Rock cross-scale parameter correlation and geometric feature matching method based on machine learning

Abstract

The invention discloses a rock cross-scale parameter correlation and geometric feature matching method based on machine learning. Comprising the steps of obtaining a rock microstructure digital image; inputting the rock microstructure digital image into a rock microstructure phase segmentation model to obtain a divided rock microstructure phase; constructing a rock microscopic-mesoscopic parameter correlation model to obtain rock mesoscopic parameters; constructing a rock microcosmic-microcosmic-macroscopic cross-scale deep learning generalized generative model, and matching rock microcosmic parameters with rock geometric structure features to obtain a rock microcosmic parameter-geometric feature matching model; constructing a rock microcosmic-macroscopic parameter correlation model to obtain rock macroscopic parameters; and based on a rock microscopic-microscopic-macroscopic cross-scale deep learning generalized generative model, matching the rock macroscopic parameters with rock geometric structure features to obtain a rock macroscopic parameter-geometric feature matching model. According to the method, cross-scale correlation of rock microscopic-microscopic-macroscopic physical and mechanical parameters and geometrical characteristics is realized.

Description

Technical Field

[0001] The present invention relates to the technical fields of rock mechanics and machine learning, and particularly relates to a method, device, storage medium and electronic device for cross-scale parameter correlation and geometric feature matching of rocks based on machine learning. Background Art

[0002] The geometric features of the microscopic material phase structure of rocks and their physical and mechanical properties determine the physical, mechanical and fracture characteristics of the overall rock, and play a very important role in deeply revealing the full-scale fracture mechanism of rocks from the microscopic to the mesoscopic to the macroscopic scale. In-depth research on the whole-process fracture evolution law and mechanism of rocks can greatly improve the efficiency and technological development of deep-earth resource exploitation (such as shale gas and geothermal energy, etc.), and at the same time promote the early warning and prevention and control technology development of deep engineering disasters (such as tunnel collapses and leakage of buried CO 2 - leakage of nuclear waste materials, etc.), and ultimately has great significance for ensuring the safety of talents and property and the stable and sustainable development of the national economy.

[0003] At present, modern high-precision imaging technologies with multi-scale imaging resolutions, such as X-ray computed tomography topology imaging (XCT), electron backscatter diffraction (EBSD), and focused ion beam scanning electron microscopy (FIB-SEM), etc., have greatly promoted the research on the microscopic structure of geotechnical engineering materials. Such technologies can clearly and intuitively capture the microscopic pore, crack and mineral matrix structure characteristics of rocks, and then deeply study the microscopic mechanical properties and fracture mechanisms of rocks. However, the research on the geometric structure characteristics and physical and mechanical properties of rocks at a specific scale can only reveal the fracture behavior and mechanism of rocks at the current scale, resulting in an inability to form an effective unified correlation understanding of the full-scale fracture mechanism and mechanical properties of multi-scale rock materials from the microscopic to the mesoscopic to the macroscopic scale, and thus unable to be further applied to engineering scale research, seriously affecting the development of deep-earth resource exploitation technologies and deep engineering disaster monitoring and prevention and control technologies. Its deficiencies are mainly reflected in the fact that the existing technologies cannot establish the correlation characteristics of rock physical and mechanical parameters at the microscopic-mesoscopic-macroscopic scales, resulting in an inability to effectively characterize the inhomogeneous characteristics of rock physical and mechanical parameters. Summary of the Invention

[0004] Embodiments of the present invention provide a method, device, storage medium and electronic device for cross-scale parameter correlation and geometric feature matching of rocks based on machine learning, which can establish the correlation characteristics of rock physical and mechanical parameters and the matching of geometric structure characteristics at the microscopic-mesoscopic-macroscopic scales, and effectively characterize the inhomogeneous characteristics of rock physical and mechanical parameters.

[0005] Embodiments of the present invention provide a method for cross-scale parameter correlation and geometric feature matching of rocks based on machine learning, including: Generating a digital image of the rock microscopic structure based on the rock microscopic specimen; Input the digital image of the rock microstructure into the rock microstructure phase segmentation model to obtain the well-divided rock microstructure phases; Construct a rock micro-meso parameter correlation model, and obtain rock meso parameters based on the rock microstructure phases and the rock micro-meso parameter correlation model; Construct a rock micro-meso-macro cross-scale deep learning generalized generative model, match the rock meso parameters with the rock geometric structure features, and obtain a rock meso parameter-geometric feature matching model; Construct a rock meso-macro parameter correlation model, and obtain rock macro parameters based on the rock microstructure phases and the rock meso-macro parameter correlation model; Based on the rock micro-meso-macro cross-scale deep learning generalized generative model, match the rock macro parameters with the rock geometric structure features, and obtain a rock macro parameter-geometric feature matching model.

[0006] Furthermore, in the above-mentioned method for rock cross-scale parameter correlation and geometric feature matching based on machine learning, where the step of inputting the digital image of the rock microstructure into the rock microstructure phase segmentation model to obtain the well-divided rock microstructure phases includes: Input the digital image of the rock microstructure into the rock microstructure phase segmentation model, and perform feature extraction, upsampling, and convolution operations respectively to obtain the rock microstructure prediction classification result; Perform digital encoding on the rock microstructure prediction classification result to obtain the well-divided rock microstructure phases.

[0007] Furthermore, in the above-mentioned method for rock cross-scale parameter correlation and geometric feature matching based on machine learning, where the step of inputting the digital image of the rock microstructure into the rock microstructure phase segmentation model, performing feature extraction, upsampling, and convolution operations respectively to obtain the rock microstructure prediction classification result is represented by the following formula:

[0008] where, is the i layer feature map, is upsampling, , is the side convolution kernel function, is the convolution operation and the i layer feature map of the underlying network, represents that in the i th convolution layer, the input feature map is subjected to convolution operation, is the output rock microstructure prediction classification result.

[0009] Furthermore, for the above-mentioned method for cross-scale parameter correlation and geometric feature matching of rocks based on machine learning, the rock micro-mesoscopic parameter correlation model is as follows:

[0010] Wherein, is the mesoscopic parameter of the rock, is the physical parameter of the i th rock microstructural phase itself, is the content of the i th rock microstructural phase, is the number of rock microstructural phases.

[0011] Furthermore, for the above-mentioned method for cross-scale parameter correlation and geometric feature matching of rocks based on machine learning, the rock mesoscopic-macroscopic parameter correlation model is as follows:

[0012]

[0013] Wherein, is the mesoscopic-macroscopic heterogeneity characterization parameter, is the macroscopic parameter of the rock.

[0014] Furthermore, for the above-mentioned method for cross-scale parameter correlation and geometric feature matching of rocks based on machine learning, the method further includes: Calculating the probability density function based on the mesoscopic-macroscopic heterogeneity characterization parameter, and determining the distribution information of each basic unit and the value information of each unit of the rock macroscopic parameter-geometric feature matching model based on the probability density function.

[0015] Furthermore, for the above-mentioned method for cross-scale parameter correlation and geometric feature matching of rocks based on machine learning, the probability density function is:

[0016] Wherein, x is a smallest basic unit of the rock macroscopic parameter-geometric feature matching model.

[0017] Furthermore, for the above-mentioned method for cross-scale parameter correlation and geometric feature matching of rocks based on machine learning, the joint loss function in the training process of the rock micro-mesoscopic-macroscopic cross-scale deep learning generalized generative model is:

[0018] Wherein, , , , They are the predicted macroscopic parameters of the rock, the true macroscopic parameters of the rock, the predicted mesoscopic parameters of the rock, and the true mesoscopic parameters of the rock respectively.

[0019] An embodiment of the present invention further provides a device for rock cross-scale parameter correlation and geometric feature matching based on machine learning, including: A digital image generation module for rock microstructures, configured to generate a digital image of a rock microstructure based on a rock micro-specimen; A rock microstructure phase segmentation module, configured to input the digital image of the rock microstructure into a rock microstructure phase segmentation model to obtain the divided rock microstructure phases; A rock mesoscopic parameter calculation module, configured to construct a rock micro-mesoscopic parameter correlation model, and obtain rock mesoscopic parameters based on the rock microstructure phases and the rock micro-mesoscopic parameter correlation model; A rock mesoscopic parameter-geometric feature matching module, configured to construct a rock micro-mesoscopic-macroscopic cross-scale deep learning generalized generative model, match the rock mesoscopic parameters with the rock geometric structure features, and obtain a rock mesoscopic parameter-geometric feature matching model; A rock macroscopic parameter calculation module, configured to construct a rock mesoscopic-macroscopic parameter correlation model, and obtain rock macroscopic parameters based on the rock microstructure phases and the rock mesoscopic-macroscopic parameter correlation model; A rock macroscopic parameter-geometric feature matching module, configured to match the rock macroscopic parameters with the rock geometric structure features based on the rock micro-mesoscopic-macroscopic cross-scale deep learning generalized generative model, and obtain a rock macroscopic parameter-geometric feature matching model.

[0020] An embodiment of the present invention further provides a computer-readable storage medium, in which multiple instructions are stored, and the instructions are suitable for being loaded by a processor to execute any one of the above-mentioned methods for rock cross-scale parameter correlation and geometric feature matching based on machine learning.

[0021] An embodiment of the present invention further provides an electronic device, including a processor and a memory, the processor is electrically connected to the memory, the memory is used to store instructions and data, and the processor is used for the steps in any one of the above-mentioned methods for rock cross-scale parameter correlation and geometric feature matching based on machine learning.

[0022] The method, device, storage medium and electronic device for cross-scale parameter correlation and geometric feature matching of rocks based on machine learning provided by the present invention first divides the microscopic structure phases of rocks, constructs a microscopic-mesoscopic parameter correlation model to calculate the mesoscopic parameters of rocks, constructs a mesoscopic-macroscopic parameter correlation model to calculate the macroscopic parameters of rocks, and then constructs a microscopic-mesoscopic-macroscopic cross-scale deep learning generalized generative model to perform the matching of mesoscopic structure and parameter features of rocks to generate a mesoscopic parameter-geometric feature matching model of rocks, and perform the matching of macroscopic units and parameter features of rocks to generate a macroscopic parameter-geometric feature matching model of rocks. The present invention greatly improves the recognition efficiency and accuracy of microscopic structure phases, and realizes the cross-scale correlation of microscopic-mesoscopic-macroscopic physical and mechanical parameters of rocks; the present invention also breaks through the scale limitation of the rock geometric structure, improves the accuracy of the heterogeneous rock geometric structure characterization and the correlation of multi-scale physical and mechanical parameters, and significantly improves the accuracy and efficiency of numerical simulation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The following, in conjunction with the drawings, through a detailed description of the specific embodiments of the present invention, will make the technical solutions and other beneficial effects of the present invention obvious.

[0024] Figure 1 It is a flowchart of the method for cross-scale parameter correlation and geometric feature matching of rocks based on machine learning provided by an embodiment of the present invention.

[0025] Figure 2 It is a schematic diagram of the segmentation of the sandstone microscopic structure by using the rock microscopic structure phase segmentation model provided by an embodiment of the present invention.

[0026] Figure 3 It is a flowchart of the matching of the mesoscopic parameters of rocks and the geometric structure features of rocks provided by an embodiment of the present invention.

[0027] Figure 4 It is a flowchart of the matching of the macroscopic parameters of rocks and the geometric structure features of rocks provided by an embodiment of the present invention.

[0028] Figure 5 It is a schematic diagram of the structure of the device for cross-scale parameter correlation and geometric feature matching of rocks based on machine learning provided by an embodiment of the present invention.

[0029] Figure 6 It is a schematic diagram of the structure of the electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0030] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0031] At present, modern high-precision imaging technologies with multi-scale imaging resolution, such as X-ray computed tomography topology imaging (XCT), electron backscatter diffraction (EBSD), and focused ion beam scanning electron microscopy (FIB-SEM), etc., have greatly promoted the research on the microstructure of geotechnical engineering materials. Such technologies can clearly and intuitively capture the microscopic pore, crack, and mineral matrix structure characteristics of rocks, and then deeply study the microscopic mechanical properties and fracture mechanisms of rocks. However, the research on the geometric structure characteristics and physical and mechanical properties of rocks at a specific scale can only reveal the fracture behavior and mechanism of rocks at the current scale, resulting in an inability to form an effective unified correlation understanding of the micro-meso-macro full-scale fracture mechanism and mechanical properties of multi-scale rock materials. Therefore, it cannot be further applied to engineering scale research, seriously affecting the development of deep earth resource extraction technologies and deep engineering disaster monitoring and prevention technologies. Its deficiencies are mainly reflected in the following aspects: (1) The existing technologies cannot establish the correlation characteristics of rock physical and mechanical parameters at the micro-meso-macro scales, resulting in an inability to effectively characterize the inhomogeneous characteristics of rock physical and mechanical parameters.

[0032] (2) The existing technologies cannot establish the characteristic matching relationship between the irregular geometric structures and physical and mechanical parameters of rocks at multiple scales, resulting in an inability to accurately establish a multi-scale inhomogeneous calculation model consistent with the rock structure.

[0033] (3) Due to the large differences between the rock physical and mechanical parameters and the geometric model in the existing numerical simulation technologies and the actual situation, the calculation accuracy is low and the accuracy is poor, and it cannot provide high-precision construction design parameters for actual engineering.

[0034] To solve the above problems, the embodiments of the present invention provide a method, device, storage medium, and electronic device for cross-scale parameter correlation and geometric feature matching of rocks based on machine learning. A device for cross-scale parameter correlation and geometric feature matching of rocks based on machine learning provided by the embodiments of the present invention can be integrated in an electronic device, and the electronic device can be a device such as a terminal or a server. Among them, the terminal can include a tablet computer, a notebook computer, a personal computer (PC), a micro-processing box, or other devices, etc.

[0035] Please refer to Figure 1 , Figure 1The flowchart of the method for cross-scale parameter association and geometric feature matching of rocks based on machine learning provided by an embodiment of the present invention is applied to an electronic device. The method for cross-scale parameter association and geometric feature matching of rocks based on machine learning includes the following steps: S1. Generate a digital image of the rock microstructure based on a microscopic rock specimen.

[0036] Specifically, modern high-precision imaging technology is used to perform non-destructive and refined capture and transparent imaging of the microscopic rock specimen to generate a digital image of the rock microstructure.

[0037] In a specific example, a microscopic rock test specimen is processed and prepared (specimen size: diameter = 5 mm, height = 10 mm), and modern high-precision imaging technology (such as X-ray computed tomography topology imaging technology with an imaging resolution of 2 μm) is used to perform non-destructive and refined capture and transparent imaging of the rock microstructure to generate a digital image of the rock microstructure.

[0038] S2. Input the digital image of the rock microstructure into the rock microstructure phase segmentation model to obtain the segmented rock microstructure phases.

[0039] Specifically, based on the deep learning U-Net segmentation model as the basic framework, a rock microstructure phase segmentation model (FPN-U-Net) based on the feature pyramid algorithm is established.

[0040] In one embodiment, step S2 includes the following steps: S21. Input the digital image of the rock microstructure into the rock microstructure phase segmentation model, and perform feature extraction, upsampling, and convolution operations respectively to obtain the predicted classification result of the rock microstructure.

[0041] Specifically, it is represented by the following formula:

[0042] Among them, is the feature map of the i layer, is upsampling, , is the side convolution kernel function, is the convolution operation and the feature map of the i layer of the underlying network, represents that in the i th convolution layer, the input feature map is subjected to a convolution operation, is the output predicted classification result of the rock microstructure.

[0043] The rock microstructure phase segmentation model is iteratively trained based on a loss function based on the fractal dimension. The expression of the loss function based on the fractal dimension is:

[0044] Among them, is the loss function based on the fractal dimension, is the predicted fractal dimension value, is the true fractal dimension value.

[0045] Furthermore, the difference between the predicted classification result of the rock microstructure and the true rock microstructure image is measured by calculating the contour similarity parameter to provide the accuracy of rock microstructure identification and division. The calculation formula of the contour similarity is:

[0046] Among them, is the contour similarity parameter, is the image dimension, and are the predicted classification result of the rock microstructure and the true classification image of the rock microstructure, respectively.

[0047] S22, the predicted classification result of the rock microstructure is digitally encoded to obtain the divided rock microstructure phases.

[0048] Figure 2 This is the schematic diagram of the sandstone microstructure segmented by the rock microstructure phase segmentation model provided by the embodiment of the present invention. As shown in Figure 2 By segmenting the original digital map of the sandstone microstructure with the rock microstructure phase segmentation model, the predicted classification result of the sandstone microstructure is obtained. As an example, a total of 4 microstructure phases are identified, which are 0 - pore phase, 1 - feldspar phase, 2 - quartz phase, and 3 - muscovite phase in sequence.

[0049] S3, construct a rock micro - meso parameter correlation model, and obtain the rock meso parameters based on the rock microstructure phases and the rock micro - meso parameter correlation model.

[0050] The rock micro - meso parameter correlation model is:

[0051] Among them, is the rock meso parameter, is the physical parameter of the i th rock microstructure phase itself, is the content of the i th rock microstructure phase, is the number of rock microstructure phases.

[0052] As a specific embodiment, the mesoscopic elastic modulus (one of the rock mesoscopic parameters) can be calculated through the rock microscopic-mesoscopic parameter correlation model, and the elastic models and contents of each phase are obtained as follows: 0 - pore phase (0 GPa, 0.15), 1 - feldspar phase (7 GPa, 0.55), 2 - quartz phase (8 GPa, 0.2), 3 - muscovite phase (15 GPa, 0.1). Thus, the rock mesoscopic elastic modulus obtained by the rock microscopic-mesoscopic parameter correlation model is 0×0.15 + 7×0.55 + 8×0.2 + 15×0.1 = 6.97 GPa.

[0053] S4. Construct a rock microscopic-mesoscopic-macroscopic cross-scale deep learning generalized generative model to match the rock mesoscopic parameters with the rock geometric structure features, and obtain a rock mesoscopic parameter-geometric feature matching model.

[0054] Figure 3 The flowchart for matching the rock mesoscopic parameters with the rock geometric structure features provided by the embodiments of the present invention can be referred to Figure 3 。

[0055] Specifically, taking the generalized generative network (GAN) as the basic framework, the feature pyramid (FPN) algorithm is used to identify and extract the rock microscopic structure features, and a rock microscopic-mesoscopic-macroscopic cross-scale deep learning generalized generative (DL-GAN) model is established.

[0056] Specifically, the rock mesoscopic parameters in step S2 (as the real rock mesoscopic parameters) and the rock microscopic structure phase are input into the rock microscopic-mesoscopic-macroscopic cross-scale deep learning generalized generative model to obtain the predicted rock mesoscopic parameters. The rock microscopic-mesoscopic-macroscopic cross-scale deep learning generalized generative model is trained through the real rock mesoscopic parameters and the joint loss function, and hyperparameters such as the batch size, learning rate, and number of iterations of the DL-GAN model are adjusted, and parameters such as the standard deviation, contour similarity coefficient, and accuracy are analyzed. 30% of the preprocessed rock microscopic structure digital image dataset is used as the input parameter to obtain the optimized DL-GAN model. Then, 70% of the rock microscopic structure digital image dataset with digital coding is used as the input parameter, and the optimized DL-GAN model is used for testing and verification to obtain the DL-GAN model with the best generalization ability.

[0057] Among them, the joint loss function in the training process of the rock microscopic-mesoscopic-macroscopic cross-scale deep learning generalized generative model is:

[0058] Among them, 、 、 、 They are the predicted macroscopic parameters of the rock, the true macroscopic parameters of the rock, the predicted mesoscopic parameters of the rock, and the true mesoscopic parameters of the rock respectively.

[0059] The standard deviation of the fractal dimension is:

[0060] Among them, is the square root function.

[0061] The accuracy parameter is:

[0062] Among them, TP, TN, FP, and FN are the true positive, true negative, false positive, and false negative respectively.

[0063] Then, through the optimized rock micro-meso-macro cross-scale deep learning generalized generative model, the mesoscopic parameters of the rock are correlated with the structural characteristics in the rock structure phase to generate a rock mesoscopic parameter-geometry feature matching model.

[0064] S5. Construct a rock meso-macro parameter correlation model, and obtain the macroscopic parameters of the rock based on the rock microstructural phase and the rock meso-macro parameter correlation model.

[0065] The rock meso-macro parameter correlation model is:

[0066]

[0067] Among them, is the meso-macro heterogeneity characterization parameter, is the macroscopic parameter of the rock.

[0068] Furthermore, this method also includes: calculating the probability density function based on the above-mentioned meso-macro heterogeneity characterization parameter, and determining the distribution information of each basic unit and the value information of each unit of the rock macroscopic parameter-geometry feature matching model based on the probability density function.

[0069] The probability density function is:

[0070] Among them, x is a smallest basic unit of the rock macroscopic parameter-geometry feature matching model, is the Weibull probability density function.

[0071] S6. Based on the rock micro-meso-macro cross-scale deep learning generalized generative model, match the macroscopic parameters of the rock with the rock geometric structure characteristics to obtain a rock macroscopic parameter-geometry feature matching model.

[0072] Figure 4 This is a flowchart for matching the macroscopic parameters of rocks with the geometric structure characteristics of rocks provided by the embodiments of the present invention. Reference can be made to Figure 4 . Similarly, first, the macroscopic parameters of the rock are predicted through a microscopic-mesoscopic-macroscopic cross-scale deep learning generalized generative model of the rock, and then the macroscopic parameters of the rock are correlated with the structural characteristics in the microscopic structure of the rock to generate a macroscopic parameter-geometric feature matching model of the rock.

[0073] In the present invention, the microscopic structure phase of the rock is first divided, a microscopic-mesoscopic parameter correlation model is constructed to calculate the mesoscopic parameters of the rock, a mesoscopic-macroscopic parameter correlation model of the rock is constructed to calculate the macroscopic parameters of the rock, and then a microscopic-mesoscopic-macroscopic cross-scale deep learning generalized generative model of the rock is constructed to perform the matching of the mesoscopic structure and parameter characteristics of the rock to generate a mesoscopic parameter-geometric feature matching model of the rock, and perform the matching of the macroscopic unit and parameter characteristics of the rock to generate a macroscopic parameter-geometric feature matching model of the rock. The present invention includes the following beneficial technical effects: (1) Break through the scale limitation of the rock structure and accurately correlate the microscopic-mesoscopic-macroscopic physical and mechanical parameters of the rock; (2) Can directly obtain the shape parameters and scale parameters of the inhomogeneity distribution of the rock and quickly generate a parametric calculation model at the macroscopic scale of the rock; (3) Improve the correlation accuracy between the geometric structure characterization of inhomogeneous rocks and multi-scale physical and mechanical parameters, and ultimately improve the calculation efficiency and accuracy of rock numerical simulation.

[0074] According to the method described in the above embodiments, this embodiment will further describe from the perspective of a device for cross-scale parameter correlation and geometric feature matching of rocks based on machine learning. The device for cross-scale parameter correlation and geometric feature matching of rocks based on machine learning can be specifically implemented as an independent entity or integrated in an electronic device. The electronic device can be a terminal, a server, or other devices. Among them, the terminal can include a tablet computer, a laptop computer, a personal computer (PC), a microprocessing box, or other devices.

[0075] Please refer to Figure 5 , Figure 5 Specifically describes a device for cross-scale parameter correlation and geometric feature matching of rocks based on machine learning provided by the embodiments of the present invention, which is applied to an electronic device. The device for cross-scale parameter correlation and geometric feature matching of rocks based on machine learning can include: A microscopic rock structure digital image generation module, configured to generate a microscopic rock structure digital image based on a microscopic rock specimen; A rock microstructure phase segmentation module, which is used to input a digital image of a rock microstructure into a rock microstructure phase segmentation model to obtain the divided rock microstructure phases; A rock mesoscopic parameter calculation module, which is used to construct a rock micro-mesoscopic parameter correlation model and obtain rock mesoscopic parameters based on the rock microstructure phases and the rock micro-mesoscopic parameter correlation model; A rock mesoscopic parameter - geometric feature matching module, which is used to construct a rock micro-mesoscopic-macroscopic cross-scale deep learning generalized generative model to match the rock mesoscopic parameters with the rock geometric structure features and obtain a rock mesoscopic parameter - geometric feature matching model; A rock macroscopic parameter calculation module, which is used to construct a rock mesoscopic-macroscopic parameter correlation model and obtain rock macroscopic parameters based on the rock microstructure phases and the rock mesoscopic-macroscopic parameter correlation model; A rock macroscopic parameter - geometric feature matching module, which is used to match the rock macroscopic parameters with the rock geometric structure features based on the rock micro-mesoscopic-macroscopic cross-scale deep learning generalized generative model to obtain a rock macroscopic parameter - geometric feature matching model.

[0076] In specific implementation, each of the above modules and / or units can be implemented as an independent entity, or can be combined arbitrarily to be implemented as the same or several entities. For the specific implementation of each of the above modules and / or units, reference can be made to the method embodiments described above. For the specific beneficial effects that can be achieved, reference can also be made to the beneficial effects in the method embodiments described above, which will not be elaborated here.

[0077] In addition, an embodiment of the present invention further provides an electronic device, which can be a device such as a computer or a tablet computer. The electronic device can implement the steps in any embodiment of the method for rock cross-scale parameter correlation and geometric feature matching based on machine learning provided by the embodiments of the present invention. Therefore, it can achieve the beneficial effects that can be achieved by any method for rock cross-scale parameter correlation and geometric feature matching based on machine learning provided by the embodiments of the present invention. For details, reference can be made to the previous embodiments, which will not be elaborated here.

[0078] Figure 6 The specific structural block diagram of the electronic device provided by the embodiment of the present invention is shown. The electronic device can be used to implement the method for rock cross-scale parameter correlation and geometric feature matching based on machine learning provided in the above embodiments. The electronic device 500 can be a device such as a terminal or a server. Among them, the terminal can include a tablet computer, a notebook computer, a personal computer (PC), a micro-processing box, or other devices, etc.

[0079] The RF circuit 510 is used to receive and transmit electromagnetic waves, realizing the mutual conversion between electromagnetic waves and electrical signals, so as to communicate with a communication network or other devices. The RF circuit 510 may include various existing circuit components for performing these functions, such as antennas, radio frequency transceivers, digital signal processors, encryption / decryption chips, subscriber identity module (SIM) cards, memories, and so on. The RF circuit 510 can communicate with various networks such as the Internet, enterprise intranets, wireless networks or communicate with other devices through a wireless network. The above-mentioned wireless networks may include cellular phone networks, wireless local area networks or metropolitan area networks. The above-mentioned wireless networks can use various communication standards, protocols and technologies, including but not limited to Global System for Mobile Communication (GSM), Enhanced Data GSM Environment (EDGE), Wideband Code Division Multiple Access (WCDMA), Code Division Access (CDMA), Time Division Multiple Access (TDMA), Wireless Fidelity (Wi-Fi) (such as Institute of Electrical and Electronics Engineers standards IEEE 802.11a, IEEE 802.11b, IEEE 802.11g and / or IEEE 802.11n), Voice over Internet Protocol (VoIP), Worldwide Interoperability for Microwave Access (Wi-Max), other protocols for e-mail, instant messaging and short messages, and any other suitable communication protocols, and may even include those protocols that have not yet been developed currently.

[0080] The memory 520 can be used to store software programs and modules, such as the corresponding program instructions / modules in the above embodiments. The processor 580 executes various functional applications and data processing by running the software programs and modules stored in the memory 520, that is, to implement functions such as taking pictures with the front camera, processing the captured images, and switching the display colors of the display content on the display screen. The memory 520 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memories, or other non-volatile solid-state memories. In some instances, the memory 520 may further include a memory remotely disposed relative to the processor 580, and these remote memories can be connected to the electronic device 500 through a network. Examples of the above network include but are not limited to the Internet, enterprise intranets, local area networks, mobile communication networks, and combinations thereof.

[0081] The input unit 530 can be used to receive input digital or character information, as well as generate keyboards and mice related to user settings and function controls. The display unit 540 can be used to display information input by the user or information provided to the user, as well as various graphical user interfaces, and these graphical user interfaces can be composed of graphics, text, icons, videos, and any combination thereof. The display unit 540 may include a display panel 541. Optionally, the display panel 541 can be configured in the form of an LCD (Liquid Crystal Display) or an OLED (Organic Light-Emitting Diode).

[0082] The audio circuit 560, speaker 561, and microphone 562 can provide an audio interface between the user and the electronic device 500. The audio circuit 560 can transmit the electrical signal converted from the received audio data to the speaker 561, and the speaker 561 converts it into a sound signal for output; on the other hand, the microphone 562 converts the collected sound signal into an electrical signal, which is received by the audio circuit 560 and then converted into audio data. After the audio data is output to the processor 580 for processing, it is sent through the RF circuit 510 to, for example, another terminal, or the audio data is output to the memory 520 for further processing. The audio circuit 560 may also include an earphone jack to provide communication between the peripheral earphone and the electronic device 500.

[0083] The electronic device 500 can help the user receive requests, send information, etc. through the transmission module 570 (such as a Wi-Fi module), and it provides the user with wireless broadband Internet access. Although the transmission module 570 is shown in the figure, it can be understood that it does not belong to the essential components of the electronic device 500 and can be omitted completely as needed without changing the essence of the invention.

[0084] The processor 580 is the control center of the electronic device 500, connecting various parts of the entire mobile phone through various interfaces and circuits. By running or executing software programs and / or modules stored in the memory 520, and by invoking the data stored in the memory 520, it executes various functions of the electronic device 500 and processes data, thereby monitoring the electronic device as a whole. Optionally, the processor 580 may include one or more processing cores; in some embodiments, the processor 580 may integrate an application processor and a modem processor. Among them, the application processor mainly processes the operating system, user interface, application programs, etc., and the modem processor mainly processes wireless communication. It can be understood that the above-mentioned modem processor may not be integrated into the processor 580 either.

[0085] The electronic device 500 also includes a power source 590 (such as a battery) for supplying power to each component. In some embodiments, the power source can be logically connected to the processor 580 through a power management system, so as to realize functions such as management of charging, discharging, and power consumption management through the power management system. The power source 590 may also include any components such as one or more DC or AC power sources, a recharge system, a power failure detection circuit, a power converter or inverter, and a power status indicator.

[0086] Although not shown, the electronic device 500 also includes a camera (such as a front camera and a rear camera), a Bluetooth module, etc., which will not be elaborated here. Specifically, in this embodiment, the display unit of the electronic device is a touch screen display, and the mobile terminal also includes a memory, and one or more programs, where one or more programs are stored in the memory and are configured to be executed by one or more processors. The one or more programs include instructions for performing the following operations: Generating a digital image of the rock micro-structure based on a rock micro-specimen; Inputting the digital image of the rock micro-structure into a rock micro-structure phase segmentation model to obtain the segmented rock micro-structure phases; Constructing a rock micro-meso parameter correlation model, and obtaining rock meso parameters based on the rock micro-structure phases and the rock micro-meso parameter correlation model; Constructing a rock micro-meso-macro cross-scale deep learning generalized generative model, matching the rock meso parameters with the rock geometric structure characteristics to obtain a rock meso parameter-geometric feature matching model; Constructing a rock meso-macro parameter correlation model, and obtaining rock macro parameters based on the rock micro-structure phases and the rock meso-macro parameter correlation model; Based on the rock microscopic-mesoscopic-macroscopic cross-scale deep learning generalized generative model, the macroscopic parameters of the rock are matched with the geometric structure characteristics of the rock to obtain a macroscopic parameter-geometric feature matching model of the rock.

[0087] In specific implementation, each of the above modules can be implemented as an independent entity, or can be combined arbitrarily and implemented as the same or several entities. For the specific implementation of each of the above modules, reference can be made to the foregoing method embodiments, which will not be elaborated herein.

[0088] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructions, or the relevant hardware can be controlled by instructions. The instructions can be stored in a computer-readable storage medium and loaded and executed by a processor. To this end, an embodiment of the present invention provides a storage medium, in which multiple instructions are stored, and the instructions can be loaded by a processor to execute the steps of any one of the embodiments of the method for cross-scale parameter association and geometric feature matching of rocks based on machine learning provided by the embodiments of the present invention.

[0089] Among them, the computer-readable storage medium may include: read-only memory (ROM, Read Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disk, etc.

[0090] Since the instructions stored in the storage medium can execute the steps in any one of the embodiments of the method for cross-scale parameter association and geometric feature matching of rocks based on machine learning provided by the embodiments of the present invention, the beneficial effects that can be achieved by any one of the methods for cross-scale parameter association and geometric feature matching of rocks based on machine learning provided by the embodiments of the present invention can be realized. For details, refer to the foregoing embodiments, which will not be elaborated herein.

[0091] The above has introduced in detail a method, device, storage medium and electronic device for cross-scale parameter association and geometric feature matching of rocks based on machine learning provided by the embodiments of the present invention. Specific examples are used in this article to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those skilled in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A rock cross-scale parameter association and geometric feature matching method based on machine learning, characterized in that: The method comprises: Generate digital images of rock microstructure based on rock micro specimens; Inputting the rock microstructure digital image into a rock microstructure phase segmentation model to obtain a divided rock microstructure phase; Constructing a rock micro-meso parameter correlation model, and obtaining rock meso parameters based on the rock microstructural phase and the rock micro-meso parameter correlation model; Constructing a rock micro-meso-macro cross-scale deep learning generalized generative model, matching the rock meso parameters with the rock geometric structure characteristics, and obtaining a rock meso parameter-geometric feature matching model; Constructing a rock meso-macro parameter correlation model, and obtaining rock macro parameters based on the rock microstructure phase and the rock meso-macro parameter correlation model; Based on the rock microscopic-mesoscopic-macroscopic cross-scale deep learning generalized generative model, the rock macroscopic parameters are matched with rock geometric structure characteristics to obtain a rock macroscopic parameter-geometric characteristic matching model.

2. The rock cross-scale parameter association and geometric feature matching method based on machine learning according to claim 1 is characterized in that: The step of inputting the rock microstructure digital image into a rock microstructure phase segmentation model to obtain a divided rock microstructure phase comprises: Inputting the rock microstructure digital image into a rock microstructure phase segmentation model, performing feature extraction, upsampling and convolution operations respectively to obtain rock microstructure prediction and classification results; The rock microstructure prediction and classification results are digitally encoded to obtain the divided rock microstructure phases.

3. The rock cross-scale parameter association and geometric feature matching method based on machine learning according to claim 2 is characterized in that: The rock microstructure digital image is input into the rock microstructure phase segmentation model, and feature extraction, upsampling and convolution operations are performed to obtain the rock microstructure prediction classification result, which is expressed by the following formula: in, For the i Layer feature map, For upsampling, is the convolution kernel function, is the side convolution kernel function, The convolution operation is the first i Layer feature map, Indicated in i In the convolutional layer, the input feature map Perform convolution operation, The classification results are predicted for the output rock microstructure.

4. The rock cross-scale parameter association and geometric feature matching method based on machine learning according to claim 1 is characterized in that: The rock micro-meso parameter correlation model is: in, is the rock microscopic parameter, For the i The physical parameters of the rock microstructure phase itself, For the i The content of each rock microstructure phase, is the number of rock microstructural phases.

5. The rock cross-scale parameter association and geometric feature matching method based on machine learning according to claim 1 is characterized in that: The rock micro-macro parameter correlation model is: in, is the parameter characterizing the micro-macro heterogeneity, is the macroscopic parameter of rock.

6. The rock cross-scale parameter association and geometric feature matching method based on machine learning according to claim 5 is characterized in that: The method further comprises: Based on the micro-macroscopic heterogeneity characterization parameters, a probability density function is calculated, and based on the probability density function, the distribution information of each basic unit and the value information of each unit of the rock macroscopic parameter-geometric feature matching model are determined.

7. The rock cross-scale parameter association and geometric feature matching method based on machine learning according to claim 6 is characterized in that: The probability density function is: in, x It is the smallest basic unit of the rock macro-parameter-geometric characteristic matching model.

8. The rock cross-scale parameter association and geometric feature matching method based on machine learning according to claim 1 is characterized in that: The joint loss function of the training process of the rock micro-meso-macro cross-scale deep learning generalized generative model is: in, , , , They are predicted rock macroscopic parameters, real rock macroscopic parameters, predicted rock mesoscopic parameters, and real rock mesoscopic parameters respectively.

9. A rock cross-scale parameter association and geometric feature matching device based on machine learning, characterized in that: include: A rock microstructure digital image generation module, used for generating a rock microstructure digital image based on a rock microsample; A rock microstructure phase segmentation module is used to input the rock microstructure digital image into a rock microstructure phase segmentation model to obtain a divided rock microstructure phase; A rock mesoscopic parameter calculation module is used to construct a rock microscopic-mesoscopic parameter correlation model, and obtain rock mesoscopic parameters based on the rock microstructural phase and the rock microscopic-mesoscopic parameter correlation model; A rock mesoscopic parameter-geometric feature matching module is used to construct a rock microscopic-mesoscopic-macroscopic cross-scale deep learning generalized generative model, match the rock mesoscopic parameters with the rock geometric structure characteristics, and obtain a rock mesoscopic parameter-geometric feature matching model; A rock macro-parameter calculation module is used to construct a rock meso-macro parameter correlation model, and obtain rock macro-parameters based on the rock microstructure phase and the rock meso-macro parameter correlation model; The rock macro-parameter-geometric feature matching module is used to match the rock macro-parameters with the rock geometric structure characteristics based on the rock micro-meso-macro cross-scale deep learning generalized generative model to obtain a rock macro-parameter-geometric feature matching model.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a plurality of instructions, which are suitable for being loaded by a processor to execute the rock cross-scale parameter association and geometric feature matching method based on machine learning as described in any one of claims 1 to 8.

Images