Computer vision-based method for reconstructing three-dimensional fractures of hydraulic fracturing experimental samples

Abstract

The application relates to the technical field of oil and gas field development, and discloses a computer vision-based three-dimensional fracture reconstruction method for a hydraulic fracturing experiment sample. The method comprises the following steps: obtaining a first point cloud set of a target fracture region; determining a first normal vector corresponding to each first point cloud data; rotating the first normal vector so that a second normal vector points to the same side of the first point cloud set; determining an outlier point cloud set corresponding to the first point cloud set; dividing a region where the outlier point cloud set and the first point cloud set are located into a plurality of sub-regions according to a recursive algorithm; for each sub-region, generating a space envelope of the sub-region based on a point at the outermost side of the sub-region; obtaining a local interpolation implicit function of each space envelope through a multiple harmonic spline radial basis local interpolation algorithm; determining a global interpolation implicit function corresponding to the target fracture region according to all the local interpolation implicit functions; and determining an equal-value surface corresponding to interpolation when the global interpolation implicit function is equal to zero, so as to obtain a three-dimensional fracture image of the target fracture region.

Description

Technical Field

[0001] This application relates to the field of oil and gas field development technology, specifically to a method, processor, system and storage medium for three-dimensional fracture reconstruction of hydraulic fracturing experimental specimens based on computer vision. Background Technology

[0002] Hydraulic fracturing is a reservoir stimulation technique that uses high-pressure water to create artificial fractures in oil and gas reservoirs to improve fluid flow. After the reservoir is fracturing, the fracture morphology needs to be reconstructed to quantitatively evaluate the effectiveness of the hydraulic fracturing.

[0003] Commonly used crack morphology reconstruction methods in existing technologies include tracer methods, laser scanning, and 3D CT scanning. Tracer methods exhibit significant errors in revealing crack morphology, and the reconstructed two-dimensional crack images are insufficient to characterize the three-dimensional spatial morphology of the crack. Therefore, using tracer methods to evaluate the reconstructed crack morphology can lead to inaccurate assessments. 3D CT scanning is costly and time-consuming, and cannot scan large-sized artificial cracks. Laser scanning is affected by reflections from non-planar crack surfaces, resulting in data voids at the sampling surface and partial incomplete images in the reconstructed image, making it difficult to quantitatively evaluate the crack morphology. Summary of the Invention

[0004] The purpose of this application is to provide a method, processor, system, and storage medium for reconstructing three-dimensional cracks in hydraulic fracturing experimental specimens based on computer vision.

[0005] To achieve the above objectives, the first aspect of this application provides a method for reconstructing three-dimensional cracks in hydraulic fracturing test specimens based on computer vision, comprising:

[0006] The first point cloud of the target crack region is obtained, wherein the first point cloud includes multiple first point cloud data collected by a laser scanning device at a preset sampling interval;

[0007] Determine the first normal vector corresponding to each first point cloud data;

[0008] Rotate the first normal vector so that the second normal vector points to the same side of the first point cluster, wherein the second normal vector includes the rotated first normal vector and the unrotated first normal vector;

[0009] The outlier cloud corresponding to the first point cloud is determined based on the second normal vector and the preset outlier coefficient, wherein the outlier cloud includes multiple second point cloud data.

[0010] The regions containing the outlier cloud and the first point cloud are divided into multiple sub-regions using a recursive algorithm, wherein the number of point clouds in each sub-region is less than or equal to a preset threshold.

[0011] For each sub-region, generate the spatial envelope of the sub-region based on the outermost point of the sub-region;

[0012] The local interpolation function of each spatial envelope is obtained by interpolating each spatial envelope using a multiharmonic spline radial basis local interpolation algorithm.

[0013] The global implicit interpolation function corresponding to the target crack region is determined based on all local implicit interpolation functions;

[0014] Determine the isosurface corresponding to the interpolation when the global interpolation implicit function is equal to zero, so as to obtain a three-dimensional crack image of the target crack region.

[0015] In this embodiment of the application, rotating the first normal vector so that the second normal vector points to the same side of the first point cloud includes: determining a weighted undirected graph of multiple first point cloud data; and rotating the first normal vector according to the weighted undirected graph so that the second normal vector points to the same side of the first point cloud.

[0016] In this embodiment of the application, rotating the first normal vector according to the weighted undirected graph so that the second normal vector points to the same side of the first point cloud includes: determining the minimum spanning tree of the first point cloud according to the weighted undirected graph; determining the traversal sequence of the first point cloud data according to the minimum spanning tree; traversing the first normal vector according to the traversal sequence to determine the normal vector to be rotated in the first normal vector; and rotating the normal vector to be rotated to obtain the second normal vector.

[0017] In this embodiment of the application, dividing the region containing the outlier cloud and the first point cloud into multiple sub-regions according to a recursive algorithm includes: determining an initial cube sub-region based on the region containing the outlier cloud and the first point cloud; obtaining the total number of point clouds in the initial cube sub-region; dividing the initial cube sub-region into a preset number of sub-cube sub-regions if the total number of point cloud data exceeds a preset threshold; obtaining the amount of point cloud data in each sub-cube sub-region; and dividing the sub-cube sub-region into a preset number of sub-cube sub-regions if the amount of point cloud data in each sub-cube sub-region exceeds a preset threshold, so that the amount of point cloud data in each sub-cube sub-region is less than or equal to the preset threshold.

[0018] In this embodiment of the application, the method further includes: after obtaining the first point cloud of the target crack region, preprocessing the first point cloud to obtain the preprocessed second point cloud, wherein the preprocessing includes data registration and data cleaning.

[0019] In this embodiment of the application, the method further includes: after obtaining the three-dimensional crack image, determining the hole filling rate and the three-dimensional crack area of ​​the three-dimensional crack image, so as to determine the reconstruction effect of the three-dimensional crack image.

[0020] In this embodiment of the application, generating the spatial envelope of a sub-region based on the outermost point of the sub-region for each sub-region includes: determining the outermost point cloud data of each sub-region using the Alpha-shape algorithm; and connecting the outermost point cloud data of each sub-region into multiple triangular faces to form a spatial envelope composed of triangular faces.

[0021] A second aspect of this application provides a processor configured to perform the above-described computer vision-based three-dimensional crack reconstruction method for hydraulic fracturing experimental specimens.

[0022] A third aspect of this application provides a computer vision-based three-dimensional fracture reconstruction system for hydraulic fracturing test specimens, comprising: a laser scanning device for acquiring a first point set of a target fracture region; and the aforementioned processor.

[0023] A fourth aspect of this application provides a machine-readable storage medium storing instructions that, when executed by a processor, configure the processor to perform the aforementioned computer vision-based three-dimensional crack reconstruction method for hydraulic fracturing test specimens.

[0024] Through the above technical solution, this application can process the first point cloud acquired by the lidar and determine the outlier point cloud corresponding to the first point cloud. A recursive algorithm is used to segment the regions containing the outlier point cloud and the first point cloud, resulting in multiple sub-regions. The spatial envelope corresponding to each sub-region is determined, and interpolation is performed on each spatial envelope using a multiple harmonic spline radial basis local interpolation algorithm to obtain the local interpolation implicit function for each spatial envelope. The global interpolation implicit function is then determined based on all local interpolation implicit functions. The isosurface corresponding to the interpolation when the global interpolation implicit function is equal to zero is determined, thereby obtaining a three-dimensional crack image of the target crack region. Using this method, the obtained three-dimensional crack image can fill in the holes in the three-dimensional crack image obtained by laser scanning, resulting in a more accurate three-dimensional crack image.

[0025] Other features and advantages of the embodiments of this application will be described in detail in the following detailed description section. Attached Figure Description

[0026] The accompanying drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the following detailed description to explain the embodiments of this application, but do not constitute a limitation on the embodiments of this application. In the drawings:

[0027] Figure 1 The schematic diagram illustrates a flowchart of a computer vision-based three-dimensional fracture reconstruction method for hydraulic fracturing experimental specimens according to an embodiment of this application.

[0028] Figure 2The schematic diagram illustrates the structural block diagram of a computer vision-based three-dimensional crack reconstruction system for hydraulic fracturing test specimens according to an embodiment of this application.

[0029] Figure 3 The diagram illustrates the internal structure of a computer device according to an embodiment of this application. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for illustration and explanation of the embodiments of this application and are not intended to limit the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0031] Figure 1 The illustration schematically shows a flowchart of a computer vision-based three-dimensional fracture reconstruction method for hydraulic fracturing experimental specimens according to an embodiment of this application. Figure 1 As shown in one embodiment of this application, a method for reconstructing three-dimensional cracks in a hydraulic fracturing experimental specimen based on computer vision is provided, including the following steps:

[0032] S102, acquire the first point cloud set of the target crack region, wherein the first point cloud set includes multiple first point cloud data acquired by a laser scanning device at a preset sampling interval.

[0033] S104, determine the first normal vector corresponding to each first point cloud data.

[0034] S106, rotate the first normal vector so that the second normal vector points to the same side of the first point cluster, wherein the second normal vector includes the rotated first normal vector and the unrotated first normal vector.

[0035] S108, determine the outlier cloud corresponding to the first point cloud based on the second normal vector and the preset outlier coefficient, wherein the outlier cloud includes multiple second point cloud data.

[0036] S110, the region containing the outlier cloud and the first point cloud is divided into multiple sub-regions according to a recursive algorithm, wherein the number of point clouds in each sub-region is less than or equal to a preset threshold.

[0037] S112, For each sub-region, generate the spatial envelope of the sub-region based on the outermost point of the sub-region.

[0038] S114 uses a multi-harmonic spline radial basis local interpolation algorithm to interpolate each spatial envelope to obtain the local interpolation implicit function of each spatial envelope.

[0039] S116, Determine the global interpolation implicit function corresponding to the target crack region based on all local interpolation implicit functions.

[0040] S118, determine the isosurface corresponding to the interpolation when the global interpolation implicit function is equal to zero, so as to obtain the three-dimensional crack image of the target crack region.

[0041] Three-dimensional crack images reconstructed using laser scanning methods often contain gaps, making it difficult to quantitatively evaluate crack morphology. Therefore, this application, based on image completion theory in the field of computer vision, uses interpolation to fill in the gaps in the crack image and reconstruct a three-dimensional crack image, enabling a more refined representation of crack morphology.

[0042] First, the processor can acquire the first point cloud set of the target crack region. The first point cloud set is a collection of multiple first point cloud data points acquired by a laser scanning device at preset sampling intervals. Each first point cloud data point is located on the crack surface of the target crack region. The processor can determine the first normal vector corresponding to each first point cloud data point. Since each first point cloud data point is on the crack surface, the processor can determine the first normal vector based on the first normal vector, which is perpendicular to the crack surface. Furthermore, the processor can rotate the first normal vector so that a second normal vector points to the same side of the first point cloud set. The processor can determine the direction of the first normal vector. If it is determined that there are first normal vectors with inconsistent directions with other first normal vectors, it rotates the inconsistent first normal vectors so that all second normal vectors point to the same side of the first point cloud set, i.e., perpendicular to the crack surface inwards or perpendicular to the crack surface outwards. The second normal vector includes both the rotated and unrotated first normal vectors. The processor can determine the outlier point cloud set corresponding to the first point cloud set based on the second normal vector and a preset outlier coefficient. The outlier point cloud set includes multiple second point cloud data points. The second point cloud data all lie on the second normal vector, and each second point cloud data has a corresponding first point cloud data. The distance between each second point cloud data and its corresponding first point cloud data can be determined based on the outlier coefficient. The processor can recursively divide the regions containing the outlier point cloud set and the first point cloud set into multiple sub-regions. The number of point clouds in each sub-region is less than or equal to a preset threshold, and the point cloud data in each sub-region can include multiple first point cloud data and / or multiple second point cloud data. The processor can determine the outermost point of each sub-region and generate spatial envelopes for multiple sub-regions. Each spatial envelope contains all the point cloud data in each sub-region. The processor can perform interpolation processing on each spatial envelope using a multiharmonic spline radial basis local interpolation algorithm, also known as the Polyharmonic spline radial basis interpolation algorithm. The local interpolation implicit function of each spatial envelope can be obtained through the above algorithm. For each spatial envelope, a weighted summation of the local interpolation functions can be obtained to obtain the global interpolation implicit function. The processor can determine the isosurface corresponding to the global interpolation implicit function being equal to zero, in order to obtain a three-dimensional crack image of the target crack region.

[0043] In one specific embodiment, the processor acquires 400×400×400mm shale samples and performs three-dimensional reconstruction of the fractures after a true three-cycle hydraulic fracturing experiment. Laser scanning is performed on the fracture surface at a sampling interval of 0.1mm to obtain the first point cloud, which includes 2,123,907 first point cloud data points. The processor acquires the first normal vector of each first point cloud data point and rotates it so that the second normal vector is perpendicular to the fracture surface and points inward. The processor determines the outlier cloud of the first point cloud based on the outlier coefficient. The processor divides the regions containing the outlier cloud and the first point cloud into multiple sub-regions. The point cloud data threshold for each sub-region is 7000. For each sub-region, the processor determines the outermost point and generates the spatial envelope of the sub-region. The processor can perform interpolation processing on each spatial envelope using a multiple harmonic spline radial basis local interpolation algorithm to obtain the local interpolation implicit function for each spatial envelope. The processor determines the global interpolation implicit function corresponding to the target fracture region based on all local interpolation implicit functions. The processor can determine the isosurface corresponding to the interpolation when the global interpolation implicit function is equal to zero, so as to obtain a three-dimensional crack image of the target crack region.

[0044] Through the above technical solution, this application can process the first point cloud acquired by the lidar and determine the outlier point cloud corresponding to the first point cloud. A recursive algorithm is used to segment the regions containing the outlier point cloud and the first point cloud, resulting in multiple sub-regions. The spatial envelope corresponding to each sub-region is determined, and interpolation is performed on each spatial envelope using a multiple harmonic spline radial basis local interpolation algorithm to obtain the local interpolation implicit function for each spatial envelope. The global interpolation implicit function is then determined based on all local interpolation implicit functions. The isosurface corresponding to the interpolation when the global interpolation implicit function is equal to zero is determined, thereby obtaining a three-dimensional crack image of the target crack region. Using this method, the obtained three-dimensional crack image can fill in the holes in the three-dimensional crack image, resulting in a more accurate three-dimensional crack image.

[0045] In one embodiment, rotating the first normal vector so that the second normal vector points to the same side of the first point cloud includes: determining a weighted undirected graph of multiple first point cloud data; and rotating the first normal vector according to the weighted undirected graph so that the second normal vector points to the same side of the first point cloud. The weights in the weighted undirected graph of the point cloud data represent the degree of change in the normal vectors corresponding to two adjacent point cloud data. When the weight is greater than 1, it is determined that the angle between the normal vectors of two adjacent points is greater than 90 degrees. Because the distance between two adjacent point cloud data is extremely small in this application, the angle between the normal vectors pointing to the same side is correspondingly small. When it is determined that the angle between the normal vectors of two adjacent point cloud data is greater than 90 degrees, it can be determined that the two normal vectors point to opposite sides perpendicular to the crack surface. The processor can determine the first normal vector to be rotated according to the weighted undirected graph so that the second normal vector points to the same side of the first point cloud.

[0046] In one embodiment, rotating the first normal vector according to the weighted undirected graph so that the second normal vector points to the same side of the first point cloud includes: determining the minimum spanning tree of the first point cloud based on the weighted undirected graph; determining the traversal sequence of the first point cloud data based on the minimum spanning tree; traversing the first normal vectors according to the traversal sequence to determine the normal vector to be rotated in the first normal vectors; and rotating the normal vector to be rotated to obtain the second normal vector. When determining whether each first normal vector needs to be rotated, the processor needs to determine each one individually. If there are many first normal vectors, this can lead to excessive computational overhead and slow processing. Using the minimum spanning tree algorithm, the processor can connect each first point cloud data and generate a traversal sequence graph with the minimum computational cost. Therefore, the processor can determine the minimum spanning tree of the first point cloud data based on the weighted undirected graph. The processor determines the traversal sequence of the first point cloud data based on the minimum spanning tree, and traverses the first normal vectors according to the traversal sequence to determine the normal vector to be rotated in the first normal vectors. During the traversal, if the weight between the current point cloud data and the previous point cloud data is greater than 1, the first normal vector corresponding to the current point cloud data is determined to be the normal vector to be rotated. The processor can rotate the normal vector to be rotated so that all the normal vectors point to the same side of the first point cloud data, thus obtaining the second normal vector.

[0047] For example, the processor can acquire the first point cloud data of the target crack region and determine the first normal vector corresponding to each first point cloud data. The processor determines a weighted undirected graph among the first point cloud data and determines the weights between two adjacent point cloud data. Each point cloud data can have multiple adjacent point cloud data; therefore, the processor can determine the minimum spanning tree of the first point cloud data based on the weighted undirected graph and traverse each first point cloud data according to the minimum spanning tree. The processor determines the first normal vector to be rotated and rotates it so that the second normal vector points to the same side of the first point cloud set. Specifically, if the processor determines that the weight between the first and second first point cloud data is greater than 1, the processor rotates the first normal vector corresponding to the second first point cloud data so that the rotated first normal vector and the first normal vector corresponding to the first first point cloud data both point to the same side perpendicular to the crack surface. After the rotation, the weights associated with the second first point cloud data in the weighted undirected graph have changed. The processor can then redetermine the weights between the second and third first point cloud data to determine whether the first normal vector corresponding to the third first point cloud data needs to be rotated.

[0048] In one embodiment, dividing the region containing the outlier point cloud and the first point cloud into multiple sub-regions using a recursive algorithm includes: determining an initial cube sub-region based on the regions containing the outlier point cloud and the first point cloud; obtaining the total number of point clouds in the initial cube sub-region; dividing the initial cube sub-region into a preset number of sub-cube sub-regions if the total number of point cloud data exceeds a preset threshold; obtaining the amount of point cloud data in each sub-cube sub-region; and dividing the sub-cube sub-region into a preset number of sub-cube sub-regions if the amount of point cloud data in each sub-cube sub-region exceeds a preset threshold, so that the amount of point cloud data in each sub-cube sub-region is less than or equal to the preset threshold. The processor can determine the initial cube sub-region based on the regions containing the outlier point cloud and the first point cloud, and determine whether the total number of point clouds in the initial cube sub-region exceeds a preset threshold. If the total number of point clouds exceeds the preset threshold, the processor can divide the initial cube sub-region into a preset number of sub-cube sub-regions. The processor then determines the amount of point cloud data in each sub-cube sub-region. If the amount of point cloud data in a certain sub-cube subdomain exceeds a preset threshold, the processor divides the sub-cube subdomain into a preset number of sub-cube subdomains so that the amount of point cloud data in each sub-cube subdomain is less than or equal to the preset threshold, so as to facilitate the computer to perform interpolation calculations for each sub-cube subdomain.

[0049] In one embodiment, the method further includes: after acquiring the first point cloud dataset of the target crack region, preprocessing the first point cloud dataset to obtain a preprocessed second point cloud dataset, wherein the preprocessing includes data registration and data cleaning. After acquiring the first point cloud dataset, the processor can determine that there is abnormal point cloud data in the first point cloud dataset and remove the abnormal data to complete the data cleaning of the first point cloud dataset. The laser scanning device may sample the first point cloud dataset of the target crack region through multiple different sampling locations. Therefore, the processor can perform data registration on the first point cloud data so that each point cloud data corresponds to the same coordinate system, generating a data-registered point cloud dataset. After data registration and data cleaning of the first point cloud dataset, the second point cloud dataset can be obtained.

[0050] In one embodiment, the method further includes: after obtaining a three-dimensional fracture image, determining the cavity completion rate and the three-dimensional fracture area of ​​the three-dimensional fracture image to determine the reconstruction effect of the three-dimensional fracture image. The processor can determine the cavity completion rate and the three-dimensional fracture area of ​​the reconstructed three-dimensional fracture image. Based on the cavity completion rate, the degree of accuracy improvement of the reconstruction effect relative to the original laser scanning image can be determined. If the cavity completion rate is low, the processor can adjust the parameters in the reconstruction process to improve image accuracy. The processor can determine the three-dimensional fracture area based on the reconstructed three-dimensional fracture image so that the user can determine the hydraulic fracturing effect.

[0051] In one embodiment, generating a spatial envelope of a sub-region based on its outermost points includes: determining the outermost point cloud data of each sub-region using an Alpha-shape algorithm; and for each sub-region, concatenating the outermost point cloud data into multiple triangular faces to form a spatial envelope composed of these triangular faces. The Alpha-shape algorithm is a contour shape recognition algorithm that can be used to determine the contour of a point cloud image. The processor can determine the outermost point cloud data of each sub-region using the Alpha-shape algorithm and concatenate the outermost point cloud data of each sub-region into multiple triangular faces to form a spatial envelope composed of these triangular faces. The processor can then perform interpolation processing on each spatial envelope.

[0052] In a specific embodiment, the computer vision-based method for reconstructing three-dimensional fractures in hydraulic fracturing test specimens includes: acquiring a first point cloud set of a target fracture region, wherein the first point cloud set includes multiple first point cloud data acquired by a laser scanning device at preset sampling intervals. After acquiring the first point cloud set of the target fracture region, the first point cloud set is preprocessed to obtain a preprocessed point cloud set, wherein the preprocessing includes data registration and data cleaning. A first normal vector corresponding to each first point cloud data is determined. A minimum spanning tree of the first point cloud set is determined based on a weighted undirected graph. A traversal sequence of the first point cloud data is determined based on the minimum spanning tree. The first normal vector is traversed according to the traversal sequence to determine the normal vector to be rotated in the first normal vector. The normal vector to be rotated is rotated to obtain a second normal vector. An outlier point cloud set corresponding to the first point cloud set is determined based on the second normal vector and a preset outlier coefficient, wherein the outlier point cloud set includes multiple second point cloud data. An initial cubic subdomain is determined based on the outlier point cloud set and the region where the first point cloud set is located. The initial cube subdomain is divided into a predetermined number of sub-cube subdomains if the total number of point cloud data exceeds a preset threshold. The point cloud data volume in each sub-cube subdomain is also determined. If the point cloud data volume in each sub-cube subdomain exceeds a preset threshold, it is further divided into a predetermined number of sub-cube subdomains to ensure that the point cloud data volume in each sub-cube subdomain is less than or equal to the preset threshold. The outermost point cloud data of each sub-region is determined using the Alpha-shape algorithm. For each sub-region, the outermost point cloud data is connected to form multiple triangular faces, creating a spatial envelope composed of these faces. Each spatial envelope is interpolated using a multiharmonic spline radial basis local interpolation algorithm to obtain its local interpolation implicit function. Based on all local interpolation implicit functions, the global interpolation implicit function corresponding to the target crack region is determined. The isosurface corresponding to the interpolation when the global interpolation implicit function is equal to zero is then determined to obtain a three-dimensional crack image of the target crack region. After obtaining the three-dimensional crack image, the void filling rate and the area of ​​the three-dimensional crack image are determined to determine the reconstruction effect of the three-dimensional crack image.

[0053] Using the above method, this application can process the first point cloud acquired by the lidar and determine the outlier point cloud corresponding to the first point cloud. A recursive algorithm is used to segment the regions containing the outlier point cloud and the first point cloud, resulting in multiple sub-regions. The spatial envelope corresponding to each sub-region is determined, and interpolation is performed on each spatial envelope using a multiple harmonic spline radial basis local interpolation algorithm to obtain the local interpolation implicit function for each spatial envelope. Based on all the local interpolation implicit functions, the global interpolation implicit function is determined. The isosurface corresponding to the interpolation when the global interpolation implicit function is equal to zero is determined, thereby obtaining a three-dimensional crack image of the target crack region. Using this method, the obtained three-dimensional crack image can fill in the holes in the three-dimensional crack image, resulting in a more accurate three-dimensional crack image.

[0054] Figure 1 This is a flowchart illustrating a computer vision-based method for reconstructing three-dimensional cracks in a hydraulic fracturing experimental specimen, as shown in one embodiment. It should be understood that, although... Figure 1 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0055] In one embodiment, such as Figure 2 As shown, a computer vision-based three-dimensional crack reconstruction system 200 for hydraulic fracturing experimental specimens is provided, including a laser scanning device 201 and a processor 202, wherein:

[0056] Laser scanning device 201 is used to acquire the first point cluster of the target crack area.

[0057] Processor 202 is used to execute the above-described computer vision-based three-dimensional crack reconstruction method for hydraulic fracturing experimental specimens.

[0058] The processor contains a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured, and adjusting kernel parameters enables a computer vision-based method for reconstructing three-dimensional cracks in hydraulic fracturing experimental specimens.

[0059] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0060] This application provides a storage medium storing a program that, when executed by a processor, implements the above-described computer vision-based three-dimensional crack reconstruction method for hydraulic fracturing experimental specimens.

[0061] This application provides a processor for running a program, wherein the program executes the above-described computer vision-based three-dimensional crack reconstruction method for hydraulic fracturing experimental specimens.

[0062] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 3 As shown, the computer device includes a processor A01, a network interface A02, a memory (not shown), and a database (not shown) connected via a system bus. The processor A01 provides computing and control capabilities. The memory includes internal memory A03 and a non-volatile storage medium A04. The non-volatile storage medium A04 stores an operating system B01, a computer program B02, and a database (not shown). The internal memory A03 provides an environment for the operation of the operating system B01 and the computer program B02 stored in the non-volatile storage medium A04. The network interface A02 is used for communication with external terminals via a network connection. When the computer program B02 is executed by the processor A01, it implements a computer vision-based method for reconstructing three-dimensional cracks in hydraulic fracturing experimental specimens.

[0063] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0064] This application provides a device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it performs the following steps:

[0065] The process involves: acquiring a first point cloud set of the target crack region, comprising multiple first point cloud data collected by a laser scanning device at preset sampling intervals; determining a first normal vector corresponding to each first point cloud data; rotating the first normal vector so that a second normal vector points to the same side of the first point cloud set, wherein the second normal vector includes the rotated first normal vector and the unrotated first normal vector; determining an outlier point cloud set corresponding to the first point cloud set based on the second normal vector and a preset outlier coefficient, wherein the outlier point cloud set includes multiple second point cloud data; and connecting the outlier point cloud set with the first point cloud set. The region containing the set is divided into multiple sub-regions using a recursive algorithm, where the number of point clouds in each sub-region is less than or equal to a preset threshold. For each sub-region, a spatial envelope is generated based on the outermost point of the sub-region. Each spatial envelope is interpolated using a multiple harmonic spline radial basis local interpolation algorithm to obtain the local interpolation implicit function of each spatial envelope. The global interpolation implicit function corresponding to the target crack region is determined based on all local interpolation implicit functions. The isosurface corresponding to the interpolation when the global interpolation implicit function is equal to zero is determined to obtain the three-dimensional crack image of the target crack region.

[0066] In one embodiment, rotating the first normal vector so that the second normal vector points to the same side of the first point cloud includes: determining a weighted undirected graph of a plurality of first point cloud data; and rotating the first normal vector according to the weighted undirected graph so that the second normal vector points to the same side of the first point cloud.

[0067] In one embodiment, rotating the first normal vector according to the weighted undirected graph so that the second normal vector points to the same side of the first point cloud includes: determining the minimum spanning tree of the first point cloud according to the weighted undirected graph; determining the traversal sequence of the first point cloud data according to the minimum spanning tree; traversing the first normal vector according to the traversal sequence to determine the normal vector to be rotated in the first normal vector; and rotating the normal vector to be rotated to obtain the second normal vector.

[0068] In one embodiment, dividing the region containing the outlier cloud and the first point cloud into multiple sub-regions using a recursive algorithm includes: determining an initial cube sub-region based on the region containing the outlier cloud and the first point cloud; obtaining the total number of point clouds in the initial cube sub-region; dividing the initial cube sub-region into a preset number of sub-cube sub-regions if the total number of point cloud data exceeds a preset threshold; obtaining the amount of point cloud data in each sub-cube sub-region; and dividing the sub-cube sub-region into a preset number of sub-cube sub-regions if the amount of point cloud data in each sub-cube sub-region exceeds a preset threshold, so that the amount of point cloud data in each sub-cube sub-region is less than or equal to the preset threshold.

[0069] In one embodiment, the method further includes: after obtaining the first point cloud of the target crack region, preprocessing the first point cloud to obtain a preprocessed second point cloud, wherein the preprocessing includes data registration and data cleaning.

[0070] In one embodiment, the method further includes: after obtaining the three-dimensional crack image, determining the void filling rate and the three-dimensional crack area of ​​the three-dimensional crack image to determine the reconstruction effect of the three-dimensional crack image.

[0071] In one embodiment, generating the spatial envelope of a sub-region based on the outermost point of the sub-region for each sub-region includes: determining the outermost point cloud data of each sub-region using an Alpha-shape algorithm; and connecting the outermost point cloud data into multiple triangular faces for each sub-region to form a spatial envelope composed of triangular faces.

[0072] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0073] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0074] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0075] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0076] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0077] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0078] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0079] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0080] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for reconstructing three-dimensional cracks in hydraulic fracturing experimental specimens based on computer vision, characterized in that, The reconstruction method includes: Acquire the first point cloud set of the target crack region, wherein the first point cloud set includes multiple first point cloud data acquired by a laser scanning device at a preset sampling interval; Determine the first normal vector corresponding to each first point cloud data; The first normal vector is rotated so that the second normal vector points to the same side of the first point cluster, wherein the second normal vector includes the rotated first normal vector and the unrotated first normal vector; The outlier cloud corresponding to the first point cloud is determined based on the second normal vector and the preset outlier coefficient, wherein the outlier cloud includes multiple second point cloud data. The outlier cloud and the region where the first point cloud is located are divided into multiple sub-regions according to a recursive algorithm, wherein the number of point clouds in each sub-region is less than or equal to a preset threshold. For each sub-region, a spatial envelope of the sub-region is generated based on the outermost point of the sub-region; The local interpolation function of each spatial envelope is obtained by interpolating each spatial envelope using a multiharmonic spline radial basis local interpolation algorithm. The global interpolation implicit function corresponding to the target crack region is determined based on all local interpolation implicit functions; The isosurface corresponding to the interpolation when the global interpolation implicit function is equal to zero is determined in order to obtain a three-dimensional crack image of the target crack region.

2. The reconstruction method according to claim 1, characterized in that, The step of rotating the first normal vector so that the second normal vector points to the same side of the first point set includes: Determine a weighted undirected graph of the plurality of first point cloud data; The first normal vector is rotated according to the weighted undirected graph so that the second normal vector points to the same side of the first point set.

3. The reconstruction method according to claim 2, characterized in that, The step of rotating the first normal vector according to the weighted undirected graph so that the second normal vector points to the same side of the first point set includes: Determine the minimum spanning tree of the first point cloud based on the weighted undirected graph; The traversal sequence of the first point cloud data is determined based on the minimum spanning tree; The first normal vector is traversed according to the traversal sequence to determine the normal vector to be rotated in the first normal vector; The normal vector to be rotated is rotated to obtain the second normal vector.

4. The reconstruction method according to claim 1, characterized in that, The step of dividing the outlier cloud and the region containing the first point cloud into multiple sub-regions according to a recursive algorithm includes: The initial cube subdomain is determined based on the outlier cloud and the region where the first point cloud is located. Obtain the total number of point clouds in the initial cube subdomain; If the total amount of point cloud data exceeds the preset threshold, the initial cube subdomain is divided into a preset number of sub-cube subdomains. Obtain the amount of point cloud data in each sub-cube subdomain; If the amount of point cloud data in a sub-cube subdomain exceeds the preset threshold, the sub-cube subdomain is divided into a preset number of sub-cube subdomains so that the amount of point cloud data in each sub-cube subdomain is less than or equal to the preset threshold.

5. The reconstruction method according to claim 1, characterized in that, Also includes: After obtaining the first point cloud of the target crack region, the first point cloud is preprocessed to obtain the preprocessed second point cloud, wherein the preprocessing includes data registration and data cleaning.

6. The reconstruction method according to claim 1, characterized in that, Also includes: After obtaining the three-dimensional crack image, the void completion rate and the area of ​​the three-dimensional crack image are determined to determine the reconstruction effect of the three-dimensional crack image.

7. The reconstruction method according to claim 1, characterized in that, The step of generating the spatial envelope of each sub-region based on the outermost point of the sub-region includes: The outermost point cloud data of each sub-region is determined using the Alpha-shape algorithm; For each sub-region, the outermost point cloud data is connected into multiple triangular faces to form a spatial envelope composed of triangular faces.

8. A processor, characterized in that, It is configured to perform the computer vision-based three-dimensional crack reconstruction method for hydraulic fracturing test specimens as described in any one of claims 1 to 7.

9. A three-dimensional fracture reconstruction system for hydraulic fracturing experimental specimens based on computer vision, characterized in that, include: A laser scanning device is used to acquire the first point cluster of the target crack area; as well as The processor according to claim 8.

10. A machine-readable storage medium storing instructions thereon, characterized in that, When executed by a processor, this instruction causes the processor to be configured to perform the computer vision-based three-dimensional crack reconstruction method for hydraulic fracturing experimental specimens according to any one of claims 1 to 7.

Images