Multi-device data fusion three-dimensional fissure reconstruction method and system combining CT and three-dimensional scanner

Abstract

The application relates to the technical field of rock mechanics, and provides a multi-device data fusion three-dimensional fracture reconstruction method and system combining CT and a three-dimensional scanner, the relative positions of upper and lower fracture surface point cloud data of a prefabricated fractured rock sample obtained by a three-dimensional scanner are iteratively corrected based on a local fracture volume of a prefabricated fractured rock sample obtained by CT scanning, when the error between a local enclosed volume and the local fracture volume is less than or equal to a volume error threshold, the three-dimensional fracture of the fractured rock sample is reconstructed based on the relative positions of the upper and lower fracture surface point cloud data by converting the point cloud file format of the upper and lower fracture surface point cloud data, and the reconstruction result is judged, so that the advantages of CT and the three-dimensional scanner in fracture internal volume representation and surface geometry reconstruction are fully combined, error accumulation in sample fracture reconstruction is reduced, the geometry of a rock mass fracture is finely characterized, and the accuracy and stability of three-dimensional fracture reconstruction are improved.

Description

Technical Field

[0001] This application relates to the field of rock mechanics technology, and in particular to a method and system for three-dimensional fracture reconstruction by multi-device data fusion combining CT and three-dimensional scanners. Background Technology

[0002] Rock fissures are the main seepage channels in underground rock masses, and their geometric morphology and spatial distribution have a significant impact on underground engineering projects such as enhanced geothermal systems, shale oil and gas extraction, and low-grade carbon dioxide sequestration. The three-dimensional spatial characteristics within fissures determine the mechanical properties and seepage behavior of the rock mass, significantly controlling its stability. Furthermore, demonstrating that fissures, as weak structural planes within rock masses, also influence the evolution of damage zones and fault activation by affecting stress concentration, fissure penetration, and propagation patterns. Therefore, achieving high-precision reconstruction of rock fissures and accurately calculating their geometric characteristics is of significant engineering importance and research value for revealing the seepage-mechanical coupling process and evolution mechanism of rock masses. Summary of the Invention

[0003] The purpose of this application is to provide a method and system for three-dimensional crack reconstruction by combining multi-device data fusion of CT and 3D scanners, so as to solve or alleviate the problems existing in the prior art.

[0004] To achieve the above objectives, this application provides the following technical solution: This application provides a method for three-dimensional fracture reconstruction by combining multi-device data fusion of CT and 3D scanners, including: local fracture volume of a pre-fabricated fractured rock sample obtained from CT scans. The relative positions of the upper and lower fracture surface point cloud data of the precast fractured rock sample acquired by the 3D scanner are iteratively corrected; the local enclosed volume of the upper and lower fracture surface point cloud data is responded to. With local crack volume If the error is less than or equal to the volume error threshold, the three-dimensional fractures of the fractured rock sample are reconstructed based on the relative position of the point cloud data of the upper and lower fracture surfaces by converting the point cloud file format of the point cloud data of the upper and lower fracture surfaces, and the reconstruction results are evaluated.

[0005] Preferably, the fractured local area of ​​the pre-fabricated fractured rock sample is obtained by CT scanning, and the local fracture volume of the fractured rock sample is obtained based on an image segmentation algorithm. .

[0006] Preferably, CT scans are performed on the local fracture region of the precast fractured rock sample in a zero confining pressure fracture contact state to obtain three-dimensional grayscale voxel data of the local fracture region. A local threshold iterative segmentation method is used to achieve stable convergence segmentation of the three-dimensional gray voxel data of the local region of the crack. Based on the stable convergent partitioning results, according to the formula:

[0007] Determining the local fracture volume of a fractured rock sample In the formula, For the first The number of crack pixels in a CT slice that achieves stable convergence segmentation. For the first Voxel side length of a CT slice; This represents the total number of CT slices obtained from CT scans of a localized area of ​​the fissure. All are positive integers.

[0008] Preferably, the Ostu adaptive thresholding method is used to perform binary segmentation in the candidate crack region, and the segmentation result is subjected to morphological dilation operation to generate an extended mask region; Repeat the Ostu threshold segmentation iteration within the extended mask region until the change in the number of crack voxels in the two Ostu threshold segmentation results is less than the preset threshold, thus determining that the local crack region is stably converged and segmented.

[0009] Preferably, a point cloud dataset of the fractured upper surface of a pre-fabricated fractured rock sample is obtained using a 3D scanner. Point cloud dataset of the lower surface of the fracture ; Point cloud dataset of the upper surface of the crack Point cloud dataset of the lower surface of the fracture Perform point cloud registration so that the upper surface of the crack coincides with and is parallel to the lower surface of the crack; Vertical translation of the point cloud dataset on the upper surface of the crack By adjusting the relative positions of the point cloud data of the upper and lower fracture surfaces of the precast fractured rock specimen, the point cloud dataset of the upper fracture surface is calculated. Point cloud dataset of the lower surface of the fracture Local enclosed volume .

[0010] Preferably, the lower surface of the fractured rock sample is normalized to make the lower surface of the fractured rock sample geometrically aligned with the global coordinate system. The point cloud dataset on the upper surface of the crack was registered using the Normal Distribution Transform (NDT) point cloud registration algorithm. Point cloud dataset of the lower surface of the fracture Spatial registration is performed; among which, rigid body transformation is used to constrain the upper and lower surfaces of the crack to ensure optimal spatial alignment of the upper and lower surfaces of the crack in a unified reference coordinate system.

[0011] Preferably, the registered point cloud file format of the upper and lower surfaces is converted into equally spaced two-dimensional matrix text data, all point clouds are rasterized according to equally spaced differences, and then processed according to the formula:

[0012] Computation of point cloud dataset on the upper surface of the crack Point cloud dataset of the lower surface of the fracture Local enclosed volume ; In the formula, The first grid cell obtained by dividing the grid into equally spaced grid units In the multiple tetrahedral elements of the local grid, the first The volume of a tetrahedral unit cell The total number of local grid cells obtained by dividing the grid into equally spaced grid units. All are positive integers; The points obtained by converting point cloud file formats are respectively the first two points. The point cloud coordinates of the vertices of a tetrahedral unit.

[0013] Preferably, the local enclosed volume responds to the point cloud data of the upper and lower fracture surfaces. With local crack volume If the error is less than or equal to the volume error threshold, the relative distance between the point cloud data of the upper and lower fracture surfaces is determined as the true fracture aperture of the fractured rock sample under the condition of no initial stress. The geometric characteristics of the fractured rock sample were calculated using the equally spaced two-dimensional matrix text data obtained by converting the point cloud file format; wherein, the geometric characteristics of the fractured rock sample include at least the arithmetic mean of the fracture aperture, the standard deviation, the relative roughness coefficient, and the true contact area.

[0014] Preferably, an elastoplastic iso-contact model is used to study the three-dimensional crack in... The crack closure process under normal load was simulated to obtain the three-dimensional crack contact area distribution; The fracture surface of the three-dimensional fracture is divided into multiple uniform grids, and then processed according to the formula:

[0015] Calculate the contact uniformity index of the contact area within each grid. Among them, the contact uniformity index Used to characterize the uniformity of the contact distribution on the upper and lower surfaces of a crack; In the formula, This represents the average of the standard deviations of the contact area of ​​the grid cells along the horizontal direction. This represents the average of the standard deviations of the contact area of ​​the grid cells along the vertical direction; Response to contact uniformity index Then the three-dimensional crack is in Under normal load, the relative positions of the upper and lower surfaces of the crack do not deflect.

[0016] This embodiment also provides a multi-device data fusion three-dimensional crack reconstruction system combining CT and 3D scanners. The system reconstructs three-dimensional cracks using any of the aforementioned multi-device data fusion three-dimensional crack reconstruction methods combining CT and 3D scanners. The system includes: The point cloud correction unit is configured to measure the local fracture volume of a pre-fabricated fractured rock sample obtained from a CT scan. The relative positions of the upper and lower fracture surface point cloud data of the prefabricated fractured rock sample acquired by the 3D scanner are iteratively corrected. The fracture reconstruction and evaluation unit is configured to respond to the local enclosing volume of the upper and lower fracture surface point cloud data. With local crack volume If the error is less than or equal to the volume error threshold, the three-dimensional fractures of the fractured rock sample are reconstructed based on the relative position of the point cloud data of the upper and lower fracture surfaces by converting the point cloud file format of the point cloud data of the upper and lower fracture surfaces, and the reconstruction results are evaluated.

[0017] Beneficial effects: The present application provides a method and system for three-dimensional fracture reconstruction by combining multi-device data fusion of CT and 3D scanners, based on the local fracture volume of a pre-fabricated fractured rock sample obtained by CT scanning. The relative positions of the point cloud data of the upper and lower fracture surfaces of the precast fractured rock sample acquired by the 3D scanner are iteratively corrected. When the local enclosed volume of the point cloud data of the upper and lower fracture surfaces is... With local crack volume When the error is less than or equal to the volume error threshold, the three-dimensional fractures of the fractured rock sample are reconstructed based on the relative positions of the point cloud data of the upper and lower fracture surfaces by converting the point cloud file format of the point cloud data. The reconstruction results are then evaluated. This allows for the assessment of the locally enclosed volume. With local crack volume Under consistent constraints, the relative positions of the fracture surfaces are iteratively corrected to determine the true geometric state of the fracture under no initial stress conditions. The advantages of CT and 3D scanner in internal volume characterization and surface geometry reconstruction of fractures are fully integrated to reduce the accumulation of errors in fracture reconstruction of samples, refine the geometric features of rock fractures, and improve the accuracy and stability of 3D fracture reconstruction. Attached Figure Description

[0018] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. Wherein: Figure 1 This is a flowchart illustrating a multi-device data fusion method for three-dimensional crack reconstruction combining CT and a 3D scanner, according to some embodiments of this application. Figure 2 A logic diagram of a multi-device data fusion three-dimensional crack reconstruction method combining CT and 3D scanners provided according to some embodiments of this application; Figure 3 Point cloud dataset of the upper surface of a crack provided according to some embodiments of this application Point cloud dataset of the lower surface of the fracture A schematic diagram of spatial registration; Figure 4 This is a schematic diagram illustrating the local volume of a crack obtained by CT scanning a specimen, according to some embodiments of this application. Figure 5 The local fracture volume is defined by vertically translating the point cloud data of the upper and lower fracture surfaces provided according to some embodiments of this application. A schematic diagram; Figure 6 The local fracture volume enclosed by point cloud data of upper and lower fracture surfaces provided according to some embodiments of this application. The calculation principle diagram; Figure 7 This is a comparative schematic diagram showing the results of reconstructing three-dimensional cracks using different reconstruction methods according to some embodiments of this application; Figure 8 This is a schematic diagram illustrating the calculation of crack contact area uniformity according to some embodiments of this application; Figure 9 This is a schematic diagram of a multi-device data fusion three-dimensional crack reconstruction system combining CT and a three-dimensional scanner, according to some embodiments of this application. Detailed Implementation

[0019] The present application will now be described in detail with reference to the accompanying drawings and embodiments. Various examples are provided by way of explanation and not by way of limitation. In fact, those skilled in the art will understand that modifications and variations can be made to the present application without departing from the scope or spirit of the present application. For example, a feature shown or described as part of one embodiment may be used in another embodiment to produce yet another embodiment. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention should fall within the scope of protection of the embodiments of the present invention.

[0020] Currently, there are two main methods for reconstructing three-dimensional fractures in rocks: one is to use CT scans to examine fractured samples and extract the fractures by image segmentation of the scan results; the other is to use a 3D scanner to scan the upper and lower fractures and the outer reference point, and then stitch the upper and lower fracture surfaces to the outer reference frame by point cloud stitching, and use the relative position of the stitched upper and lower fracture surfaces as the true fracture opening.

[0021] However, on the one hand, the standard specimens commonly used in rock mechanics experiments have a diameter of... ,high For cylindrical samples, after using Brazilian splitting to pre-fabricate fractures, the overall size of the sample is relatively large compared to the CT scan's field of view, resulting in limited pixel resolution in the reconstructed fracture results. The fracture aperture exhibits obvious pixel dispersion characteristics, leading to poor accuracy in subsequent fracture image segmentation and quantitative extraction. Furthermore, when scanning large-sized samples, 3D scanners also require multiple scans and stitching, causing errors to accumulate and resulting in relatively large errors in the obtained fracture aperture.

[0022] Based on this, this embodiment provides a multi-device data fusion method for three-dimensional crack reconstruction that combines CT and 3D scanner data. By utilizing CT and 3D scanner data, the three-dimensional crack is reconstructed through local equivalent volume. This fully combines the advantages of CT in quantitative extraction of the internal volume of the crack with the advantages of 3D scanner in high-precision acquisition of aggregate features on the crack surface. This allows for high-precision reconstruction of the overall three-dimensional geometric structure of the crack by performing CT scanning on a local area of ​​the crack, significantly improving the accuracy and reliability of three-dimensional crack reconstruction.

[0023] like Figures 1 to 8 As shown, this multi-device data fusion method for three-dimensional crack reconstruction combining CT and 3D scanners includes: Step S101: Local fracture volume of prefabricated fractured rock sample obtained based on CT scan. The relative positions of the upper and lower fracture surface point cloud data of the prefabricated fractured rock sample acquired by the 3D scanner are iteratively corrected.

[0024] In this embodiment, rock samples with fractures prepared by the Brazilian splitting test are selected, or rock samples with natural or artificially interconnected fractures are selected as pre-fractured rock samples. In the Brazilian splitting test, standard cylindrical rock samples are subjected to Brazilian splitting to form pre-prepared tensile fractures.

[0025] In a specific example, choose the diameter ,high Cylindrical fine-grained granite rock sample (average mineral grain size approximately) The Brazilian splitting test was used to form tensile through-cracks inside the sample. Generally, the smaller the rock mineral grain size, the lower the initial aperture of the crack, and the higher the accuracy of identifying and reconstructing the crack geometry.

[0026] The rock sample was split into two matching fracture surfaces, which were then labeled as the upper fracture surface. With the lower surface of the crack Next, the upper surface of the fracture in the pre-fabricated fractured rock sample was... With the lower surface of the crack The two fracture surfaces are fitted together to restore their spatial correspondence to that before splitting. Simultaneously, adhesive tape is lightly wrapped around the upper and lower ends of the cylindrical rock sample to restrict overall sample displacement without imposing additional normal constraints on the fracture surfaces. This simulates the true contact state of the fractures under initial zero confining pressure without introducing external loads. Here, the above-mentioned method of fixing the rock sample is only used to maintain the relative stability of the fracture surfaces and does not affect the natural distribution of fracture aperture and geometry.

[0027] Next, the bonded rock sample was placed in a CT scanning device, and the central region of the rock sample (approximately [height missing]) was scanned. The scanning range was used to acquire three-dimensional grayscale voxel data containing the local area of ​​the fracture. In a specific example, a total of 1400 TIFF format cross-sectional grayscale images were obtained through CT scanning, with a single image size of [size missing]. The pixel resolution is Here, 1300 grayscale images located in the middle of the scanned area are selected as the image dataset for subsequent fracture body data processing to reduce imaging errors of cone-beam CT in the scanned edge areas and improve data quality.

[0028] Then, image segmentation processing was performed on the cross-sectional grayscale images exported from CT scans to extract the fracture region. First, beam hardening correction was performed on the grayscale images to eliminate grayscale unevenness during imaging; then, median filtering and contrast-limited adaptive histogram equalization (CLAHE) were applied sequentially to reduce noise interference and enhance the grayscale contrast between the fracture and the matrix.

[0029] In this process, a local threshold iterative segmentation method is used to extract the fractures. Specifically, the three-dimensional grayscale voxel data of the local fracture region is stably converged using an image segmentation algorithm, and the local fracture volume of the fractured rock sample is obtained based on the segmentation results. In a specific example, when using the local threshold iterative segmentation method to stably converge the segmentation of the 3D grayscale voxel data of the local region of the crack, the initial threshold is first preset. The grayscale image is initially segmented to obtain candidate regions containing cracks. Since the range of the initial segmented region is larger than the actual crack region, the Ostu adaptive thresholding method is used for binary segmentation within the crack candidate region. The binary segmentation result is then subjected to morphological dilation to remove isolated small-scale discrete regions and enhance the connectivity of the crack region, generating an extended mask region. Subsequently, the Ostu thresholding segmentation iteration is repeated within the extended mask region until the change in the number of crack voxels between the two Ostu thresholding segmentation results is less than a preset threshold, thus determining that the local crack region is stably converged.

[0030] Furthermore, based on the stable convergent partitioning results, according to the formula:

[0031] Determining the local fracture volume of a fractured rock sample In the formula, For the first The number of crack pixels in a CT slice that achieves stable convergence segmentation. For the first Voxel side length of a CT slice; This represents the total number of CT slices obtained from CT scans of a localized area of ​​the fissure. All are positive integers.

[0032] In this embodiment, a 3D scanner with calibrated equipment and corrected parameters is used to examine the upper surface of the rock sample's fractures. With the lower surface of the crack Three-dimensional scanning was performed to obtain the fractured upper surface of the pre-fabricated fractured rock sample. With the lower surface of the crack The initial point cloud data is used, and denoising, outlier removal, data smoothing, downsampling, and mesh reconstruction are performed on the initial point cloud data to obtain a continuous and complete geometric model of the fracture surface, corresponding to the point cloud dataset of the fracture surface. Point cloud dataset of the lower surface of the fracture Before scanning the surface of the fractures formed by splitting, the surface of the fractures is first pre-processed: specifically, a thin, uniform coating of developer is applied to the surface of the fractures to reduce interference from mineral reflections and improve the quality and completeness of the point cloud data acquired from the fracture surface.

[0033] Point cloud dataset of the upper surface of the crack obtained by scanning Point cloud dataset of the lower surface of the fracture In a spatial coordinate system, there are differences in attitude and position. In this embodiment, the point cloud dataset on the upper surface of the crack is used. Point cloud dataset of the lower surface of the fracture Point cloud registration is performed to ensure that the upper and lower surfaces of the fracture coincide and are parallel. Firstly, the orientation of the lower surface of the precast fractured rock sample is normalized to align it with the global coordinate system. Planar geometric alignment. Specifically, for the point cloud dataset of the fracture's lower surface. Perform best-fit plane fitting and spatial coordinate transformation based on the best-fit plane to obtain the point cloud dataset of the fracture lower surface. The fitting plane and the global coordinate system The planes remain parallel, thus achieving orientation standardization of the lower surface of the crack.

[0034] In a specific example, the point cloud dataset of the lower surface of the crack. Optimal plane fitting calculation is performed. Using a least-squares plane fitting method, the point cloud covariance matrix is ​​constructed and eigenvalue decomposition is performed. The eigenvector corresponding to the smallest eigenvalue is taken as the normal vector of the optimal fitting plane for the fracture's lower surface. This determines the plane equation. Furthermore, based on the normal vector of the best-fit plane and its coordinates in the global coordinate system... The spatial relationship between the plane normal vectors is used to construct a rotation transformation matrix for the point cloud dataset on the lower surface of the crack. Perform overall attitude adjustments to make the best-fit plane align with... By keeping the planes parallel, the pose of the point cloud on the lower surface of the fracture is standardized, providing a unified reference coordinate system for the subsequent spatial registration and aperture calculation of the upper and lower surfaces of the fracture.

[0035] Since both the upper and lower fracture surfaces are formed by Brazilian splitting, they have strong complementary features in terms of geometry. In this embodiment, the point cloud set on the upper surface of the fracture is directly spliced ​​to the point cloud set on the lower surface of the fracture through point cloud registration, thereby ensuring that the upper and lower fracture surfaces remain relatively parallel in space.

[0036] Specifically, the Normal Distribution Transform (NDT) point cloud registration algorithm is used to register the point cloud dataset on the upper surface of the crack. Point cloud dataset of the lower surface of the fracture Spatial registration is performed. This involves constraining the upper and lower surfaces of the fracture through rigid body transformation to ensure optimal spatial alignment between them in a unified reference coordinate system. Specifically, the rigid body transformation constrains the upper and lower surfaces of the fracture to maintain a parallel relationship, ensuring that the average normal distance between the registered upper and lower surfaces is close to zero.

[0037] In a specific example, by dividing the reference point cloud into a regular 3D voxel mesh, a normal distribution model is constructed within each voxel cell using the statistical properties of the point cloud, with the probability distribution field serving as the registration objective function. For any point within a voxel, its spatial coordinates... It follows a multivariate normal distribution, and its probability density function is expressed as:

[0038] In the formula, Let be the mean vector of the point cloud within the voxel. This corresponds to the covariance matrix. Based on this, a rigid body pose change matrix is ​​constructed. The objective function that maximizes the overall likelihood function of the transformed point cloud in the probability distribution field is expressed as:

[0039] in, Represents the rigid body pose change matrix Point cloud coordinates after the action. Objective function. The optimal rigid body transformation parameters are obtained by solving the problem using Newton's iterative method. Compared with traditional registration methods based on point-to-point or point-to-plane methods, the orthogonal distribution transformation point cloud registration algorithm utilizes local statistical distribution information for matching, which is more robust to the incomplete consistency of the local morphology of the upper and lower surfaces of the fracture and can effectively reduce the risk of getting trapped in local optima. After point cloud registration, the transformed upper surface point cloud set closely fits the lower surface point cloud set. At this point, it can be assumed that there is no significant relative deflection or lateral displacement between the upper and lower surfaces of the fracture, and the overall fracture aperture approaches zero. In this way, by ensuring the parallel poses of the upper and lower fracture surfaces, the distribution consistency of the fracture aperture is effectively improved, overcoming the problem of uneven local aperture distribution caused by fracture surface pose deviation in traditional methods, and improving the reliability of the lightning system geometric reconstruction structure.

[0040] Finally, the point cloud dataset on the upper surface of the crack was vertically translated. By adjusting the relative positions of the point cloud data of the upper and lower fracture surfaces of the precast fractured rock specimen, the point cloud dataset of the upper fracture surface is calculated. Point cloud dataset of the lower surface of the fracture Local enclosed volume Specifically, the point cloud coordinate data of the registered upper and lower surfaces of the crack are exported separately, and the point cloud files of the registered upper and lower surfaces are converted into equally spaced two-dimensional matrix text data. In the equally spaced two-dimensional matrix text data, each value represents the corresponding position. The coordinate values ​​of the direction, the row and column indices of the matrix correspond to the fracture surface. Spatial position in a direction.

[0041] Furthermore, based on equally spaced two-dimensional matrix text data, according to the formula:

[0042] Computation of point cloud dataset on the upper surface of the crack Point cloud dataset of the lower surface of the fracture Local enclosed volume In the formula, The first grid cell obtained by dividing the grid into equally spaced grid units In the multiple tetrahedral elements of the local grid, the first The volume of a tetrahedral unit cell The total number of local grid cells obtained by dividing the grid into equally spaced grid units. All are positive integers; The points obtained by converting point cloud file formats are respectively the first two points. The point cloud coordinates of the vertices of a tetrahedral unit.

[0043] In a specific example, adopt The equally spaced grid is used to generate the matrix corresponding to the fracture surface using the Kriging interpolation method. The coordinate values ​​of the direction; the row and column indices of the matrix correspond to the fracture surface. The planar spatial position in the direction. In this way, irregularly distributed 3D point cloud data is transformed into structured regular raster data, so as to realize the correspondence between the upper and lower surfaces of the crack in the same coordinate system, and provide a unified data foundation for crack aperture calculation, volume integral summation and quantitative analysis of related geometric parameters based on raster cells.

[0044] Step S102: Local enclosed volume in response to the point cloud data of the upper and lower fracture surfaces With local crack volume The error is less than or equal to the volume error pre-formulation. Based on the conversion result of the point cloud file format of the upper and lower fracture surface point cloud data, the three-dimensional fracture of the fractured rock sample is reconstructed according to the relative position of the upper and lower fracture surface point cloud data, and the reconstruction result is evaluated.

[0045] After point cloud registration of the upper and lower surfaces of the fracture, the upper and lower fracture surfaces are spatially parallel and aligned, with an average fracture aperture close to zero. To obtain the accurate true geometric state of the fracture under zero confining pressure initial conditions, in this embodiment, while ensuring that the spatial range of the local volume calculation area obtained through 3D scanning is consistent with the fracture volume data obtained through CT scanning, the translation distance of the upper surface of the fracture is iteratively adjusted, and the local confining volume calculation is repeated. The calculation continues until the local enclosed volume is obtained. Local fracture volume obtained from CT scan If the error is less than or equal to the volume error threshold, the relative distance between the point cloud data of the upper and lower fracture surfaces is determined as the true fracture opening of the fractured rock sample under the condition of no initial stress.

[0046] Specifically, while maintaining the parallel relationship between the upper and lower fracture surfaces, a gradual vertical translation is applied to the upper fracture surface along the normal (Z direction). At each translation step, the fracture volume enclosed by the upper and lower fracture surfaces within a selected local area of ​​the CT scan (a 50 mm × 46 mm rectangular area in this embodiment) is calculated. When the calculated local fracture volume matches the fracture volume extracted based on CT grayscale data segmentation, it can be determined that the relative position of the upper and lower fracture surfaces at this time is the true fracture geometry under the initial state of zero confining pressure, thus obtaining a reliable initial fracture aperture distribution.

[0047] By using iterative volume constraints, the systematic errors introduced by repeated scanning due to relying solely on point cloud registration are effectively avoided, providing reliable initial conditions for the accurate calculation of fracture aperture distribution, volume evolution, and related geometric parameters. Compared to existing methods that require additional scanning of the outer rock wall and the contact state reference coordinate system, the method in this embodiment only needs to perform three-dimensional scanning of the upper and lower surfaces of the fracture to complete fracture reconstruction and aperture calculation, significantly reducing the number of scans, effectively reducing the error accumulation caused by multiple scans, and improving the accuracy of fracture aperture calculation.

[0048] After obtaining the true fracture aperture, matrix data of the upper and lower surfaces of the fracture are obtained separately. The geometric characteristics of the fractured rock sample are calculated by converting the point cloud file format to obtain equidistant two-dimensional matrix text data. The geometric characteristics of the fractured rock sample include at least the arithmetic mean and standard deviation of the fracture aperture, the relative roughness coefficient, and the true contact area.

[0049] In this embodiment, by comparing and analyzing different reconstruction methods on the fractures of fine-grained granite, the corresponding local cross-sectional morphology of the fractures was obtained. Furthermore, by injecting epoxy resin into the fractures and preparing thin rock sections, the fracture aperture was statistically measured under a microscope, and the root mean square error was used as the metric. The reconstruction accuracy of different 3D crack reconstruction methods was compared as an evaluation index.

[0050] In the formula, To reconstruct the crack at a uniform sampling interval, the first... The aperture value at each discrete point This represents the true aperture at the corresponding position as measured by the microscope. This represents the total number of data points after the thin section profile is discretized at the same sampling interval.

[0051] Simultaneously, by calculating the uniformity of the contact area distribution under a three-dimensional crack normal load, the accuracy of crack reconstruction can be quickly determined. First, an elastoplastic iso-contact model is used to assess the uniformity of the contact area distribution under a three-dimensional crack. The crack closure process under normal load was simulated to obtain the contact area distribution of the three-dimensional crack. Specifically, the contact state of the three-dimensional crack under different confining pressures was calculated using boundary element method based on Fourier transform, and the contact distribution of the crack under different confining pressures was obtained.

[0052] Then, the fracture surface of the three-dimensional fracture is divided into multiple uniform meshes. In a specific example, the fracture surface is obtained in... The contact area distribution of the fracture surface under confining pressure is shown, where the contact area and the non-contact area are marked with 1 and 0 respectively, dividing the entire fracture surface into uniform areas. For each grid cell, the area ratio of the contact region within that cell is statistically analyzed and calculated. Based on this, the formula is used:

[0053] Calculate the contact uniformity index of the contact area within each grid. The spatial distribution characteristics of the crack contact area are quantitatively characterized. Here, the contact uniformity index... Used to characterize the uniformity of contact distribution on the upper and lower surfaces of a crack; when the contact uniformity index Then the three-dimensional crack is in Under normal load, the relative positions of the upper and lower surfaces of the crack remain unchanged. In the formula, This represents the average of the standard deviations of the contact area of ​​the grid cells along the horizontal direction. This represents the average of the standard deviations of the contact area of ​​the grid cells along the vertical direction.

[0054] In this embodiment, by fusing the volume data inside the fracture from CT scans with the geometric data of the fracture surface obtained from a 3D scanner, the advantages of CT in volume calculation and the scanner in surface morphology methods are fully utilized to achieve multi-source data fusion, improve the accuracy of fracture reconstruction, and avoid the limitation of CT scanning range and resolution. It effectively overcomes the technical problems of reconstructing 3D fractures in existing fracture reconstruction methods, such as the large number of scans, easy accumulation of errors, and limited CT scanning range and resolution. It significantly improves the accuracy and stability of the reconstruction of the 3D geometric structure and aperture distribution of fractures, and realizes the efficient acquisition of the real geometric parameters of fractures under multi-source data conditions, making the characterization of rock fracture structure at both laboratory and field scales more reliable.

[0055] This embodiment also provides a multi-device data fusion three-dimensional fracture reconstruction system combining CT and 3D scanners, such as... Figure 9As shown, a three-dimensional crack reconstruction method combining CT and 3D scanner data fusion is used in any of the above embodiments to perform three-dimensional crack reconstruction. The system includes: Point cloud correction unit 901 is configured to measure the local fracture volume of a pre-fabricated fractured rock sample obtained from CT scans. The relative positions of the upper and lower fracture surface point cloud data of the prefabricated fractured rock sample acquired by the 3D scanner are iteratively corrected. The fracture reconstruction and evaluation unit 902 is configured to respond to the local enclosing volume of the upper and lower fracture surface point cloud data. With local crack volume If the error is less than or equal to the volume error threshold, the three-dimensional fractures of the fractured rock sample are reconstructed based on the relative position of the point cloud data of the upper and lower fracture surfaces by converting the point cloud file format of the point cloud data of the upper and lower fracture surfaces, and the reconstruction results are evaluated.

[0056] The multi-device data fusion three-dimensional crack reconstruction system combining CT and 3D scanners in this embodiment can realize the steps and processes of the multi-device data fusion three-dimensional crack reconstruction method combining CT and 3D scanners in any of the above embodiments, and achieve the same technical effect, which will not be described in detail here.

[0057] In the description of this invention, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0058] The above description is merely a preferred embodiment of this application and is not intended to limit 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 protection scope of this application.

Claims

1. A method for three-dimensional crack reconstruction by multi-device data fusion combining CT and 3D scanners, characterized in that, include: Local fracture volume of pre-fabricated fractured rock specimens obtained from CT scans The relative positions of the upper and lower fracture surface point cloud data of the prefabricated fractured rock sample acquired by the 3D scanner are iteratively corrected. Local enclosed volume in response to point cloud data of upper and lower fracture surfaces With local crack volume If the error is less than or equal to the volume error threshold, the three-dimensional fractures of the fractured rock sample are reconstructed based on the relative position of the point cloud data of the upper and lower fracture surfaces by converting the point cloud file format of the point cloud data of the upper and lower fracture surfaces, and the reconstruction results are evaluated.

2. The method according to claim 1, characterized in that, The fractured rock specimen was examined using CT scans of local fractured regions, and the local fracture volume was obtained based on an image segmentation algorithm. .

3. The method according to claim 2, characterized in that, CT scans were performed on the local fracture region of a precast fractured rock sample in a zero-containment-pressure fracture contact state to obtain three-dimensional grayscale voxel data of the local fracture region. A local threshold iterative segmentation method is used to achieve stable convergence segmentation of the three-dimensional gray voxel data of the local region of the crack. Based on the stable convergent partitioning results, according to the formula: ； Determining the local fracture volume of a fractured rock sample In the formula, For the first The number of crack pixels in a CT slice that achieves stable convergence segmentation. For the first Voxel side length of a CT slice; This represents the total number of CT slices obtained from CT scans of a localized area of ​​the fissure. All are positive integers.

4. The method according to claim 3, characterized in that, Ostu adaptive thresholding method is used to perform binary segmentation in the candidate crack region, and the segmentation result is subjected to morphological dilation operation to generate an extended mask region. Repeat the Ostu threshold segmentation iteration within the extended mask region until the change in the number of crack voxels in the two Ostu threshold segmentation results is less than the preset threshold, thus determining that the local crack region is stably converged and segmented.

5. The method according to claim 1, characterized in that, A point cloud dataset of the fractured upper surface of a pre-fabricated fractured rock sample was obtained using a 3D scanner. Point cloud dataset of the lower surface of the fracture ; Point cloud dataset of the upper surface of the crack Point cloud dataset of the lower surface of the fracture Perform point cloud registration so that the upper surface of the crack coincides with and is parallel to the lower surface of the crack; Vertical translation of the point cloud dataset on the upper surface of the crack By adjusting the relative positions of the point cloud data of the upper and lower fracture surfaces of the precast fractured rock specimen, the point cloud dataset of the upper fracture surface is calculated. Point cloud dataset of the lower surface of the fracture Local enclosed volume .

6. The method according to claim 5, characterized in that, The orientation of the lower surface of the precast fractured rock specimen is normalized so that the lower surface of the precast fractured rock specimen is geometrically aligned with the global coordinate system. The point cloud dataset on the upper surface of the crack was registered using the Normal Distribution Transform (NDT) point cloud registration algorithm. Point cloud dataset of the lower surface of the fracture Spatial registration is performed; among which, rigid body transformation is used to constrain the upper and lower surfaces of the crack to ensure optimal spatial alignment of the upper and lower surfaces of the crack in a unified reference coordinate system.

7. The method according to claim 5, characterized in that, The registered point cloud files of the upper and lower surfaces are converted into equally spaced two-dimensional matrix text data. All point clouds are rasterized according to equally spaced differences, and then processed according to the formula: ； Computation of point cloud dataset on the upper surface of the crack Point cloud dataset of the lower surface of the fracture Local enclosed volume ; In the formula, The first grid cell obtained by dividing the grid into equally spaced grid units In the multiple tetrahedral elements of the local grid, the first The volume of a tetrahedral unit cell The total number of local grid cells obtained by dividing the grid into equally spaced grid units. All are positive integers; The points obtained by converting point cloud file formats are respectively the first two points. The point cloud coordinates of the vertices of a tetrahedral unit.

8. The method according to claim 7, characterized in that, Local enclosed volume in response to point cloud data of upper and lower fracture surfaces With local crack volume If the error is less than or equal to the volume error threshold, the relative distance between the point cloud data of the upper and lower fracture surfaces is determined as the true fracture aperture of the fractured rock sample under the condition of no initial stress. The geometric characteristics of the fractured rock sample were calculated using the equally spaced two-dimensional matrix text data obtained by converting the point cloud file format; wherein, the geometric characteristics of the fractured rock sample include at least the arithmetic mean of the fracture aperture, the standard deviation, the relative roughness coefficient, and the true contact area.

9. The method according to claim 1, characterized in that, Using an elastoplastic iso-contact model, the three-dimensional crack in... The crack closure process under normal load was simulated to obtain the three-dimensional crack contact area distribution; The fracture surface of the three-dimensional fracture is divided into multiple uniform grids, and then processed according to the formula: ； Calculate the contact uniformity index of the contact area within each grid. Among them, the contact uniformity index Used to characterize the uniformity of the contact distribution on the upper and lower surfaces of a crack; In the formula, This represents the average of the standard deviations of the contact area of ​​the grid cells along the horizontal direction. This represents the average of the standard deviations of the contact area of ​​the grid cells along the vertical direction; Response to contact uniformity index Then the three-dimensional crack is in Under normal load, the relative positions of the upper and lower surfaces of the crack do not deflect.

10. A multi-device data fusion three-dimensional crack reconstruction system combining CT and three-dimensional scanners, characterized in that, The system employs any one of the multi-device data fusion three-dimensional crack reconstruction methods combining CT and a three-dimensional scanner as described in claims 1-9, and comprises: The point cloud correction unit is configured to measure the local fracture volume of a pre-fabricated fractured rock sample obtained from a CT scan. The relative positions of the upper and lower fracture surface point cloud data of the prefabricated fractured rock sample acquired by the 3D scanner are iteratively corrected. The fracture reconstruction and evaluation unit is configured to respond to the local enclosing volume of the upper and lower fracture surface point cloud data. With local crack volume If the error is less than or equal to the volume error threshold, the three-dimensional fractures of the fractured rock sample are reconstructed based on the relative position of the point cloud data of the upper and lower fracture surfaces by converting the point cloud file format of the point cloud data of the upper and lower fracture surfaces, and the reconstruction results are evaluated.

Images