Layered rock mass preparation method and device based on digital image

Abstract

The invention discloses a layered rock mass preparation method and device based on a digital image, and the method comprises the following specific steps: S1, carrying out the intelligent recognition and parameter extraction of the stratified rock mass bedding of a target layer rock sample, obtaining the bedding structure feature information of the sample, carrying out the high-precision digital three-dimensional reconstruction of the bedding structure, and obtaining the high-precision digital three-dimensional reconstruction of the stratified rock mass; a corresponding digital reconstruction model is obtained; and S2, according to the digital reconstruction model and bedding structure feature information of the sample, preparing the layered rock-like test piece based on a 3D printing technology. According to the invention, the preparation process of the layered rock mass test piece is simplified, the sample printing efficiency is improved, the sample difference is reduced, the consistency and surface flatness of the test piece are greatly improved, the discreteness among the samples is reduced, and the rock mass characteristics under the real geological condition are better simulated; meanwhile, the period of stratified rock mass research and engineering application is effectively shortened, and the reliability and practicability of a simulation experiment are improved.

Description

Technical Field

[0001] The present invention relates to the field of digital geotechnical engineering technology, and in particular to a method and device for preparing layered rock mass based on intelligent recognition and parameter extraction of digital images and sand mold 3D printing technology. Background Art

[0002] The layered rock mass formed in nature has uneven thickness distribution of different rock layers and great differences in the degree of cementation between layers due to the influence of long-term complex geological structures. As a result, the test results of layered rock specimens sampled on site are often affected by multiple variables, which in turn leads to large discreteness and non-repeatability of the test results. The use of 3D printing technology can save materials, eliminate the need to remove scraps, improve material utilization, and achieve higher precision. Compared with traditional production methods, 3D printed samples have less discreteness, a smoother surface, and are closer to the actual engineering rock state.

[0003] The existence of bedding dip makes the mechanical properties and failure mode of rock mass closely related to the force direction. In addition to the directional characteristics caused by the bedding dip, the diversity of bedding roughness, spacing, weak interlayers and mechanical properties also increase the complexity of layered rock mass, and even control the deformation coordination and instability characteristics of the rock mass to a certain extent. Therefore, it is very necessary to quantitatively analyze the influence of bedding characteristics on the mechanical behavior of layered rock mass and reveal its failure mechanism.

[0004] In addition, due to the three-dimensional spatial distribution characteristics of bedding and the complexity of the existing stress environment in deep underground projects, the macro- and micro-destruction mechanisms of deep layered rock masses have not been fully revealed. In particular, there is a lack of new methods of layered rock visualization and digital mechanical testing based on deep learning and 3D printing reconstruction to study the anisotropic mechanical properties and micro-destruction mechanisms of deep layered rock masses under complex stress states.

[0005] To this end, the present application specifically proposes a layered rock preparation method and device based on digital images to solve the above technical problems. Summary of the invention

[0006] The main purpose of the present invention is to provide a method and device for preparing layered rock mass based on digital images, so as to solve the technical problems raised in the background technology.

[0007] The present invention adopts the following technical solutions to solve the above technical problems:

[0008] A method for preparing layered rock mass based on digital images comprises the following specific steps:

[0009] S1. Intelligently identify and extract parameters of the layered rock mass of the target layer rock sample, obtain the layered structure characteristic information of the sample, and perform high-precision digital three-dimensional reconstruction of the layered structure to obtain the corresponding digital reconstruction model;

[0010] S2. According to the digital reconstruction model and bedding structure characteristic information of the sample, layered rock specimens are prepared based on 3D printing technology.

[0011] Preferably, the specific operation steps of performing intelligent identification and parameter extraction in step S1 include:

[0012] S11. Establishing a layered structure image database for model training through scanning results of layered rock specimens, the database comprising training set images and test set images;

[0013] S12. Use the improved MD-LinkNet model to accurately identify and segment the bedding structure in the database digital image;

[0014] S13. extracting the bedding skeleton in the bedding recognition result map, and performing vector processing on the skeleton;

[0015] S14. The key features of bedding parameters are extracted from the vectorized skeleton, including bedding inclination, length, fractal dimension and spacing, to provide a basis for the preparation of layered rock specimens.

[0016] Preferably, in the step S11, the database is further added with joint structure images as a supplementary training set to enhance the generalization ability of the MD-LinkNet model.

[0017] Preferably, the MD-LinkNet model in step S12 is based on the traditional D-LinkNet network model framework to establish a fusion of multi-scale features by cascaded dilated convolution operations and residual multi-core pooling operations, so as to realize the effective extraction of multi-scale feature information of the layered structure.

[0018] Preferably, the residual multi-core pooling operation is performed in the encoder to adapt to the changes in the scale of the bedding structure picture information and extract rich bedding structure feature information.

[0019] Preferably, the MD-LinkNet model uses DUpsample bilinear interpolation for upsampling to optimize the low-resolution image of the bedding structure after convolution pooling, and fuses the global and local features of the bedding structure through pixel neighborhood information to improve the accuracy of bedding structure recognition.

[0020] Preferably, the specific operation steps for preparing the rock-like specimen in step S2 include:

[0021] S21. Sand mold 3D technology is used to prepare layered rock specimens, and the proportion of rock-like materials is determined through proportion tests and triaxial compression, uniaxial compression and Brazilian splitting tests;

[0022] S22. Use the control variable method to generate three-dimensional layered rock data information with different bedding roughness and spacing, and import the model data into Solidworks to convert it into a data format recognized by the 3D printer

[0023] S23. Reconstructing cubic layered rock blocks with different bedding roughness and spacing using a 3D printing system;

[0024] S24. The obtained cubic layered rock blocks are used to prepare layered rock specimens under different loading conditions.

[0025] Preferably, the sand mold 3D technology in step S21 is a 3D printing operation performed by inkjet powder bonding technology, the printing material is quartz sand powder, ceramsite sand powder or a combination of the two, and the binder is furan resin.

[0026] A digital image-based layered rock mass preparation device, applied to any of the above methods, comprising:

[0027] The intelligent identification and parameter extraction module is used to intelligently identify and extract parameters of the layered rock mass of the target layer rock sample to obtain the layered structure characteristic information of the sample;

[0028] A digital 3D reconstruction module is used to perform high-precision digital 3D reconstruction according to the layered structure to obtain a corresponding digital reconstruction model;

[0029] The printing module is used to prepare layered rock specimens according to the digital reconstruction model and layer structure characteristic information of the sample.

[0030] Preferably, the intelligent identification and parameter extraction module is provided with:

[0031] The layered structure image database contains scanning results of multiple layered rock specimens as training set images and test set images;

[0032] The MD-LinkNet model is used to accurately identify and segment the bedding structure of image data in the bedding structure image database, extract the bedding skeleton in the bedding recognition result map, vectorize the skeleton, and extract the key features of the bedding parameters based on the vectorized skeleton.

[0033] It can be seen from the above technical solution that the present invention provides a method and device for preparing layered rock mass based on digital images. Compared with the prior art, the present invention has the following advantages:

[0034] 1. The present invention adopts an improved MD-LinkNet model in the bedding identification stage, integrates cascaded void convolution and residual multi-core pooling technology, can more finely capture and analyze the multi-scale characteristics of layered rock mass, significantly improve the accuracy and efficiency of bedding structure identification, and thus provide a more accurate bedding structure model for subsequent 3D printing.

[0035] 2. The present invention optimizes the quality of low-resolution images in the image post-processing stage by introducing the DUpsample bilinear interpolation upsampling technology, ensures the effective fusion of the global and local features of the bedding structure, improves the detail restoration of the image, and makes the 3D printed rock specimen structure closer to the actual situation in nature.

[0036] 3. The present invention constructs a set of characteristic points of the bedding structure in the process of intelligent identification, uses intersection recognition and skeleton vectorization calculation to accurately extract key parameters of the bedding, realizes quantitative analysis of the bedding structure recognition image, and provides solid data support for rock mechanics analysis.

[0037] 4. The present invention combines precise digital models with sand mold 3D printing technology, which not only simplifies the preparation process of layered rock specimens, improves the sample printing efficiency while reducing sample differences, but also greatly improves the consistency and surface flatness of the specimens, reduces the discreteness between samples, and better simulates the rock characteristics under real geological conditions.

[0038] 5. The present invention closely combines intelligent image analysis with high-precision 3D printing technology, effectively shortening the cycle of layered rock research and engineering application, improving the reliability and practicality of simulation experiments, and facilitating further in-depth exploration of the mechanical behavior, anisotropic characteristics and microscopic failure mechanism of deep layered rock masses.

[0039] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present invention, nor is it intended to limit the scope of the present invention. Other features of the present invention will become easy to understand through the following description. Of course, it is not necessary to achieve all of the advantages described above simultaneously for any product implementing the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The drawings constituting a part of the present application are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0041] Figure 1 It is a schematic diagram of the overall process of the method of the present invention;

[0042] Figure 2 Schematic diagram of the bedding identification process of the method of the present invention;

[0043] Figure 3 It is a statistical schematic diagram of bedding spacing parameters of the method of the present invention;

[0044] Figure 4 It is a statistical schematic diagram of bedding roughness parameters of the method of the present invention;

[0045] Figure 5 It is a three-dimensional schematic diagram of the structure of layered rock blocks with different bedding spacings according to the present invention;

[0046] Figure 6 It is a schematic diagram of the structure of layered rock blocks with different bedding roughness according to the present invention;

[0047] Figure 7 The schematic diagram of the overall process of preparing layered rock specimens by the method of the present invention;

[0048] Figure 8 Schematic diagram of the preparation process of cube specimens with different inclination angles of the present invention;

[0049] Fig. 9 Schematic diagram of the preparation process of cylindrical specimens with different inclination angles of the present invention;

[0050] Fig.10 Schematic diagram of the preparation process of disc specimens with different bedding inclination angles of the present invention;

[0051] Fig.11 Schematic diagram of the preparation process of SCB specimens with different inclination angles of the present invention;

[0052] Fig.12 Schematic diagram of the module structure of the device of the present invention. DETAILED DESCRIPTION

[0053] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. In the absence of conflict, the embodiments in this application and the features in the embodiments can be combined with each other. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0054] In the embodiment, see Figures 1 to 12 .

[0055] like Figure 1 The method for preparing a layered rock mass based on digital images proposed in the embodiment of the present invention obtains the three-dimensional spatial distribution characteristics and digital information of the layered rock mass through the intelligent recognition and characterization algorithm of the bedding structure, and realizes the establishment of a three-dimensional physical model of the layered rock mass by using 3D printing technology, and includes the following specific steps:

[0056] S1. Intelligently identify and extract parameters of the layered rock mass of the target layer rock sample, obtain the layered structure characteristic information of the sample, and perform high-precision digital three-dimensional reconstruction of the layered structure to obtain the corresponding digital reconstruction model;

[0057] The specific steps for intelligent identification and parameter extraction include:

[0058] S11. Establishing a layered structure image database for model training through scanning results of layered rock specimens, the database comprising training set images and test set images;

[0059] S12. Use the improved MD-LinkNet model to accurately identify and segment the bedding structure in the database digital image;

[0060] In addition, it can be further supplemented that joint structure images can be added to the referenced layer structure image database as a supplementary training set to enhance the generalization ability of the MD-LinkNet model;

[0061] Among them Figure 2 As shown in the figure, the MD-LinkNet model is based on the traditional D-LinkNet network model framework, which uses the cascaded hole convolution fusion operation and the remaining multi-core pooling operation to fuse multi-scale features, so as to effectively extract multi-scale feature information of the layered structure.

[0062] At this time, by adopting the improved MD-LinkNet model in the bedding identification stage and integrating cascaded void convolution and residual multi-core pooling technology, the multi-scale characteristics of the layered rock mass can be captured and analyzed more finely, and the accuracy and efficiency of bedding structure identification can be significantly improved, thereby providing a more accurate bedding structure model for subsequent 3D printing;

[0063] Furthermore, the MD-LinkNet model uses DUpsample bilinear interpolation for upsampling to optimize the low-resolution image of the bedding structure after convolution pooling, and fuses the global and local features of the bedding structure through pixel neighborhood information to improve the recognition accuracy of the bedding structure.

[0064] Therefore, in summary, the improved MD-LinkNet model adds residual multicore pooling (RMP) in the encoder to adapt to the changes in the scale of the bedding structure image information and extract rich bedding structure feature information; the low-resolution image after the pooling operation is upsampled by DUpsample (Data-dependent Up-sample) to enhance the combination of bedding structure pixels and domain pixels, and fuse the global features of the bedding structure with the local features, thereby improving the calculation rate and sensitivity of the model as well as the segmentation accuracy;

[0065] At this time, by introducing the DUpsample bilinear interpolation upsampling technology, the quality of low-resolution images is optimized in the image post-processing stage, ensuring the effective fusion of the global and local features of the bedding structure, improving the image detail restoration, and making the 3D printed rock specimen structure closer to the actual situation in nature;

[0066] S13. extracting the bedding skeleton in the bedding recognition result map, and performing vector processing on the skeleton;

[0067] Here, the bedding structure feature point set is constructed through the bedding structure intelligent recognition structure map, and the intersection recognition and bedding skeleton vectorization calculation methods are proposed. The bedding parameters such as bedding inclination, length, fractal dimension and spacing are accurately extracted, and the quantitative analysis of the bedding structure recognition image is realized, providing solid data support for rock mass mechanics analysis.

[0068] S14. Extract key features of bedding parameters from the vectorized skeleton. The bedding parameters here can be referenced Figure 3 and Figure 4 , including bedding dip, length, fractal dimension, spacing and roughness, which are used to provide a basis for the preparation of layered rock specimens;

[0069] S2. According to the digital reconstruction model and bedding structure characteristic information of the sample, layered rock specimens are prepared based on 3D printing technology;

[0070] The specific steps for preparing the rock specimens include:

[0071] S21. Use sand mold 3D technology to prepare layered rock specimens. After proportion test and triaxial compression, uniaxial compression and Brazilian splitting test, determine the uniaxial compressive strength, elastic modulus, tensile strength and failure mode of layered rock with different bedding plane inclination angles. Compare the test results with the test results of layered rock sampled on site to determine the proportion of rock-like materials.

[0072] It should be noted that the equipment used in the sand mold 3D technology here is the industrial-grade 3D printing equipment VX2000, which can economically produce sand molds and sand cores for casting. It has a printing volume of 2000L and can put sand molds of different sizes and structures in the same molding box for fast and flexible printing. The manufacturing method without mold opening can produce small and medium batches of parts faster. The technical process at this time is to perform 3D printing operations through inkjet powder bonding technology. The printing material is quartz sand powder, ceramsite sand powder or a combination of the two, and the binder is furan resin;

[0073] S22. Using the control variable method, three-dimensional layered rock mass data information with different bedding roughness and spacing is generated respectively, and the model data is imported into Solidworks to convert it into a data format recognized by the 3D printer, and then imported into a virtual molding cylinder of the computer, which corresponds to the actual molding box volume of the VX2000, and then the layered rock blocks with different bedding roughness and spacing are reconstructed using the 3D printing system;

[0074] S23. Figure 5 and Figure 6 As shown, a 3D printing system is used to reconstruct a cubic layered rock-like block with different bedding roughness and spacing;

[0075] The 3D printing process at this point can be divided into the following steps:

[0076] (1) Push the VX2000 molding box into the printing device;

[0077] (2) Scraper powdering, the printing material is quartz sand powder, ceramsite sand powder or a combination of the two;

[0078] (3) The high-definition print head selectively sprays a binder, which is a furan resin;

[0079] (4) After each layer is printed, the forming platform is lowered by one layer thickness, and the above steps are repeated until the printing is nearly completed. During the printing process, the printing interval between layers is about 20 seconds;

[0080] (5) Unprinted powder is removed in the shell breaking station, and the printed layered rock blocks can be directly used for specimen preparation, saving a lot of rework time;

[0081] S24. Figure 7 As shown, the obtained cubic layered rock blocks are cut, cored, polished, etc. to prepare layered rock specimens for use under different loading conditions;

[0082] refer to Figure 8Here, the cubic layered rock specimen can be ground flat by a grinder to make a layered rock specimen (length × width × height = 50 mm × 50 mm × 100 mm) that meets the requirements of the true triaxial test.

[0083] refer to Fig. 9 Alternatively, a core drill with a diameter of 50 mm may be used to drill and core at a certain angle to the bedding plane to obtain a corresponding cylindrical specimen. A handheld cutting machine may be used to cut the long cylinder along the length of the cylinder. The cut cylindrical specimen may then be polished to produce a cylindrical specimen size (50 mm in diameter and 100 mm in height) that meets the requirements of uniaxial compression test, triaxial compression test, triaxial unloading test, and triaxial compression test with constant amplitude cyclic loading and unloading.

[0084] refer to Fig.10 , similarly, a coring machine with a diameter of 50 mm was used to drill and coring along the direction of the bedding plane, and then the obtained long cylindrical specimens were cut along the length direction of the cylinder using a handheld cutting machine, and then the cut disc specimens were polished to produce disc specimens with different bedding plane inclination angles that meet the requirements of the Brazilian splitting test (diameter 50 mm, thickness 25 mm);

[0085] refer to Fig.11 In addition, similarly, the prepared cubic layered rock specimens were first cored along the bedding plane using a coring drill (with a diameter of 50 mm) to form a long cylinder; the long cylinder was then cut using a handheld cutter to prepare a Brazilian disc, and then the prepared Brazilian disc was polished using a polisher to reduce the unevenness of the disc surface, and then each prepared Brazilian disc specimen was cut into two semicircular specimens using a wire cutting machine according to the plane angle of the cushion layer, and then the notch was cut using the wire cutting machine, and finally an SCB rock specimen with a bedding plane that met the requirements of the three-point bending test was obtained.

[0086] In summary, the present invention combines precise digital models with sand mold 3D printing technology, which not only simplifies the preparation process of layered rock specimens, improves the printing efficiency of specimens and reduces the differences between specimens, but also greatly improves the consistency and surface flatness of specimens, reduces the discreteness between specimens, and better simulates the rock characteristics under real geological conditions. It also closely combines intelligent image analysis with high-precision 3D printing technology, effectively shortens the cycle of layered rock research and engineering application, improves the reliability and practicality of simulation experiments, and facilitates further in-depth exploration of the mechanical behavior, anisotropic characteristics and microscopic failure mechanism of deep layered rock masses. It provides a powerful tool and platform for scientific researchers and engineers to deeply explore the complex mechanical response, anisotropic properties and subtle fracture laws of deep layered rock masses, which is of great significance for guiding underground engineering construction and improving safety.

[0087] On the other hand, Fig.12 The present invention also discloses a layered rock preparation device based on digital images, which is applied to the above method and includes:

[0088] The intelligent identification and parameter extraction module is used to intelligently identify and extract parameters of the layered rock mass of the target layer rock sample to obtain the layered structure characteristic information of the sample;

[0089] A digital 3D reconstruction module is used to perform high-precision digital 3D reconstruction according to the layered structure to obtain a corresponding digital reconstruction model;

[0090] The printing module is used to prepare layered rock specimens according to the digital reconstruction model and layer structure characteristic information of the sample.

[0091] Furthermore, the intelligent recognition and parameter extraction module is provided with:

[0092] The bedding structure image database contains multiple layered rock specimen scanning results as training set images and test set images. Here, a bedding structure image database is set up in the intelligent recognition and parameter extraction module to collect a variety of layered rock scanning images as training and test data sets, which can significantly enhance the breadth and depth of model learning, ensure the accuracy and generalization ability of bedding recognition, and ultimately achieve efficient recognition and analysis of various bedding structures;

[0093] The MD-LinkNet model mainly includes a cascaded dilated convolution module, a residual multi-core pooling module and a DUpsample upsampling module, which are used to accurately identify and segment the bedding structure of image data in the bedding structure image database, extract the bedding skeleton in the bedding recognition result map, vectorize the skeleton, and extract the key features of the bedding parameters based on the vectorized skeleton.

[0094] In another embodiment provided in the present application, a computer program product including instructions is also provided, which, when executed on a computer, enables the computer to execute any of the digital image-based layered rock preparation methods in the above-mentioned embodiments.

[0095] It is understandable that the system provided by the embodiment of the present invention corresponds to the method provided by the embodiment of the present invention, and the explanation, examples and beneficial effects of the relevant contents can refer to the corresponding parts in the above method.

[0096] The embodiment of the present application also provides an electronic device, including a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus.

[0097] Memory, used to store computer programs;

[0098] The processor is used to implement the above-mentioned layered rock mass preparation method based on digital images when executing the program stored in the memory.

[0099] The communication bus mentioned in the above electronic device can be a peripheral component interconnect standard bus or an extended industrial standard architecture bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc.

[0100] The communication interface is used for communication between the above electronic device and other devices.

[0101] The memory may include a random access memory, or may include a non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0102] The above-mentioned processor can be a general-purpose processor, including a central processing unit, a network processor, etc.; it can also be a digital signal processor, an application-specific integrated circuit, a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component.

[0103] It should also be noted that electronic devices also include terminal devices, which can also be called terminals, user equipment, mobile platforms, mobile terminals, etc. Terminal devices can be mobile phones, smart TVs, wearable devices, tablet computers, computers with wireless transceiver functions, virtual reality terminal devices, augmented reality terminal devices, wireless terminals in industrial control, wireless terminals in unmanned driving, wireless terminals in remote surgery, wireless terminals in smart grids, wireless terminals in transportation safety, wireless terminals in smart cities, wireless terminals in smart homes, etc. The embodiments of this application do not limit the specific technology and specific device form used by the terminal devices.

[0104] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) mode. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server, data center, etc. that contains one or more available media integrated. The available medium can be a magnetic medium, (e.g., a floppy disk, a hard disk, a tape), an optical medium, or a semiconductor medium, etc.

[0105] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

[0106] In addition, the meaning of "and / or" appearing in the full text includes three parallel solutions. Taking "A and / or B" as an example, it includes solution A, solution B, or solutions that satisfy both A and B. In addition, in the embodiments of the present invention, "multiple" refers to more than two. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in this field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the protection scope required by the present invention.

Claims

1. A method for preparing layered rock mass based on digital images, characterized in that: The specific steps include: S1. Intelligently identify and extract parameters of the layered rock mass of the target layer rock sample, obtain the layered structure characteristic information of the sample, and perform high-precision digital three-dimensional reconstruction of the layered structure to obtain the corresponding digital reconstruction model; S2. According to the digital reconstruction model and bedding structure characteristic information of the sample, layered rock specimens are prepared based on 3D printing technology.

2. The method for preparing layered rock mass based on digital images according to claim 1, characterized in that: The specific operation steps for performing intelligent identification and parameter extraction in step S1 include: S11. Establishing a layered structure image database for model training through scanning results of layered rock specimens, the database comprising training set images and test set images; S12. Use the improved MD-LinkNet model to accurately identify and segment the bedding structure in the database digital image; S13. extracting the bedding skeleton in the bedding recognition result map, and performing vector processing on the skeleton; S14. Extract key features of bedding parameters from the vectorized skeleton, including bedding inclination, length, fractal dimension and spacing, to provide a basis for the preparation of layered rock specimens.

3. The method for preparing layered rock mass based on digital images according to claim 2, characterized in that: In the step S11, the database is also added with joint structure images as a supplementary training set to enhance the generalization ability of the MD-LinkNet model.

4. The method for preparing layered rock mass based on digital images according to claim 2, characterized in that: The MD-LinkNet model in step S12 is based on the traditional D-LinkNet network model framework, which enables the cascaded hole convolution fusion operation and the remaining multi-core pooling operation to fuse multi-scale features, so as to effectively extract multi-scale feature information of the layered structure.

5. The method for preparing layered rock mass based on digital images according to claim 4, characterized in that: The residual multi-core pooling operation is performed in the encoder to adapt to the changes in the scale of the layered structure picture information and extract rich layered structure feature information.

6. The method for preparing layered rock mass based on digital images according to claim 4, characterized in that: The MD-LinkNet model uses DUpsample bilinear interpolation for upsampling to optimize the low-resolution image of the bedding structure after convolution pooling, and fuses the global and local features of the bedding structure through pixel neighborhood information to improve the accuracy of bedding structure recognition.

7. The method for preparing layered rock mass based on digital images according to claim 1, characterized in that: The specific operation steps for preparing the rock-like specimen in step S2 include: S21. Sand mold 3D technology is used to prepare layered rock specimens, and the proportion of rock-like materials is determined through proportion tests and triaxial compression, uniaxial compression and Brazilian splitting tests; S22. Using the control variable method, three-dimensional layered rock mass data information with different bedding roughness and spacing is generated, and the model data is imported into Solidworks and converted into a data format recognized by the 3D printer; S23. Reconstructing cubic layered rock blocks with different bedding roughness and spacing using a 3D printing system; S24. The obtained cubic layered rock blocks are used to prepare layered rock specimens under different loading conditions.

8. The method for preparing layered rock mass based on digital images according to claim 7, characterized in that: The sand mold 3D technology in the step S21 is a 3D printing operation performed by inkjet powder bonding technology, the printing material is quartz sand powder, ceramsite sand powder or a combination of the two, and the binder is furan resin.

9. A digital image-based layered rock preparation device, applied to the method according to any one of claims 1 to 7, characterized in that: include: The intelligent identification and parameter extraction module is used to intelligently identify and extract parameters of the layered rock mass of the target layer rock sample to obtain the layered structure characteristic information of the sample; A digital 3D reconstruction module is used to perform high-precision digital 3D reconstruction according to the layered structure to obtain a corresponding digital reconstruction model; The printing module is used to prepare layered rock specimens according to the digital reconstruction model and layer structure characteristic information of the sample.

10. The method for preparing layered rock mass based on digital images according to claim 9, characterized in that: The intelligent identification and parameter extraction module is provided with: The layered structure image database contains scanning results of multiple layered rock specimens as training set images and test set images; The MD-LinkNet model is used to accurately identify and segment the bedding structure of image data in the bedding structure image database, extract the bedding skeleton in the bedding recognition result map, vectorize the skeleton, and extract the key features of the bedding parameters based on the vectorized skeleton.

Images