Fault attribute enhancement method based on matched filtering, electronic equipment and storage medium
By using matched filtering and dynamic time warping algorithms, the fault direction is estimated and the optimal path is traced, which solves the problems of low fault identification accuracy and low signal-to-noise ratio, and achieves enhanced fault features and clear continuity.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-30
- Publication Date
- 2026-03-31
AI Technical Summary
Existing fault attribute analysis techniques are affected by analysis parameters and data noise, making it difficult to improve the accuracy and signal-to-noise ratio of fault identification, and the spatial characteristics of faults are unclear and have poor continuity.
The matching filter method is used to estimate the main direction of fault attribute data, automatically find seed points and define local windows, thereby tracing the optimal fault path along the main fault direction, and enhancing fault characteristics through nonlinear smoothing and dynamic time warping algorithms.
It enhances the clarity and continuity of fault features, improves the signal-to-noise ratio of fault identification, and makes azimuth and dip estimation more accurate, which helps in automatic fault extraction.
Smart Images
Figure CN121763394A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of seismic data structural interpretation, and more specifically, to a method for enhancing fault attributes based on matched filtering, an electronic device, and a storage medium. Background Technology
[0002] Currently, researchers in this field have made great efforts to accurately describe fracture systems, proposing and applying many descriptive methods. They have also successfully extracted many discontinuous attribute volumes that highlight fault information from 3D seismic data, such as fault slicing technology, edge enhancement attribute technology, coherence volume attribute technology, variance volume attribute technology, dip angle and azimuth attribute, and maximum likelihood attribute technology. The introduction of these fault attribute analysis techniques has enabled fault identification to reach a high level.
[0003] However, the identification effects of the above-mentioned conventional fault attribute analysis techniques are affected to varying degrees by the selected analysis parameters and data noise. Moreover, it is impossible to fundamentally eliminate the influence of noise and human parameter selection. Therefore, it is urgent to develop fault attribute enhancement technology to process the fault identification results and improve the accuracy of fault identification. Summary of the Invention
[0004] The purpose of this invention is to propose a fault attribute enhancement method, electronic device, and storage medium based on matched filtering, which can improve the signal-to-noise ratio of fault identification results, highlight fault spatial features, and enhance the spatial continuity of fault identification results.
[0005] To achieve the above objectives, in a first aspect, the present invention proposes a method for enhancing tomographic attributes based on matched filtering, comprising:
[0006] Acquire fault attribute data, which includes a two-dimensional fault attribute image and a three-dimensional fault attribute data volume;
[0007] The main fault orientation of the fault attribute data is estimated using a matched filtering method.
[0008] Seed points are automatically found from all sample points of the fault attribute data, and a local window is defined with the seed points as the center.
[0009] Using the seed point as a control point, the optimal fault path is traced from the fault attribute data within the local window along the estimated main fault direction, thereby enhancing the fault attribute features.
[0010] Optionally, estimating the main fault direction of the fault attribute data using a matched filtering method includes:
[0011] Smooth each sample point in the fault attribute data along all possible fault directions, find the maximum smoothed response of each sample point and use it as the attribute value of that sample point, and record the direction that produces the maximum smoothed response as the estimated main fault direction at that sample point.
[0012] Optionally, when the fault attribute data is a two-dimensional fault attribute image, for each sample point in the two-dimensional fault attribute image, recursive Gaussian smoothing is performed along all fault directions, and smoothing values in different directions are accumulated in the smoothing volume.
[0013] Record the maximum smooth value of each sample point to obtain the maximum smooth value volume of the entire data. At the same time, record the azimuth volume and dip volume corresponding to the maximum smooth value. The azimuth volume and dip volume represent the azimuth and dip estimates of the fault attribute data, which are used to estimate the main fault direction.
[0014] Optionally, when the fault attribute data is a three-dimensional fault attribute data volume, the fault direction is determined by both the dip angle and the azimuth angle;
[0015] Find the fault azimuth and dip combination that produces the maximum smoothness value by following different plane smoothing fault property data defined by all possible combinations of azimuth and dip.
[0016] Optionally, automatically finding seed points from all sample points of the fault attribute data includes:
[0017] Collect all sample points whose attribute values are greater than or equal to a set threshold;
[0018] Seed points are determined according to the attribute values from highest to lowest.
[0019] When determining the final seed point, the distance between the current sample point and all the already determined seed points is calculated. The current sample point is determined as a seed point if and only if the minimum distance is greater than or equal to the predefined distance.
[0020] Optionally, when calculating seed points in a two-dimensional tomographic attribute image, before collecting all sample points with attribute values greater than or equal to a set threshold, the method further includes:
[0021] The non-maximum suppression method is applied to the original fault attribute image to calculate the refined fault attribute image. In the non-maximum suppression method, only the attribute values perpendicular to the estimated fault direction are retained, and the attribute values in other places are set to zero to obtain the refined fault attribute image.
[0022] Optionally, defining a local window centered on the seed point includes:
[0023] When the fault attribute data is a two-dimensional fault attribute image, for each seed point, a local rectangular window is defined. This window is centered on the seed point and aligned with the uv space coordinates. The u-axis and v-axis are perpendicular to and parallel to the estimated main fault direction, respectively.
[0024] When the fault attribute data is a three-dimensional fault attribute data volume, a local 3D window is defined for each seed point. The local 3D window is centered on the seed point and is spatially aligned with the estimated azimuth and dip angle of the seed point.
[0025] Optionally, tracing the optimal fault path from the fault attribute data within the local window along the estimated main fault direction includes:
[0026] The following constraint maximization problem is solved by a three-step dynamic time warping algorithm consisting of nonlinear smoothing, forward accumulation, and backward tracking, to find the optimal fault path and enhance fault attribute features.
[0027]
[0028] |j[i+1]-j[i]|≤ε(0<ε≤1)
[0029] In the formula, j[i] represents the optimal fault path, g[i,j] represents the fault attribute image, ε represents the set threshold, and the second formula represents the fault slope constraint.
[0030] Secondly, the present invention provides an electronic device, the electronic device comprising:
[0031] At least one processor; and,
[0032] A memory communicatively connected to the at least one processor; wherein,
[0033] The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the matched filtering-based tomographic attribute enhancement method described in the first aspect.
[0034] Thirdly, the present invention proposes a non-transitory computer-readable storage medium storing computer instructions for causing a computer to execute the matched filtering-based tomographic attribute enhancement method described in the first aspect.
[0035] The beneficial effects of this invention are as follows:
[0036] The method of this invention uses matched filtering to enhance fault features. First, it smooths each sample point in the fault attribute data along all possible fault directions, finds the maximum smooth response of each sample point and uses it as the attribute value of that point. The direction that produces the maximum smooth response is recorded as the estimated main fault direction at that sample point. Then, by automatically calculating seed points and tracing the optimal fault path along the estimated fault direction, the fault features are enhanced. Compared with the fault attribute data before enhancement, the data processed by this method has noise suppressed, the fault features become clearer and more continuous, and the fault plane is easier to trace. In addition, the azimuth and dip angles of the fault can be accurately estimated based on the maximum smooth response, which helps in the overall fault analysis and automatic extraction of the fault plane.
[0037] The system of the present invention has other features and advantages that will be apparent from or will be set forth in detail in the accompanying drawings and following detailed description, which together serve to explain the particular principles of the invention. Attached Figure Description
[0038] The above and other objects, features and advantages of the present invention will become more apparent from the accompanying drawings, in which like reference numerals generally denote like parts.
[0039] Figure 1 A flowchart illustrating the steps of a tomographic property enhancement method based on matched filtering according to the present invention is shown.
[0040] Figures 2a-2c The enhanced fault property volume, dip volume, and azimuth volume are shown respectively.
[0041] Figure 3a and Figure 3b The original fault properties and the enhanced fault properties are shown respectively. Detailed Implementation
[0042] This invention proposes a fault attribute enhancement method based on matched filtering. Fault information identified based on three-dimensional fault attribute data suffers from low signal-to-noise ratio, unclear fault features, and poor continuity. To highlight fault spatial features and improve fault spatial continuity, a fault feature enhancement technology is developed. This invention introduces a fault attribute enhancement method based on dynamic time warping. By processing fault attribute data, the influence of noise is reduced, thereby improving the signal-to-noise ratio of fault identification results, highlighting fault spatial features, and enhancing fault spatial continuity.
[0043] The invention will now be described in more detail with reference to the accompanying drawings. While preferred embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that the invention will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.
[0044] Example 1
[0045] like Figure 1 As shown, this embodiment provides a method for enhancing tomographic attributes based on matched filtering, including:
[0046] S1: Obtain fault attribute data, which includes a two-dimensional fault attribute image and a three-dimensional fault attribute data volume;
[0047] S2: Use a matched filtering method to estimate the main fault orientation of the fault attribute data;
[0048] In this step, estimating the main fault direction of the fault attribute data using the matched filtering method (also known as dip-azimuth scanning) includes:
[0049] Smooth each sample point in the fault attribute data along all possible fault directions, find the maximum smoothed response of each sample point and use it as the attribute value of that sample point, and record the direction that produces the maximum smoothed response as the estimated main fault direction at that sample point.
[0050] When the fault attribute data is a two-dimensional fault attribute image, for each sample point in the two-dimensional fault attribute image, recursive Gaussian smoothing is performed along all fault directions, and smoothing values in different directions are accumulated in the smoothing volume.
[0051] Record the maximum smooth value of each sample point to obtain the maximum smooth value volume of the entire data. At the same time, record the azimuth volume and dip volume corresponding to the maximum smooth value. The azimuth volume and dip volume represent the azimuth and dip estimates of the fault attribute data, which are used to estimate the main fault direction.
[0052] When the fault attribute data is a three-dimensional fault attribute data volume, the fault direction is determined by both the dip angle and the azimuth angle;
[0053] Find the fault azimuth and dip combination that produces the maximum smoothness value by following different plane smoothing fault property data defined by all possible combinations of azimuth and dip.
[0054] S3: Automatically find seed points from all sample points of the fault attribute data, and define a local window centered on the seed points;
[0055] In this step, seed points are automatically found from all sample points of the fault attribute data, including:
[0056] Collect all sample points whose attribute values are greater than or equal to a set threshold;
[0057] Seed points are determined according to the attribute values from highest to lowest.
[0058] When determining the final seed point, the distance between the current sample point and all the already determined seed points is calculated. The current sample point is determined as a seed point if and only if the minimum distance is greater than or equal to the predefined distance.
[0059] Preferably, when calculating seed points in a two-dimensional tomographic attribute image, before collecting all sample points with attribute values greater than or equal to a set threshold, the method further includes:
[0060] The non-maximum suppression method is applied to the original fault attribute image to calculate the refined fault attribute image. In the non-maximum suppression method, only the attribute values perpendicular to the estimated fault direction are retained, and the attribute values in other places are set to zero to obtain the refined fault attribute image.
[0061] In this step, defining a local window centered on the seed point includes:
[0062] When the fault attribute data is a two-dimensional fault attribute image, for each seed point, a local rectangular window is defined. This window is centered on the seed point and aligned with the uv space coordinates. The u-axis and v-axis are perpendicular to and parallel to the estimated main fault direction, respectively.
[0063] When the fault attribute data is a three-dimensional fault attribute data volume, a local 3D window is defined for each seed point. The local 3D window is centered on the seed point and is spatially aligned with the estimated azimuth and dip angle of the seed point.
[0064] S4: Using the seed point as a control point, the optimal fault path is traced from the fault attribute data within the local window along the estimated main fault direction to achieve fault attribute feature enhancement.
[0065] In this step, tracing the optimal fault path from the fault attribute data within the local window along the estimated main fault direction includes:
[0066] The following constraint maximization problem is solved by a three-step dynamic time warping algorithm consisting of nonlinear smoothing, forward accumulation, and backward tracking, to find the optimal fault path and enhance fault attribute features.
[0067]
[0068] |j[i+1]-j[i]|≤ε(0<ε≤1)
[0069] In the formula, j[i] represents the optimal fault path, g[i,j] represents the fault attribute image, ε represents the set threshold, and the second formula represents the fault slope constraint.
[0070] Example 2
[0071] This embodiment provides a fault attribute enhancement method based on matched filtering. Due to the low signal-to-noise ratio, unclear fault features, and poor continuity of conventional three-dimensional fault attribute data, this method introduces a fault attribute enhancement method based on dynamic time warping to reduce the impact of noise, improve the signal-to-noise ratio of fault attributes, and enhance the spatial features and continuity of faults.
[0072] The specific method of this embodiment is as follows:
[0073] 1) Tilt and azimuth scanning
[0074] This step uses matched filtering (also known as dip-azimuth scan) to estimate the fault orientation. For each sample point in the fault attribute data, the maximum smoothed response is found along all possible fault orientations and used as the attribute value for that point. The direction that produces the maximum smoothed response is recorded as the estimated dominant fault orientation at that sample point.
[0075] For each sample point x in the fault attribute image, recursive Gaussian smoothing is performed along all fault directions, and the smoothing values in different directions are accumulated in the smoothing volume c(x). c(x) is calculated as follows:
[0076]
[0077] In the formula, θ s ={θ min ,θ min +d θ ,θ min +2d θ ,…,θ max}, d φ and d θ These represent the azimuth and tilt scan intervals, respectively, and are specified by the user, typically greater than 1. θ min and θ max The range of dip angles in the three-dimensional fault property volume is defined. Indicates along the azimuth angle Φ k and tilt angle θ k Defined fault direction smoothing volume f(x).
[0078] Recording the maximum smoothed value for each sample point yields the maximum smoothed volume of the entire dataset. Simultaneously, the azimuth volume Φ(x) and dip volume θ(x) corresponding to the generation of this maximum smooth value are recorded. Φ(x) and θ(x) represent the azimuth and dip estimates of the fault attribute data. Since the fault features are most consistent in the directions defined by the azimuth and dip, the enhanced fault attribute volume h(x) can be calculated by the following formula:
[0079]
[0080] In the formula, N represents the number of dip and azimuth combinations in the matched filtering method. The computational cost of the scanning method depends on the number of fault directions scanned, i.e., N in the formula. Since this step does not require accurate calculation of the fault directions, only sparse fault angles need to be scanned.
[0081] In the case of three-dimensional fault data volumes, the fault direction is determined by both the dip angle and the strike (azimuth). Therefore, it is necessary to estimate both the dip angle and the azimuth simultaneously to determine the fault direction. The specific method is as follows:
[0082] The fault attribute data is smoothed along different planes defined by all possible combinations of azimuth and dip angles to find the combination of azimuth and dip angles that produces the maximum smoothing value. Assuming the fault plane is locally planar, the 2D smoothing filter (oriented by the combination of dip and azimuth angles) will produce the maximum smoothing response if the smoothing surface coincides with the local fault plane. Assuming the possible fault azimuth and dip angles are between [0°, 180°] and [65°, 90°] ∪ [-90°, -65°], typically only 18 × 10 combinations of fault azimuth and dip angles are scanned, i.e., the azimuth is scanned every 10° and the dip every 5°. The estimated fault direction provides a general trend of fault extension to guide the next step of local fault plane tracing. A more accurate fault direction can be calculated from the traced fault planes. The enhanced fault attribute volume, dip volume, and azimuth volume scanned in this step are shown below. Figures 2a-2c As shown.
[0083] 2) Local optimal fault plane tracking
[0084] Next, the seed point is automatically calculated, and the optimal fault path is traced along the estimated main fault direction.
[0085] When calculating seed points in a 2D fault attribute image, non-maximum suppression (NMS) is first applied to the original fault attribute image to obtain a refined fault attribute image. In NMS, only attribute values along the fault ridges (perpendicular to the estimated fault direction) are retained, while attribute values elsewhere are set to zero to obtain the refined attribute image. Then, all sample points with attribute values greater than or equal to a certain threshold are collected. Finally, seed points are determined in descending order of attribute values. When determining the final seed points, the distance between the current sample point and all previously determined seed points is calculated. The current sample point is determined as a seed point only if the minimum distance is greater than or equal to a predefined distance. This significantly reduces the number of seed points, thus saving computation time for subsequent optimal fault path tracing at the seed points. Smaller distances will generate more seed points, which can provide more fault information for subsequent fault path integration, but this requires more time to calculate the fault path.
[0086] For each seed point, a local rectangular window is defined, centered on the seed point and aligned with the uv spatial coordinates. The u-axis and v-axis are perpendicular to and parallel to the estimated fault direction, respectively. Since the fault to be tracked is expected to extend mostly along the v-axis, the rectangular window is typically wider in the v-direction than in the u-direction. In the dynamic time warping algorithm, the seed point is used as a control point to track the optimal fault path j[i] from the original fault attribute image located within the rectangular window, i.e., solving the constrained maximization problem shown in the following equation.
[0087]
[0088] |j[i+1]-j[i]|≤ε(0<ε≤1)
[0089] In the formula, g[i,j] represents the fault attribute image, and the second formula represents the fault slope constraint.
[0090] The above constraint maximization problem is solved by a three-step dynamic time warping algorithm consisting of nonlinear smoothing, forward accumulation, and backward tracking, to find the optimal fault path j[i] and achieve fault attribute feature enhancement.
[0091] The three-step dynamic time warping algorithm includes the following steps:
[0092] ① Nonlinear smoothing is achieved by applying nonlinear accumulation to the input tomographic attribute image in the positive (from left to right) and negative (from right to left) directions, as shown in the following formula.
[0093] s[i,j]=f[i,j]+b[i,j]-g[i,j]
[0094] In the formula, f[i,j] represents the forward smoothing result, b[i,j] represents the reverse smoothing result, and g[i,j] represents the input fault attribute image.
[0095] ② On the nonlinearly smoothed image, the optimal fault path is selected through forward accumulation and backtracking. The forward accumulation process is as follows.
[0096] a[0,j]=s[0,j]
[0097]
[0098] for i = 1, 2, ..., N-1
[0099] ③ The optimal fault path is calculated on the forward-accumulated image a[i,j] through a backtracking process. The backtracking starts from the last position of the path and ends at the first position. The backtracking method is as follows: j[N-1]=argmax j a[N-1,j]
[0100] l = j[i]
[0101]
[0102] for i = N-1, N-1, ..., 1
[0103] In the case of 3D data, seed points are first automatically found using the same method as for selecting 2D seed points. Then, a local 3D window is defined for each seed point, centered on the seed point and spatially aligned with the estimated azimuth and dip angles of that seed point. Similar to the 2D window definition, this 3D window is also defined to extend further in the estimated fault azimuth and dip directions than in the directions perpendicular to the azimuth and dip directions. In this 3D window oriented by azimuth and dip, the actual fault passing through the seed point should be a surface patch slightly tilted along the azimuth and dip directions. Then, the aforementioned three-step dynamic time warping algorithm is used to trace the fault surface from the fault attribute volume within the local 3D window, thereby enhancing the fault features.
[0104] Comparison of original fault properties and fault properties enhanced by this method: Figure 3a and Figure 3b As shown, the data processed using this method exhibits significantly improved noise suppression, clearer and more continuous fault features, and easier fault plane tracking compared to the data before enhancement. Furthermore, the maximum smooth response method can accurately estimate the azimuth and dip angles of the fault, aiding in overall fault analysis and automatic fault plane extraction.
[0105] Example 3
[0106] This embodiment provides a fault attribute enhancement device based on matched filtering, including:
[0107] The data acquisition module is used to acquire fault attribute data, which includes two-dimensional fault attribute images and three-dimensional fault attribute data volumes.
[0108] The fault estimation module is used to estimate the main fault direction of the fault attribute data using a matched filtering method.
[0109] The window definition module is used to automatically find seed points from all sample points of the fault attribute data and define a local window centered on the seed points.
[0110] The fault enhancement module is used to use the seed point as a control point to trace the optimal fault path from the fault attribute data within the local window along the estimated main fault direction, thereby enhancing the fault attribute features.
[0111] Example 4
[0112] This embodiment provides an electronic device, the electronic device comprising:
[0113] At least one processor; and,
[0114] A memory communicatively connected to the at least one processor; wherein,
[0115] The memory stores instructions that can be executed by the at least one processor, which enables the at least one processor to perform the matched filtering-based tomographic attribute enhancement method described in the above embodiments.
[0116] An electronic device according to embodiments of the present disclosure includes a memory and a processor. The memory is used to store non-transitory computer-readable instructions. Specifically, the memory may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may, for example, include random access memory (RAM) and / or cache memory. The non-volatile memory may, for example, include read-only memory (ROM), hard disk, flash memory, etc.
[0117] The processor may be a central processing unit (CPU) or other form of processing unit with data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions. In one embodiment of this disclosure, the processor is used to execute computer-readable instructions stored in the memory.
[0118] Those skilled in the art will understand that, in order to solve the technical problem of how to achieve a good user experience, this embodiment may also include well-known structures such as communication buses and interfaces, and these well-known structures should also be included within the protection scope of this disclosure.
[0119] For a detailed description of this embodiment, please refer to the corresponding descriptions in the foregoing embodiments, which will not be repeated here.
[0120] Example 5
[0121] This embodiment provides a non-transitory computer-readable storage medium that stores computer instructions for causing a computer to execute the matched filtering-based tomographic attribute enhancement method described in the above embodiment.
[0122] A computer-readable storage medium according to embodiments of the present disclosure stores non-transitory computer-readable instructions. When these non-transitory computer-readable instructions are executed by a processor, all or part of the steps of the methods described in the foregoing embodiments of the present disclosure are performed.
[0123] The aforementioned computer-readable storage media include, but are 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 portable hard drive), media with built-in rewritable non-volatile memory (e.g., memory card), and media with built-in ROM (e.g., ROM cartridge).
[0124] The various embodiments of the present invention have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments.
Claims
1. A matched filter based fault attribute enhancement method, characterized in that, The method comprises: acquiring fault attribute data, the fault attribute data comprising a two-dimensional fault attribute image and a three-dimensional fault attribute data volume; estimating a main fault direction of the fault attribute data using a matched filter method; automatically finding a seed point from all sample points of the fault attribute data and defining a local window centered at the seed point; using the seed point as a control point, tracking an optimal fault path from the fault attribute data in the local window along the estimated main fault direction to achieve fault attribute feature enhancement.
2. The method of claim 1, wherein, The estimation of the main fault direction of the fault attribute data using the matched filter method comprises: smoothing each sample point in the fault attribute data along all possible fault directions to find a maximum smoothing response of each sample point, taking the maximum smoothing response as the attribute value of the sample point, and recording the direction generating the maximum smoothing response as the estimated main fault direction at the sample point.
3. The method of claim 2, wherein, When the fault attribute data is a two-dimensional fault attribute image, for each sample point in the two-dimensional fault attribute image, recursive Gaussian smoothing is performed along all fault directions, and the smoothing values of different directions are accumulated in the smoothing volume. The maximum smoothing value of each sample point is recorded to obtain a maximum smoothing value volume of the entire data, and the corresponding azimuth angle volume and inclination angle volume are recorded, which represent the azimuth angle and inclination angle estimation of the fault attribute data and are used to estimate the main fault direction.
4. The method of claim 3, wherein, When the fault attribute data is a three-dimensional fault attribute data volume, the fault direction is determined by the inclination angle and the azimuth angle. The fault attribute data is smoothed along different planes defined by all possible combinations of the azimuth angle and the inclination angle to find the combination of the fault azimuth angle and the inclination angle generating the maximum smoothing value.
5. The method of claim 2, wherein, The automatic finding of the seed point from all sample points of the fault attribute data comprises: collecting all sample points with attribute values greater than or equal to a set threshold value; determining the seed point in the order of high to low attribute value; when determining the final seed point, calculating the distance between the current sample point and all determined seed points, and determining the current sample point as a seed point only when the minimum distance is greater than or equal to a predefined distance.
6. The method of claim 5, wherein, When calculating the seed point in the two-dimensional fault attribute image, before collecting all sample points with attribute values greater than or equal to the set threshold value, the method further comprises: applying a non-maximum suppression method to the original fault attribute image to calculate a refined fault attribute image, in which only the attribute values perpendicular to the estimated fault direction are retained, and the attribute values in other places are set to zero to obtain the refined fault attribute image.
7. The method of claim 4, wherein, The definition of the local window centered at the seed point comprises: when the fault attribute data is a two-dimensional fault attribute image, for each seed point, a local rectangular window is defined, which is centered at the seed point and aligned with the u-v space coordinates, with the u-axis and the v-axis being perpendicular and parallel to the estimated main fault direction, respectively; when the fault attribute data is a three-dimensional fault attribute data volume, a local 3D window is defined for each seed point, which is centered at the seed point and spatially aligned with the estimated azimuth angle and inclination angle of the seed point.
8. The method of claim 7, wherein, The tracking of the optimal fault path along the estimated main fault direction from the fault attribute data within the local window comprises: The optimal fault path is found by solving a constrained maximization problem with a three-step dynamic time warping algorithm of non-linear smoothing, forward accumulation and backward tracking to achieve the enhancement of fault attribute features. |j[i+1]-j[i]|≤ε(0<ε≤1) In the formula, j[i] represents the optimal fault path, g[i,j] represents the fault attribute image, and ε represents a set threshold. The second formula represents a fault slope constraint.
9. An electronic device, comprising: The electronic device comprises: at least one processor; and a memory connected with the at least one processor in communication; wherein The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the matching filter-based fault attribute enhancement method of any one of claims 1-8.
10. A non-transitory computer-readable storage medium, comprising: The non-transitory computer readable storage medium stores computer instructions for causing a computer to perform the matching filter-based fault attribute enhancement method of any one of claims 1-8.