4. The method for constructing a low-frequency background model guided by seismic data structure according to claim 2, characterized in that, The structure tensor of the seismic data is used to estimate the seismic structure and formation direction. For three-dimensional seismic data, its structure tensor T(x) is expressed as: g 1 (x), g 2 (x) and g 3 (x) represent the gradients of the intercepted seismic data in the z, x, and y directions, respectively. <·> represents applying three-dimensional smoothing filtering to each component of the structure tensor.

5. The method for constructing a low-frequency background model guided by seismic data structure according to claim 4, characterized in that, Estimating the reflection dip field based on the structure tensor includes calculating the reflection dip fields in the x direction and y direction, respectively: where p(x) represents the inclination angle in the x direction, q(x) represents the inclination angle in the y direction, and u 1 (x), u 2 (x), and u 3 (x) are the components of the seismic reflection feature vector μ(x) along the z direction, x direction, and y direction, respectively.

6. The method for constructing a low-frequency background model guided by seismic data structure according to claim 3, characterized in that, Projecting well logging data into the interpolation grid space includes: Suppose there are N l wells in a certain work area, and the parameter logging curve L j (j ∈ N l ) of one of the wells consists of N(j) logging data sampling points: L j = {(x 1，j , f 1，j ), (x 2，j , f 2，j ),..., (x i，j , f i，j ),..., (x N(j)，j , f N(j)，j )} (3) Among them, (x i,j , f i,j ) represents the i-th (i ∈ N(j)) logging observation value f j in the logging curve L i,j at the position of x i,j ; According to a pre-given interpolation resolution, resampling each well logging curve into the corresponding adjacent regular interpolation grid so that the vertical resolution of the interpolation model along the well logging direction is equal to or less than the original well logging observation data: wherein represents the parameter logging curve after resampling, represents the i-th (i ∈ N(j)) logging observation value at the corresponding adjacent regular interpolation grid in the logging curve, 7. The method for constructing a low-frequency background model guided by seismic data structure according to claim 6, characterized in that, Find a corresponding relative geological age u(x i,j ) = T i,j at each grid-based logging sampling point. The definition of the logging curve in the relative geological age domain is expressed in the following form:

8. The method for constructing a low-frequency background model guided by seismic data structure according to claim 7, characterized in that, Using the relative geological age body to guide the spatial interpolation of well logging data, and the parameter at any interpolation grid point x is expressed in the following form: Among them, r(x) represents the interpolation result of well logging data, and τ i,j = u(z i,j ) represents the relative geological age at the well logging sampling point z i,j , N j represents the number of parameter sampling points of the j-th well logging, and g(τ i,j ) = f i,j represents the mapping relationship between the relative geological age and the well logging observation value, and h(τ i,j ) is a one-dimensional weight function used to represent the influence range of the local interpolation field at each well logging sampling point in the relative geological age domain. is a two-dimensional weight function related to the spatial distance, and the distance between the interpolation grid point x and the well logging sampling point is used as the weight for weighted average. ∈ is a radial parameter factor.

9. An electronic device, characterized in that, the electronic device includes: A memory storing executable instructions; A processor that runs the executable instructions in the memory to implement the method for constructing a low-frequency background model guided by seismic data structure according to any one of claims 1-8.

10. A non-transitory computer-readable storage medium, on which a computer program is stored, characterized in that, when the computer program is executed by a processor, it implements the method for constructing a low-frequency background model guided by seismic data structure according to any one of claims 1-8.

Images