CN115546116A - Full coverage rock mass discontinuity extraction and spacing calculation method and system - Google Patents

Abstract

The invention provides a method and a system for extracting discontinuous surfaces and calculating intervals of a fully-covered rock mass, which are used for acquiring fracture image data exposed in a linear form and fracture image data exposed in a planar form; obtaining a fitted linear discontinuous surface according to the crack image data exposed in a linear form; obtaining a planar discontinuous surface according to the crack image data exposed in a planar form; obtaining a discontinuous surface interval according to the obtained linear discontinuous surface and the planar discontinuous surface; the invention divides the discontinuous surface of the rock mass into two discontinuous surfaces which are exposed in a surface shape and exposed in a trace line, extracts the fracture through intelligent identification and reconstruction, converts the image into fracture characteristic points, extracts the discontinuous surface through an algorithm, and calculates the distance between the discontinuous surfaces, thereby realizing the representation of discontinuous distribution.

Description

Full coverage rock mass discontinuity extraction and spacing calculation method and system

technical field

The invention relates to the technical field of discontinuous surface analysis of rock mass, in particular to a method and system for extracting discontinuous surfaces of rock mass and calculating distances with full coverage.

Background technique

The statements in this section merely provide background art related to the present invention and do not necessarily constitute prior art.

Rock mass joint planes include joints, faults, bedding planes, etc. The presence of joint plane properties (i.e., spacing, persistence, roughness, filling, weathering, and presence of water) has a significant impact on the mechanical behavior and permeability of rock mass . At present, there are few studies on the extraction of discontinuity spacing in rock mass, but the size of discontinuity spacing has a great influence on the discontinuity of rock mass. The same discontinuity size and different distribution states have different influences on engineering construction. The traditional method uses a tape measure compass to characterize the joint surface of rock mass, especially the acquisition of joint spacing takes a lot of time and energy, and the efficiency is low. Multiple conditional restrictions.

The inventors found that most of the existing solutions obtain rock mass discontinuities through 3D point clouds, but this method requires a large amount of data and requires high calculation equipment, and the 3D point cloud lacks RGB values, which is less representative of rock mass fissures. Poor; on the other hand, the discontinuous surface of the tunnel face or slope rock mass not only has cracks exposed in the form of traces, but also cracks exposed in the form of planes. Poor adaptability, which in turn leads to poor accuracy of characterization results.

Contents of the invention

In order to solve the deficiencies of the prior art, the present invention provides a method and system for extracting discontinuous surfaces of rock mass and calculating distances with full coverage, which divides the discontinuous surfaces of rock mass into two types: planar outcropping and trace outcropping. A kind of discontinuous surface, through intelligent identification and reconstruction, cracks are extracted, and the image is converted into crack feature points, the discontinuous surface is extracted through the algorithm, and the distance is calculated to realize the characterization of the discontinuous distribution.

In order to achieve the above object, the present invention adopts the following technical solutions:

The first aspect of the present invention provides a method for extracting discontinuous surfaces of a fully covered rock mass and calculating distances.

A full-coverage discontinuity surface extraction and distance calculation method of rock mass, including the following processes:

Obtain image data of cracks exposed in a linear form and image data of cracks exposed in a planar form;

According to the crack image data exposed in a linear form, combined with the semantic segmentation model, the rock mass crack recognition result map is obtained, and the crack pixels in the rock mass crack recognition result map are replaced with the pixels corresponding to the original image to obtain the crack recognition result fusion map According to the fusion map of the crack identification results, the 3D reconstruction of the rock mass cracks is carried out. According to the 3D reconstruction results, the 3D coordinates of the crack points are extracted, and the feature points of the cracks are fitted by a random convex polygon fitting algorithm to obtain the fitted linear irregularities. continuous surface;

According to the crack image data exposed in planar form, the neural network model is used to identify the cracks, and the feature points of the exposed discontinuous surface are obtained through the reconstruction algorithm, and the coplanar parameter calculation, plane fitting, plane generation and coplanarity After the surface test, a planar discontinuous surface is obtained;

From the obtained linear discontinuous surfaces and planar discontinuous surfaces, the distance between discontinuous surfaces is obtained.

As an optional implementation method, the fusion map of the crack recognition results is used as the input of the 3D reconstruction algorithm, and sparse modeling is performed after feature point extraction, feature point matching, matching optimization, triangulation, pose estimation, and BA optimization. , after depth map estimation and optimization, the 3D reconstruction result of dense modeling is obtained.

As an optional implementation method, the crack feature points are extracted using the different RGB values of the crack feature points in the fusion image, and the three-dimensional coordinates of the crack feature points are obtained.

As an optional implementation method, lock a point on the discontinuity surface, use the KNN algorithm to search for nearby feature points according to the search function and Euclidean distance, named as Q _i , after obtaining k nearby feature points, use the PCA algorithm to cluster the plane Perform detection, obtain plane normal vectors (a,b,c) and eigenvalues (λ ₁ ,λ ₂ ,λ ₃ ), and define deviation parameters

When η is less than the maximum allowable maximum deviation parameter, the coplanar condition is met, and the plane equation of the clustered discontinuous surface is obtained by SVD singular value algorithm decomposition and fitting;

The PCA algorithm obtains the plane normal vector of each cluster, converts the plane normal vector into a stereo projection, observes the density of feature points, and obtains the local maximum value of the density, and uses the normal vector at the local maximum value of the density as a discontinuous surface The main pole direction of the discontinuity surface is determined as the main plane of the cluster discontinuity surface, and then the master equation of the discontinuity surface is obtained.

As an optional implementation method, according to the obtained linear discontinuous surface and planar discontinuous surface, the discontinuous surface spacing is obtained, including:

The clustering feature point of the linear discontinuous surface and the planar discontinuous surface is P, the D value of all points is named D _p , and the clustering attribute is defined as D _c ;

Obtain the cluster list according to the ascending order of the D value of P points, randomly select a discontinuous surface and define it as cl ₁ , create a feature point data set R except cl ₁ , search the clusters in R, and find the feature point closest to cl ₁ The discontinuous surface where it is located is cl _j ;

After obtaining cl _j , define all the feature points contained in cl _j as Q, and the value of the equation D of the discontinuous surface where it is located is D ₁ , create another data set R (except cl _j and cl ₁ ), and retrieve the M to Q plane The nearest point, the cluster where it is located is defined as cl _i , and the D value of the discontinuous surface equation where it is located is D ₂ , output cl _j , cl _i , D ₁ , D ₂ , D ₁ -D ₂ , and calculate cl _j , cl _i is the distance between two discontinuous surfaces, and the above process is repeated for other discontinuous surfaces until all the discontinuous feature points are retrieved.

Further, the rock mass fracture density is calculated through the area of the linear discontinuity surface and the planar discontinuity surface.

The second aspect of the present invention provides a full-coverage rock mass discontinuity surface extraction and spacing calculation system.

A full-coverage rock mass discontinuity extraction and spacing calculation system, including:

The data acquisition module is configured to: acquire the crack image data exposed in a linear form and the crack image data exposed in a planar form;

The linear discontinuity surface extraction module is configured to: obtain the rock mass crack recognition result map based on the crack image data exposed in a linear form, combined with the semantic segmentation model, and combine the crack pixels in the rock mass crack recognition result map with the original image The corresponding pixels are replaced to obtain the fusion map of the crack recognition results. According to the fusion map of the crack recognition results, the 3D reconstruction of rock mass cracks is carried out. According to the 3D reconstruction results, the 3D coordinates of the crack points are extracted, and the crack feature points are fitted with random convex polygons. Algorithm for fitting, get the fitted linear discontinuity surface;

The planar discontinuous surface extraction module is configured to: use the neural network model to identify the results of cracks based on the crack image data exposed in the planar form, obtain the feature points of the exposed discontinuous surface through the reconstruction algorithm, and perform sequential After calculating the coplanar parameters, fitting the plane, generating the plane and checking the coplanarity, a planar discontinuous surface is obtained;

The discontinuous surface distance calculation module is configured to: obtain the discontinuous surface distance according to the obtained linear discontinuous surface and planar discontinuous surface.

As an optional implementation method, according to the obtained linear discontinuous surface and planar discontinuous surface, the discontinuous surface spacing is obtained, including:

The clustering feature point of the linear discontinuous surface and the planar discontinuous surface is P, the D value of all points is named D _p , and the clustering attribute is defined as D _c ;

Obtain the cluster list according to the ascending order of the D value of P points, randomly select a discontinuous surface and define it as cl ₁ , create a feature point data set R except cl ₁ , search the clusters in R, and find the feature point closest to cl ₁ The discontinuous surface where it is located is cl _j ;

After obtaining cl _j , define all the feature points contained in cl _j as Q, and the value of the equation D of the discontinuous surface where it is located is D ₁ , create another data set R (except cl _j and cl ₁ ), and retrieve the M to Q plane The nearest point, the cluster where it is located is defined as cl _i , and the D value of the discontinuous surface equation where it is located is D ₂ , output cl _j , cl _i , D ₁ , D ₂ , D ₁ -D ₂ , and calculate cl _j , cl _i is the distance between two discontinuous surfaces, repeat the above process for other discontinuous surfaces until all the discontinuous feature points are retrieved

The third aspect of the present invention provides a computer-readable storage medium, on which a program is stored, and when the program is executed by a processor, the full-coverage rock mass discontinuity extraction and distance calculation as described in the first aspect of the present invention are realized. steps in the method.

The fourth aspect of the present invention provides an electronic device, including a memory, a processor, and a program stored on the memory and operable on the processor. Steps in the full coverage rock mass discontinuity extraction and distance calculation method.

Compared with prior art, the beneficial effect of the present invention is:

1. The full-coverage method and system for extracting discontinuous surfaces of rock mass and calculating distances according to the present invention divides the discontinuous surfaces of rock mass into two types of discontinuous surfaces, namely planar exposure and trace exposure. Recognize and reconstruct, extract cracks, convert images into crack feature points, extract discontinuous surfaces through algorithms, and calculate their spacing, and finally realize the characterization of discontinuous distribution, with a high degree of intelligence.

2. The full-coverage discontinuity surface extraction and spacing calculation method and system of the present invention propose a discontinuity surface based on image combined with deep learning technology for the discontinuity surface of the rock mass that is completely exposed. The automatic extraction method can realize the automatic extraction of the discontinuity surface of the rock mass surface, which is beneficial to directly analyze the discontinuity of the rock mass.

3. The full-coverage method and system for extracting discontinuous rock mass discontinuities and calculating distances according to the present invention aims at incompletely exposed rock mass discontinuous faces, combined with deep learning technology to realize rock mass intelligent identification, and proposes a The convex polygon fitting algorithm realizes the fitting of the discontinuous surface of the rock mass, and finally realizes the extraction of the discontinuous surface of the rock mass and the acquisition of three-dimensional parameters.

4. The method and system for extracting and calculating distance between discontinuous surfaces of a full-coverage rock mass according to the present invention calculates the distance between discontinuous surfaces according to the D value of the discontinuous surface equation and the attributes of the discontinuous surfaces, and realizes the automatic distribution of discontinuous surfaces. characterization.

Advantages of additional aspects of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention.

Description of drawings

The accompanying drawings constituting a part of the present invention are used to provide a further understanding of the present invention, and the schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations to the present invention.

Fig. 1 is a schematic flowchart of the intelligent automatic extraction and distance calculation method for discontinuous surfaces of rock mass provided by Embodiment 1 of the present invention.

FIG. 2 is a schematic flow diagram of a system for calculating distances between discontinuous surfaces provided by Embodiment 1 of the present invention.

FIG. 3 is a schematic diagram of a random convex polygon fitting method provided in Embodiment 1 of the present invention.

detailed description

The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

It should be noted that the following detailed description is exemplary and intended to provide further explanation of the present invention. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

It should be noted that the terminology used here is only for describing specific embodiments, and is not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly dictates otherwise, the singular is intended to include the plural, and it should also be understood that when the terms "comprising" and/or "comprising" are used in this specification, they mean There are features, steps, operations, means, components and/or combinations thereof.

In the case of no conflict, the embodiments and the features in the embodiments of the present invention can be combined with each other.

Example 1:

As shown in Figures 1 and 2, Embodiment 1 of the present invention provides a full-coverage rock mass discontinuity extraction and spacing calculation method, including linear discontinuity extraction, planar discontinuity extraction, and discontinuity extraction. Spacing calculation and rock mass discontinuity analysis four processes.

Linear discontinuity extraction includes crack identification, 3D reconstruction of rock mass, random convex polygon fitting, DBSCAN geometric detection and denoising, more specifically, including:

The cracks in the rock mass are intelligently identified through the semantic segmentation model, and the cracks in the crack map are marked as the target by using the labeling tool, and other rock masses are divided into the background, and the rock mass crack image data set is input into the semantic segmentation model at a ratio of 9:1. Training, output rock mass fissure identification result map;

Replace the crack pixels in the rock mass crack recognition result map with the corresponding pixels in the original picture to obtain the fusion map of the crack recognition results, and use the fusion map as the input of the 3D reconstruction algorithm, and then perform feature point extraction, feature point matching, matching optimization, Triangulation, pose estimation, BA optimization, etc. realize sparse modeling. After sparse modeling, depth map estimation and depth map optimization are performed to obtain dense modeling;

Using the different RGB values of the crack feature points in the fusion image, extract the crack feature points, obtain the three-dimensional coordinates of the crack feature points, and use the random convex polygon fitting algorithm to fit the crack feature points.

The random convex polygon fitting algorithm is as follows, including:

倾向为

其中，a，b，c分别是平面在三个坐标轴上的截距。数据集每个点到裂缝面的距离通过截距参数定义为欧氏距离：Assuming that the normal point subset of the crack point is M _i , and the discrete point subset is M ₀ , the crack fitting surface can be defined as

tend to be

Among them, a, b, c are the intercepts of the plane on the three coordinate axes respectively. The distance from each point in the data set to the fracture surface is defined as the Euclidean distance by the intercept parameter:

式中为目标裂缝面，因为由回归生成的各裂缝面候选随机抽取的样本不能直接用M_i验证。因此，本实施提供了一个特定参数进行验证，假设支持度为s，距离阈值为τ，该参数可以表示为：Among them, x _i represents any crack point, crack detection can be simplified as an optimization problem of parameter selection, which can be expressed as

where is the target fracture surface, because the randomly selected samples of each fracture surface candidate generated by regression cannot be directly verified by _Mi. Therefore, this implementation provides a specific parameter for verification. Assuming that the support degree is s and the distance threshold is τ, this parameter can be expressed as:

If the amount of support exceeds a predefined threshold within the maximum iteration step, the iteration termination criterion is satisfied and the crack candidate is determined to be the optimal solution. After each detection, the points of the fitting plane in the corresponding data set are deleted, and then iterated to find the next optimal crack until the entire point data set is small enough. Assuming q is the probability of accurately estimating the fracture parameters from the dataset, then the probability of picking a sample with at least one outlier is 1-q. Therefore, the minimum number of iterations to obtain an accurate fracture surface satisfies the relation:

其中

in

As shown in Fig. 3, the fitted polygon can be obtained, and then the plane equation of the plane can be obtained.

The planar discontinuity extraction system first uses the neural network model to identify the crack surface exposed in the planar form, and further obtains the feature points of the exposed discontinuity surface through the reconstruction algorithm, and then performs coplanar parameter calculation and plane fitting. , Plane generation, coplanarity test, and extraction of planar discontinuities.

First, lock a point P _i on the discontinuous surface, use the Knn (K-nearest neighbors) algorithm to search for nearby feature points according to the search function and Euclidean distance, and name them Q _i , after obtaining k nearby feature points, use PCA (Principal Component Analysis) The algorithm detects the clustering plane, obtains the plane normal vector (a,b,c) and eigenvalues (λ ₁ ,λ ₂ ,λ ₃ ), and defines the deviation parameter η:

Set a maximum allowable maximum deviation parameter η _max , according to experience, when η<η _max , meet the coplanar condition, and further decompose and fit through the SVD singular value (singular value decomposition) algorithm to obtain the clustering discontinuous surface plane equation ax+by+ cz+d=0.

The PCA algorithm can obtain the plane normal vector of each cluster, convert the plane normal vector into a stereo projection, observe the density of the feature points, and obtain the local maximum value of the density, and use the normal vector here as the discontinuous surface Main polar direction, determine the discontinuous surface as the main plane of this clustered discontinuous surface, then the main equation Ax+By+Cz+D=0 of the discontinuous surface can be obtained.

The distance calculation system for discontinuous surfaces is shown in the figure. The clustering feature point of linear discontinuous surfaces and planar discontinuous surfaces is P, and the D value of all points is named D _p , and its clustering attribute is defined as D _c , get the cluster list according to the ascending order of D value of P points, randomly select a discontinuous surface and define it as cl ₁ , create a feature point data set R except cl ₁ , search the clusters in R, and find the feature closest to cl ₁ The discontinuous surface where the point is located is cl _j , after obtaining cl _j , define all the feature points contained in cl _j as Q, and the value of the equation D of the discontinuous surface where it is located is D ₁ , create another data set R (except cl _j and cl ₁ ), retrieve the nearest point from M to Q plane, the cluster where it is located is defined as cl _i , and the value of the equation D of the discontinuous surface where it is located is D ₂ , output cl _j , cl _i , D ₁ , D ₂ , D ₁ -D ₂ , that is, calculate the distance between two discontinuous surfaces cl _j and cl _i , and repeat the above process for other discontinuous surfaces until all discontinuous feature points are retrieved.

The discontinuous surface spacing obtained by the discontinuous surface spacing calculation system and the discontinuous surface area obtained by the discontinuous surface extraction system are used to calculate the rock mass fracture density P ₃₂ (m/m ² ).

Applying this method to realize the intelligent automatic extraction and spacing calculation of discontinuities in rock mass, the operation method includes the following steps:

1) First, the linear discontinuity extraction system uses a camera or mobile phone to take images of rock mass fissures, establishes a data set of rock mass fissures images, uses labeling tools to mark the image fissures, and expands the data set through preprocessing such as segmentation, rotation, and mirroring , according to the ratio of 9:1, input the semantic segmentation model to intelligently identify rock mass fissures, and output the identification result map;

2) Matrix replacement between the rock mass crack recognition result image and the corresponding crack pixel points in the original image to realize the fusion of crack recognition results, and sequentially perform feature point extraction, feature point matching, matching optimization, triangulation, pose estimation, BA optimization and depth Figure estimation, complete the three-dimensional reconstruction of rock mass, realize the conversion of all feature points from two-dimensional to three-dimensional, and obtain their three-dimensional coordinates;

3) The linear discontinuity surface extraction system extracts the crack feature points according to the difference in the RGB values in the fusion image of the crack recognition results, and obtains the three-dimensional coordinates of the crack feature points, and uses a random convex polygon fitting algorithm to fit the discontinuous surface. Further obtain the optimal equation of the discontinuity surface;

4) The planar discontinuity extraction system repeats steps 1) and 2), identifying the planar discontinuity and realizing the three-dimensional conversion of the feature points of the planar discontinuity;

5) The surface discontinuity extraction system calculates point curvature calculations and point normal vectors, extracts clustered discontinuity surfaces according to SVD singular value decomposition, and further calculates density distribution. Continuous surface, obtain the plane equation and properties of the discontinuous surface;

6) Input the discontinuous surface equation and attributes into the discontinuous surface distance calculation system, divide the discontinuous surface according to the D value, and iteratively calculate the distance between two discontinuous surfaces according to the method, until all the discontinuous feature points are retrieved.

7) The discontinuous surface analysis system calculates the unit volume density P ₃₂ (m/m ² ).

Example 2:

Embodiment 2 of the present invention provides a full-coverage rock mass discontinuity extraction and spacing calculation system, including:

The data acquisition module is configured to: acquire the crack image data exposed in a linear form and the crack image data exposed in a planar form;

The linear discontinuity surface extraction module is configured to: obtain the rock mass crack recognition result map based on the crack image data exposed in a linear form, combined with the semantic segmentation model, and combine the crack pixels in the rock mass crack recognition result map with the original image The corresponding pixels are replaced to obtain the fusion map of the crack recognition results. According to the fusion map of the crack recognition results, the 3D reconstruction of rock mass cracks is carried out. According to the 3D reconstruction results, the 3D coordinates of the crack points are extracted, and the crack feature points are fitted with random convex polygons. Algorithm for fitting, get the fitted linear discontinuity surface;

The planar discontinuous surface extraction module is configured to: use the neural network model to identify the results of cracks based on the crack image data exposed in the planar form, obtain the feature points of the exposed discontinuous surface through the reconstruction algorithm, and perform sequential After calculating the coplanar parameters, fitting the plane, generating the plane and checking the coplanarity, a planar discontinuous surface is obtained;

The discontinuous surface distance calculation module is configured to: obtain the discontinuous surface distance according to the obtained linear discontinuous surface and planar discontinuous surface.

The working method of the system is the same as that of the full-coverage discontinuity surface extraction and spacing calculation method provided in Embodiment 1, and will not be repeated here.

Example 3:

Embodiment 3 of the present invention provides a computer-readable storage medium, on which a program is stored, and when the program is executed by a processor, the full-coverage rock mass discontinuity extraction and distance calculation as described in Embodiment 1 of the present invention are realized. steps in the method.

Example 4:

Embodiment 4 of the present invention provides an electronic device, including a memory, a processor, and a program stored in the memory and operable on the processor. When the processor executes the program, the implementation as described in Embodiment 1 of the present invention Steps in the full coverage rock mass discontinuity extraction and distance calculation method.

Those skilled in the art should understand that the embodiments of the present invention may be provided as methods, systems, or computer program products. Accordingly, the present invention can take the form of a hardware embodiment, a software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) having computer-usable program code embodied therein.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each procedure and/or block in the flowchart and/or block diagram, and a combination of procedures and/or blocks in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions may be provided to a general purpose computer, special purpose computer, embedded processor, or processor of other programmable data processing equipment to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing equipment produce a An apparatus for realizing the functions specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to operate in a specific manner, such that the instructions stored in the computer-readable memory produce an article of manufacture comprising instruction means, the instructions The device realizes the function specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions can also be loaded onto a computer or other programmable data processing device, causing a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby The instructions provide steps for implementing the functions specified in the flow chart or blocks of the flowchart and/or the block or blocks of the block diagrams.

Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be implemented through computer programs to instruct related hardware, and the programs can be stored in a computer-readable storage medium. During execution, it may include the processes of the embodiments of the above-mentioned methods. Wherein, the storage medium may be a magnetic disk, an optical disk, a read-only memory (Read-Only Memory, ROM) or a random access memory (Random Access Memory, RAM) and the like.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and changes. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims (10)

1. A fully covered rock mass discontinuity extraction and spacing calculation method, characterized in that: Including the following process: Obtain image data of cracks exposed in a linear form and image data of cracks exposed in a planar form; According to the crack image data exposed in a linear form, combined with the semantic segmentation model, the rock mass crack recognition result map is obtained, and the crack pixels in the rock mass crack recognition result map are replaced with the pixels corresponding to the original image to obtain the crack recognition result fusion map According to the fusion map of the crack identification results, the 3D reconstruction of the rock mass cracks is carried out. According to the 3D reconstruction results, the 3D coordinates of the crack points are extracted, and the feature points of the cracks are fitted by a random convex polygon fitting algorithm to obtain the fitted linear irregularities. continuous surface; According to the crack image data exposed in planar form, the neural network model is used to identify the cracks, and the feature points of the exposed discontinuous surface are obtained through the reconstruction algorithm, and the coplanar parameter calculation, plane fitting, plane generation and coplanarity After the surface test, a planar discontinuous surface is obtained; From the obtained linear discontinuous surfaces and planar discontinuous surfaces, the distance between discontinuous surfaces is obtained. 2. The method for extracting discontinuous surfaces of a fully covered rock mass and calculating distances as claimed in claim 1, characterized in that: The fusion map of the crack recognition results is used as the input of the 3D reconstruction algorithm, and sparse modeling is performed after feature point extraction, feature point matching, matching optimization, triangulation, pose estimation and BA optimization, and depth map estimation and optimization are obtained. 3D reconstruction results from dense modeling. 3. The method for extracting discontinuous surfaces of a fully covered rock mass and calculating distances as claimed in claim 1, characterized in that: Using the different RGB values of the crack feature points in the fusion image, extract the crack feature points and obtain the three-dimensional coordinates of the crack feature points. 4. The full-coverage discontinuity surface extraction and spacing calculation method of claim 1, characterized in that: Lock a point on the discontinuous surface, use the KNN algorithm to search for nearby feature points according to the search function and Euclidean distance, and name them Q _i , after obtaining k nearby feature points, use the PCA algorithm to detect the clustering plane, and obtain the plane normal vector (a ,b,c) and eigenvalues (λ ₁ ,λ ₂ ,λ ₃ ), define the deviation parameter

When η is less than the maximum allowable maximum deviation parameter, the coplanar condition is met, and the plane equation of the clustered discontinuous surface is obtained by SVD singular value algorithm decomposition and fitting; The PCA algorithm obtains the plane normal vector of each cluster, converts the plane normal vector into a stereo projection, observes the density of feature points, and obtains the local maximum value of the density, and uses the normal vector at the local maximum value of the density as a discontinuous surface The main pole direction of the discontinuity surface is determined as the main plane of the cluster discontinuity surface, and then the master equation of the discontinuity surface is obtained. 5. The method for extracting discontinuous surfaces of a fully-covered rock mass and calculating distances as claimed in claim 1, characterized in that: According to the obtained linear discontinuity surface and planar discontinuity surface, the distance between discontinuity surfaces is obtained, including: The clustering feature point of the linear discontinuous surface and the planar discontinuous surface is P, the D value of all points is named D _p , and the clustering attribute is defined as D _c ; Obtain the cluster list according to the ascending order of the D value of P points, randomly select a discontinuous surface and define it as cl ₁ , create a feature point data set R except cl ₁ , search the clusters in R, and find the feature point closest to cl ₁ The discontinuous surface where it is located is cl _j ; After obtaining cl _j , define all the feature points contained in cl _j as Q, and the value of the equation D of the discontinuous surface where it is located is D ₁ , create another data set R (except cl _j and cl ₁ ), and retrieve the M to Q plane The nearest point, the cluster where it is located is defined as cl _i , and the D value of the discontinuous surface equation where it is located is D ₂ , output cl _j , cl _i , D ₁ , D ₂ , D ₁ -D ₂ , and calculate cl _j , cl _i is the distance between two discontinuous surfaces, and the above process is repeated for other discontinuous surfaces until all the discontinuous feature points are retrieved. 6. The full-coverage discontinuity surface extraction and spacing calculation method of claim 5, characterized in that: The fracture density of the rock mass is calculated by the area of the linear discontinuity surface and the planar discontinuity surface. 7. A full-coverage rock mass discontinuity extraction and spacing calculation system, characterized in that: include: The data acquisition module is configured to: acquire the crack image data exposed in a linear form and the crack image data exposed in a planar form; The linear discontinuity surface extraction module is configured to: obtain the rock mass crack recognition result map based on the crack image data exposed in a linear form, combined with the semantic segmentation model, and combine the crack pixels in the rock mass crack recognition result map with the original image The corresponding pixels are replaced to obtain the fusion map of the crack recognition results. According to the fusion map of the crack recognition results, the 3D reconstruction of rock mass cracks is carried out. According to the 3D reconstruction results, the 3D coordinates of the crack points are extracted, and the crack feature points are fitted with random convex polygons. Algorithm for fitting, get the fitted linear discontinuity surface; The planar discontinuous surface extraction module is configured to: use the neural network model to identify the results of cracks based on the crack image data exposed in the planar form, obtain the feature points of the exposed discontinuous surface through the reconstruction algorithm, and perform sequential After calculating the coplanar parameters, fitting the plane, generating the plane and checking the coplanarity, a planar discontinuous surface is obtained; The discontinuous surface distance calculation module is configured to: obtain the discontinuous surface distance according to the obtained linear discontinuous surface and planar discontinuous surface. 8. The full-coverage rock mass discontinuity extraction and spacing calculation system as claimed in claim 7, characterized in that: According to the obtained linear discontinuity surface and planar discontinuity surface, the distance between discontinuity surfaces is obtained, including: The clustering feature point of the linear discontinuous surface and the planar discontinuous surface is P, the D value of all points is named D _p , and the clustering attribute is defined as D _c ; Obtain the cluster list according to the ascending order of the D value of P points, randomly select a discontinuous surface and define it as cl ₁ , create a feature point data set R except cl ₁ , search the clusters in R, and find the feature point closest to cl ₁ The discontinuous surface where it is located is cl _j ; After obtaining cl _j , define all the feature points contained in cl _j as Q, and the value of the equation D of the discontinuous surface where it is located is D ₁ , create another data set R (except cl _j and cl ₁ ), and retrieve the M to Q plane The nearest point, the cluster where it is located is defined as cl _i , and the D value of the discontinuous surface equation where it is located is D ₂ , output cl _j , cl _i , D ₁ , D ₂ , D ₁ -D ₂ , and calculate cl _j , cl _i is the distance between two discontinuous surfaces, and the above process is repeated for other discontinuous surfaces until all the discontinuous feature points are retrieved. 9. A computer-readable storage medium, on which a program is stored, characterized in that, when the program is executed by a processor, the full-coverage rock mass discontinuity extraction and Steps in the spacing calculation method. 10. An electronic device comprising a memory, a processor, and a program stored on the memory and operable on the processor, wherein the processor implements any one of claims 1-6 when executing the program Steps in the method for extracting discontinuous surfaces of a fully covered rock mass and calculating distances.

Images