Rock identification method, system, device, terminal and readable storage medium

Abstract

The present application relates to a rock identification method, system, device, terminal and readable storage medium, and relates to the field of rock identification technology. The method includes: receiving a rock thin section image sent by an image acquisition device; generating geometric features, mineral features and structural features corresponding to the rock thin section based on the rock thin section image; generating an identification result of the rock thin section based on the geometric features, mineral features and structural features. After obtaining the rock thin section image, feature extraction based on three dimensions of geometric features, mineral features and structural features is performed on the rock thin section image, the properties of the rock are determined from multiple feature dimensions, and finally an identification result including a textual description is generated. After obtaining the microscopic visualization image corresponding to the rock thin section, feature extraction of multiple dimensions is performed on the image, and the rock thin section is identified with reference to the features of multiple dimensions, thereby improving the accuracy of rock identification.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims the priority of Chinese patent application CN 202110181773.4, filed on February 8, 2021, entitled “A rock thin section intelligent identification device and method” and Chinese patent application CN 202110182027.7, filed on February 10, 2021, entitled “A method for identifying organisms in carbonate rock thin section images”, the entire contents of which are incorporated herein by reference. Technical Field

[0003] The present application relates to the technical field of rock identification, and in particular to a rock identification method, device, terminal and readable storage medium. Background Art

[0004] Applying optical principles to cut and grind rocks into 30-micron thick slices, and using a polarizing microscope to observe and study the rock slices is one of the most basic, simplest and cheapest methods in scientific research and production in the geological industry, and is also one of the basic knowledge and basic skills that geologists must master.

[0005] The traditional rock thin section identification method is still used today. In the traditional rock thin section identification method, researchers usually identify the rock thin sections through manual identification and microscopic observation of the rock thin sections. At the same time, researchers also use equipment and computers to collect microscopic images to make more detailed observations of microscopic images.

[0006] However, in the related art, whether it is direct observation of rock slices or inputting rock slice images into a computer for observation, they are both affected by large subjective factors and have unavoidable errors, resulting in a low accuracy rate in rock identification. Summary of the invention

[0007] This application is about a rock identification method, system, device, terminal and readable storage medium, which can improve the accuracy of rock identification. The technical solution is as follows:

[0008] In one aspect, a rock identification method is provided. The method is applied to a computer device of a rock identification system. The rock identification system includes an image acquisition device and a computer device. The image acquisition device is connected to the computer device. The method includes:

[0009] receiving a rock thin section image sent by an image acquisition device, wherein the rock thin section image is an image obtained by photographing a rock thin section, wherein the rock thin section is a thin section obtained by cutting a rock sample, and the rock thin section image includes at least one component region;

[0010] Generate geometric features, mineral features and structural features corresponding to the rock thin section based on the rock thin section image, wherein the geometric features are used to indicate the regional division of the rock thin section based on the components, the mineral features are used to indicate the mineral identification results obtained by mineral identification of the rock thin section, and the structural features are used to indicate the spatial structure of the rock sample;

[0011] Based on geometric, mineral and structural characteristics, the identification results of rock thin sections are generated, including a textual description of the characteristics of the rock samples.

[0012] In an optional embodiment, the method further includes:

[0013] Acquire a sample slice image, where the sample slice image is annotated with a sample segmentation result;

[0014] The sample thin-section image is input into the rock thin-section image segmentation model, and the predicted segmentation result is output;

[0015] Compare the predicted segmentation result with the sample segmentation result to obtain the segmentation difference;

[0016] The rock thin section image segmentation model is adjusted based on the segmentation difference.

[0017] In an optional embodiment, after comparing the predicted segmentation result with the sample segmentation result, the method further includes:

[0018] In response to the segmentation difference indicating a difference threshold between the predicted segmentation result and the sample segmentation result, determining that the training of the rock thin section image segmentation model is completed;

[0019] or,

[0020] In response to the number of sample thin section image inputs to the rock thin section image segmentation model in a segmentation training cycle reaching a segmentation number threshold, it is determined that the training of the rock thin section image segmentation model is completed.

[0021] In an optional embodiment, the method further includes:

[0022] Acquire a sample component image, where the sample component image is annotated with sample component type features;

[0023] The sample component image is input into the component type recognition model, and the predicted component type feature is output;

[0024] Compare the predicted component type characteristics with the sample component type characteristics to obtain the identification difference;

[0025] The component type identification model was adjusted based on the identification differences.

[0026] In an optional embodiment, the non-single-crystalline component set image includes an amorphous component set image and a polycrystalline component set image, the non-single-crystalline component region includes an amorphous component region and a polycrystalline component region, the amorphous component region corresponds to an amorphous component identification, and the polycrystalline component region corresponds to a polycrystalline component identification result.

[0027] In an optional embodiment, the rock slice is a clastic rock slice;

[0028] The single crystal component area includes the quartz feldspar debris area and the siliceous calcareous cement area, and the non-single crystal component area includes the non-terrigenous debris area, the rock debris area, the matrix filling area and the mud cement area;

[0029] The classification and recognition model includes a classification sub-model group and a recognition sub-model group;

[0030] The classification sub-model group includes particle classification sub-model, clastic classification sub-model, terrigenous clastic classification sub-model, interstitial material classification sub-model and cementation material classification sub-model;

[0031] The identification sub-model group includes non-terrigenous debris identification sub-model, rock debris identification sub-model, quartz feldspar debris identification sub-model, siliceous calcareous cement identification sub-model, argillaceous cement identification sub-model and matrix filling identification sub-model;

[0032] The rock thin section segmentation image is input into the classification recognition model, and the output is a collection image of single crystal components and a collection image of non-single crystal components, including:

[0033] Input the clastic rock thin section into the grain classification sub-model, and output the clastic sub-image and the interstitial material sub-image;

[0034] The debris sub-image is input into the debris classification sub-model, and the non-terrigenous debris sub-image and the terrigenous debris sub-image are output;

[0035] The terrigenous debris sub-image is input into the terrigenous debris classification sub-model, and the rock debris sub-image and the quartz feldspar debris sub-image are output;

[0036] Input the interstitial material sub-image into the interstitial material classification sub-model, and output the interstitial material sub-image and the cementing material sub-image;

[0037] Input the cementation sub-image into the cementation classification sub-model, and output the siliceous calcareous cementation sub-image and the argillaceous cementation sub-image;

[0038] The non-terrigenous debris sub-image is input into the non-terrigenous debris identification sub-model, and the non-terrigenous debris identification result is output, and the non-terrigenous debris area is determined based on the non-terrigenous debris identification result;

[0039] Inputting the rock-cutting sub-image into the rock-cutting debris identification sub-model, outputting the rock-cutting debris identification result, and determining the rock-cutting area based on the rock-cutting debris identification result;

[0040] The quartz feldspar sub-image is input into the quartz feldspar fragment identification sub-model, and the quartz feldspar fragment identification result is output, and the quartz feldspar fragment area is determined based on the quartz feldspar fragment identification result;

[0041] Inputting the miscellaneous interstitial material sub-image into the miscellaneous interstitial material identification sub-model, outputting the miscellaneous interstitial material identification result, and determining the miscellaneous interstitial material region based on the miscellaneous interstitial material identification result;

[0042] Inputting the siliceous calcareous cement sub-image into the siliceous calcareous cement identification sub-model, outputting the siliceous calcareous cement identification result, and determining the siliceous calcareous cement area based on the siliceous calcareous cement identification result;

[0043] Inputting the mud cementation sub-image into the mud cementation identification sub-model, outputting the mud cementation identification result, and determining the mud cementation area based on the mud cementation identification result;

[0044] Based on the quartz feldspar debris area and the siliceous calcareous cement area, the single crystal component area is determined;

[0045] The non-single crystalline component area is determined based on the non-terrigenous debris area, rock debris area, mud cement area and matrix filling area.

[0046] In an optional embodiment, the method further includes:

[0047] Acquire a sample thin slice image, wherein the sample thin slice image is annotated with sample segmentation regions and sample component type features;

[0048] The sample thin section image is input into the rock thin section image recognition model, and the predicted segmentation area and predicted component type characteristics are output;

[0049] Compare the predicted segmentation area with the sample segmentation area to obtain the sample segmentation difference, and compare the predicted component type feature with the sample component type feature to obtain the sample component difference;

[0050] Based on the differences in sample segmentation and sample composition, the rock thin section image recognition model is adjusted.

[0051] In an optional embodiment, the rock thin section identification result includes the maximum grain size identification result;

[0052] Generate rock thin section structure identification results based on geometric and mineral features, including:

[0053] Determine the maximum grain component region corresponding to the maximum grain size identification result based on the mineral characteristics;

[0054] A maximum particle size corresponding to the component region is determined based on the geometric features, and a maximum particle size identification result is generated based on the maximum particle size.

[0055] In an optional embodiment, the rock thin section identification results include particle sorting results;

[0056] Generate rock thin section structure identification results based on geometric and mineral features, including:

[0057] Determine a sorting degree component region set corresponding to the particle sorting degree identification result based on the mineral characteristics, wherein the sorting degree component region set includes at least two component regions;

[0058] Determine a particle size classification correspondence relationship based on geometric features corresponding to component regions in the sorting degree component region set, wherein the particle size classification correspondence relationship includes at least two particle size levels and the number of component regions in the particle size level;

[0059] Based on the corresponding relationship between particle size classification, the particle sorting degree result is determined.

[0060] In an optional embodiment, the rock thin section identification result includes a particle roundness identification result;

[0061] Generate rock thin section structure identification results based on geometric and mineral features, including:

[0062] Determine a roundness identification component region set corresponding to the particle roundness identification result based on the mineral characteristics, wherein the roundness identification component region set includes at least two component regions;

[0063] Determine a roundness classification correspondence relationship based on geometric features corresponding to component regions in the roundness component identification region set, wherein the roundness classification correspondence relationship includes at least two roundness levels and the number of component regions in the roundness level;

[0064] Based on the corresponding relationship between the roundness grades, the particle roundness recognition result is determined.

[0065] In an optional embodiment, the rock thin section identification results include the results of the degree of grain eumorphism;

[0066] Generate rock thin section structure identification results based on geometric and mineral features, including:

[0067] Determine, based on the mineral characteristics, a set of euhedral component regions corresponding to the particle euhedral degree result, wherein the set of euhedral component regions includes at least two euhedral component regions;

[0068] Determine the shape of the component region based on the geometric features corresponding to the component region in the set of component regions of automorphic degree;

[0069] Based on the shapes of the component regions, the particle automorphicity results are determined.

[0070] In an optional embodiment, the rock thin section identification result includes the grain contact mode identification result;

[0071] Generate rock thin section structure identification results based on geometric and mineral features, including:

[0072] Determine at least two groups of component region pairs based on the mineral characteristics, wherein the component region pairs include two adjacent component regions;

[0073] Based on the geometric features, determining the intersection-and-union ratio of two component regions in the component region pair;

[0074] Based on the intersection-and-union ratio of at least two component regions, the particle contact mode identification result is determined.

[0075] In an optional embodiment, determining the name of the rock slice based on the rock slice identification result includes:

[0076] Based on the characteristics of the subdivided rock types, determine the naming rules corresponding to the names of rock thin sections;

[0077] The rock thin section name is generated based on the naming rules and the rock thin section structure identification results.

[0078] On the other hand, a rock identification system is provided, the rock identification system comprising an image acquisition device and a computer device, the image acquisition device being connected to the computer device;

[0079] An image acquisition device, used for generating rock thin section images; and sending the rock thin section images to a computer device;

[0080] A computer device is used to receive a rock thin section image sent by an image acquisition device, wherein the rock thin section image is an image obtained by photographing the rock thin section, and the rock thin section is a thin section obtained by cutting a rock sample, and the rock thin section image includes at least one component region; based on the rock thin section image, geometric features, mineral features and structural features corresponding to the rock thin section are generated, wherein the geometric features are used to indicate the regional division of the rock thin section based on the components, the mineral features are used to indicate the mineral identification results obtained by mineral identification of the rock thin section, and the structural features are used to indicate the spatial structure of the rock sample; based on the geometric features, mineral features and structural features, an identification result of the rock thin section is generated, and the identification result includes a textual description of the characteristics of the rock sample.

[0081] In another aspect, a rock identification device is provided, the device comprising:

[0082] A receiving module, used for receiving a rock thin section image sent by an image acquisition device, wherein the rock thin section image is an image obtained by photographing a rock thin section, wherein the rock thin section is a thin section obtained by cutting a rock sample, and the rock thin section image includes at least one component region;

[0083] A generation module, used to generate geometric features, mineral features and structural features corresponding to the rock thin section based on the rock thin section image, wherein the geometric features are used to indicate the regional division of the rock thin section based on the components, the mineral features are used to indicate the mineral identification results obtained by performing mineral identification on the rock thin section, and the structural features are used to indicate the spatial structure of the rock sample;

[0084] The generation module is used to generate the identification results of rock thin sections based on geometric features, mineral features and structural features, and the identification results include a textual description of the characteristics of the rock samples.

[0085] On the other hand, a computer device is provided, which includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, code set or instruction set, and the processor can load and execute at least one instruction, at least one program, code set or instruction set to implement the rock identification method provided in the above-mentioned embodiment of the present application.

[0086] On the other hand, a computer-readable storage medium is provided, in which at least one instruction, at least one program, code set or instruction set is stored. The processor can load and execute at least one instruction, at least one program, code set or instruction set to implement the rock identification method provided in the above-mentioned embodiment of the present application.

[0087] On the other hand, a computer program product or a computer program is provided, the computer program product or the computer program comprising computer program instructions, the computer program instructions being stored in a computer-readable storage medium. A processor reads the computer instructions from the computer-readable storage medium and executes the computer instructions, so that the computer device performs the rock identification method provided in the embodiments of the present application.

[0088] The beneficial effects of the technical solution provided by this application include at least:

[0089] After obtaining the rock thin section image, the rock thin section image is subjected to feature extraction based on three dimensions of geometric features, mineral features and structural features. From the perspective of microscopic composition and macroscopic performance, the properties of the rock are determined by integrating multiple feature dimensions, and finally an identification result including a textual description is generated. In the process of rock identification, after obtaining the microscopic visualization image corresponding to the rock thin section, the image is subjected to feature extraction in multiple dimensions, and the rock thin section is identified with reference to the features in multiple dimensions, thereby improving the accuracy of rock identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0090] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0091] Figure 1 A schematic diagram of a rock identification system provided by an exemplary embodiment of the present application is shown;

[0092] Figure 2 A schematic structural diagram of an image acquisition device provided by an exemplary embodiment of the present application is shown;

[0093] Figure 3 A flow chart of a rock identification method provided by an exemplary embodiment of the present application is shown;

[0094] Figure 4 A schematic diagram of the structure of a rock slice image provided by an exemplary embodiment of the present application is shown;

[0095] Figure 5 A flow chart of a method for determining geometric features, mineral features, and structural features provided by an exemplary embodiment of the present application is shown;

[0096] Figure 6 A schematic flow chart of a method for segmenting a rock thin section image provided by an exemplary embodiment of the present application is shown;

[0097] Figure 7 A structural block diagram of an image segmentation model provided by an exemplary embodiment of the present application is shown;

[0098] Figure 8 A schematic diagram of a segmented rock slice segmentation image provided by an exemplary embodiment of the present application is shown;

[0099] Fig. 9A flow chart showing a method for identifying mineral features provided by an exemplary embodiment of the present application is shown;

[0100] Fig.10 A schematic diagram of extracting segmented areas from a rock slice segmentation image provided by an exemplary embodiment of the present application is shown;

[0101] Fig.11 A schematic diagram of another rock identification system provided by an exemplary embodiment of the present application is shown;

[0102] Fig.12 A schematic diagram of a mineral data acquisition device provided by an exemplary embodiment of the present application is shown;

[0103] Fig.13 A schematic flow chart of a method for determining mineral features corresponding to a rock slice provided by an exemplary embodiment of the present application is shown;

[0104] Fig.14 A schematic diagram of a rock slice corresponding to clastic rock provided by an exemplary embodiment of the present application is shown;

[0105] Fig.15 A schematic structural diagram of a rock identification model corresponding to clastic rock provided by an exemplary embodiment of the present application is shown;

[0106] Fig.16 A schematic flow chart of a method for determining mineral features corresponding to a component region provided by an exemplary embodiment of the present application is shown;

[0107] Fig.17 A schematic diagram of the structure of a mineral spectrum database provided by an exemplary embodiment of the present application is shown;

[0108] Fig.18 A schematic diagram of a process for determining geometric features and mineral features provided by an exemplary embodiment of the present application is shown;

[0109] Fig.19 A schematic diagram showing a method for generating identification results of a rock slice provided by an exemplary embodiment of the present application;

[0110] Fig. 20 A schematic diagram of a rock identification method provided by an exemplary embodiment of the present application is shown;

[0111] Fig.21 A schematic diagram showing the contents of a sheet identification report provided by an exemplary embodiment of the present application is shown;

[0112] Fig. 22A schematic diagram of a rock identification method provided by an exemplary embodiment of the present application is shown;

[0113] Fig.23 A schematic diagram of a rock identification method provided by an exemplary embodiment of the present application is shown;

[0114] Fig.24 A schematic diagram of an intrusive rock classification phase diagram provided by an exemplary embodiment of the present application is shown;

[0115] Fig.25 A structural block diagram of a rock identification device provided by an exemplary embodiment of the present application is shown;

[0116] Fig.26 A structural block diagram of another rock identification device provided by an exemplary embodiment of the present application is shown;

[0117] Fig. 27 A schematic diagram of the structure of a computer device for executing a rock identification method provided by an exemplary embodiment of the present application is shown. DETAILED DESCRIPTION

[0118] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.

[0119] First, the terms involved in each embodiment of the present application are explained:

[0120] Component refers to each ingredient in a mixture. In the present application, component refers to the component of a solid material. Since the solid material is a mixture, the mixture includes at least two mutually separated components. The solid material is also called a solid multi-component mixed material. In one example, the mixture is implemented as a metal wire, and the components of the metal wire include at least one of lead, cadmium, bismuth, and tin; in another example, the mixture is implemented as a rock, and the components of the rock include but are not limited to at least one of a terrigenous debris component, a non-terrigenous debris component, a cement component, and a matrix component. In the present application, a component indicates a component contained in a rock thin section, and in a rock thin section image, there are different component areas corresponding to different components.

[0121] Image processing technology is a technology that uses computers to process image information. Image processing technology mainly includes image digitization, image enhancement and restoration, image data encoding, image segmentation and image recognition. Among them, image segmentation is the technology and process of dividing an image into several specific regions with unique properties and proposing targets of interest. The image segmentation method includes at least one of a threshold-based segmentation method, a region-based segmentation method, an edge-based segmentation method, and a segmentation method based on a specific theory. Image recognition refers to the technology of using computers to process, analyze and understand images to identify targets and objects of various different patterns. It is a practical application of deep learning algorithms. At this stage, image recognition technology is generally divided into face recognition and commodity recognition. Face recognition is mainly used in security inspections, identity verification and mobile payments; commodity recognition is mainly used in the process of commodity circulation, especially in unmanned retail fields such as unmanned shelves and smart retail cabinets. In the embodiment of this application, both image recognition and image segmentation technologies will be applied to the field of rock identification.

[0122] Artificial Intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. Artificial Intelligence attempts to understand the essence of intelligence and produce an intelligent machine that can respond in a similar way to human intelligence. The purpose of artificial intelligence is to enable machines to have the functions of perception, reasoning and decision-making.

[0123] Artificial intelligence technology is a comprehensive discipline covering a wide range of fields. Basic artificial intelligence technologies include, but are not limited to, sensor technology, artificial intelligence chip technology, cloud computing technology, big data processing technology, and mechatronics technology. The artificial intelligence technology used in the embodiments of this application is machine learning technology, and the machine learning device is applied to computer devices.

[0124] Machine Learning (ML) is a multi-disciplinary subject involving probability theory, statistics, algorithm complexity theory and other disciplines. The discipline of machine learning is specifically used to study how computers simulate or implement human learning behavior, so that computers can acquire new knowledge, reorganize existing knowledge structures, and thus improve their own performance. Machine learning is usually combined with deep learning. Machine learning and deep learning usually include artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and teaching learning.

[0125] Spectrum is a pattern in which the dispersed monochromatic light is arranged once according to the wavelength after the complex light is split by the dispersion system. The full name of spectrum is optical spectrum. In the embodiment of the present application, the spectrum is collected by the spectral data acquisition device for the measurement points in the component area. Taking the component as a mineral component as an example, when the mineral is irradiated with electromagnetic radiation, the molecules inside it will produce transitions between quantized energy levels, which will lead to the generation of emission, absorption or scattering, and radiation phenomena, thereby generating a spectrum. The spectrum is realized as a visual representation of spectral data. That is, while generating the spectrum, the spectral data acquisition device will also generate corresponding spectral data at the same time, and send such spectral data to other computer devices.

[0126] A database is a virtual warehouse that organizes, stores, and manages data according to a data structure. It is a collection of large amounts of data that is stored in a computer for a long time, organized, shareable, and uniformly managed. In the embodiment of the present application, corresponding to molecular spectroscopy, the computer device can access the spectral data database to confirm and analyze the spectral data.

[0127] Surface scanning, or surface scanning analysis, is a mainstream method for accurately identifying material components using test instruments. The usual practice is to cut the material components into thin slices or make samples with a smooth surface, and then use surface scanning to accurately identify the components therein. The test instruments used for surface scanning include but are not limited to molecular spectroscopy such as infrared spectroscopy and Raman spectroscopy, X-ray fluorescence spectroscopy (XRF), X-ray energy dispersive X-ray spectroscopy (EDS), etc., and in some embodiments of the present application, the surface scanning process can be combined with a microscope or a scanning electron microscope to combine microscopic observation with micro-area analysis.

[0128] In view of the fact that the manual rock identification method in the related art has a complicated identification process, extremely high professional requirements, but a low accuracy rate, this application provides a rock identification system and a rock identification method. With the help of artificial intelligence-based image recognition technology, multi-dimensional features are generated in the rock identification process, thereby improving the accuracy of rock identification.

[0129] Figure 1 A schematic diagram of a rock identification system provided by an exemplary embodiment of the present application is shown. The device includes an image acquisition device 101 and a computer device 102 , and the image acquisition device 101 is connected to the computer device 102 .

[0130] In the embodiment of the present application, the computer device is a terminal device for executing the rock identification method. The computer device has the functions of data transmission, data reception and data processing.

[0131] The image acquisition device is a device that provides images to the computer device during the process of the computer device executing the rock identification method. In the embodiment of the present application, the image acquisition device has the functions of image shooting and image sending. In one example, Figure 2 1 shows a schematic diagram of the structure of an image acquisition device provided by an exemplary embodiment of the present application, the image acquisition device includes a first light source 201, a first stage 202, a first objective lens 203, a polarization device 204, a charge-coupled device (CCD) camera 205, a first base 206 and a first bracket 207. Please refer to Figure 2 , the first stage 202 is located in the center of the first bracket 207, the first objective lens 203 is located at the top of the first bracket, the first objective lens 203 has an analyzer, the bottom of the first objective lens 203 has a first mounted objective lens converter, and the first mounted objective lens converter is connected to the first objective lens 203. The CCD camera 205 is located at the top of the first objective lens 203, the polarization device 204 includes a polarizer and an analyzer, the polarizer and the first light source 201 are located on the first base, and the first base 206 is connected to the bottom of the first bracket 207. When the image acquisition device 200 generates a rock slice image, the first light source 201 is started and emits light, and the light path corresponding to the light passes through the polarizer and the analyzer, and is finally projected onto the rock slice. The rock slice is located at the top of the first stage 202, that is, it is placed on the first stage 202. The CCD camera is in a first working state, that is, the CCD camera corresponds to the position of the objective lens, and captures a microscopic image of the corresponding rock slice as the rock slice image.

[0132] In the embodiment of the present application, a communication connection is established between the image acquisition device and the computer device. Based on the communication connection, the image acquisition device can send the image to the computer device for storage. After the computer device receives the image sent by the image acquisition device, it processes the image to obtain the features related to the present solution and adapted to the rock identification process, and performs rock identification based on the features adapted to the rock identification process.

[0133] Corresponding to the above rock identification system, Figure 3 A flowchart of a rock identification method provided by an exemplary embodiment of the present application is shown, and the method is described by applying the method to a computer device of a rock identification system. The method includes:

[0134] Step 301, receiving a rock slice image sent by an image acquisition device.

[0135] In one example, before receiving a rock thin section image sent by an image acquisition device, a computer device sends an image acquisition instruction to the image acquisition device. After receiving the image acquisition instruction, the image acquisition device acquires an image in accordance with the image acquisition instruction, and sends the acquired image to the computer device. The computer device screens and processes the image sent by the image acquisition device, and finally obtains a rock thin section image. A rock thin section is a thin section obtained by cutting a rock sample, that is, the components in the rock thin section can represent and characterize the components of the rock sample. Correspondingly, the rock thin section image includes at least one component area, and each component area includes a component. In the embodiment of the present application, the component indicates the component corresponding to the rock. Please refer to Figure 4 Since the rock slice image 400 is an image obtained by microscopically photographing the rock slice, the rock slice image includes a component region 401. Figure 4 In the embodiment of the present application, the number of component areas is 18, and the 18 component areas can represent the same component, or different components, or at least two different components, and there is a situation where the components in the component areas are the same. The embodiment of the present application does not limit the content of the component area. In the embodiment of the present application, the components included in the component area are implemented as mineral components, or as structural components. Optionally, when the content included in the component area is a mineral component, the component area can indicate a class of minerals. When the content included in the component area is a structural component, the component area can indicate a class of particles, or, a class of fillers.

[0136] It should be noted that the number of rock slice images in the embodiment of the present application is 2, which are a single polarization image and an orthogonal polarization image corresponding to the rock slice; or, the number of rock slices in the embodiment of the present application is 1, and the image is an image obtained by combining the single polarization image and the orthogonal polarization image corresponding to the rock slice. The embodiment of the present application does not limit the actual form of the rock slice image, but the rock slice image must reflect the anisotropy of the rock slice under ordinary light and polarized light sources.

[0137] Step 302: Generate geometric features, mineral features, and structural features corresponding to the rock slice based on the rock slice image.

[0138] In the embodiment of the present application, when a rock thin section image is received, the computer device will extract features of the rock thin section image corresponding to the properties of the rock thin section, corresponding to the geometric features, mineral features and structural features of the rock thin section. Optionally, the geometric features are used to indicate the regional division of the rock thin section based on the components, the mineral features are used to indicate the mineral identification results obtained by performing mineral identification on the rock thin section, and the structural features are used to indicate the spatial structure of the rock sample. In the embodiment of the present application, the distribution image of the component area in the rock thin section image can characterize the geometric features, the specific mineral type of the component area in the rock thin section image can characterize the mineral features, and the spatial structural properties reflected by the rock thin section image can characterize the structural features. The embodiment of the present application does not limit the specific contents of the geometric features, mineral features and structural features.

[0139] Step 303, generating identification results of rock slices based on geometric features, mineral features and structural features.

[0140] In an embodiment of the present application, after the computer device determines the rock thin section image representation, that is, the geometric features, mineralogical features and structural features corresponding to the rock thin section, the identification result of the rock thin section is generated based on the above features. The identification result includes a textual description of the characteristics of the rock sample. In one example, the identification result includes the naming of the rock sample, and a textual summary of the geometric features, mineralogical features and structural features. Optionally, the identification result is presented in the form of a list, or the identification result also includes visual image content, which is presented in the form of a chart. The present application does not limit the method for generating the identification structure.

[0141] In summary, the method provided in the embodiment of the present application, after obtaining the rock thin section image, performs feature extraction based on three dimensions of geometric features, mineral features and structural features on the rock thin section image, determines the properties of the rock from the perspective of microscopic composition and macroscopic performance, and finally generates an identification result including a textual description. In the process of identifying the rock, after obtaining the microscopic visualization image corresponding to the rock thin section, the image is subjected to feature extraction in multiple dimensions, and the rock thin section is identified with reference to the features in multiple dimensions, thereby improving the accuracy of rock identification.

[0142] In some embodiments of the present application, there is a correlation between geometric features, mineral features and structural features. In one example, the structural features indicate the specific type of rock, and the geometric features and mineral features can provide guidance for the identification of the specific type of rock; in another example, the mineral features indicate the minerals corresponding to the component region, so before determining the mineral features, the computer device needs to prioritize the determination of the geometric features. Figure 5A flowchart of a method for determining geometric features, mineral features and structural features provided by an exemplary embodiment of the present application is shown. The method can replace Figure 3 Step 302 shown is implemented as steps 501 to 506. The method includes:

[0143] Step 501, generating preliminary rock structural features based on rock thin section images.

[0144] In the embodiment of the present application, the structural features include the structural features of the primary rock type and the structural features of the subdivided rock type. The structural features of the primary rock type are used to classify the rock samples corresponding to the rock slices and classify them into the primary rock types, namely, sedimentary rocks, igneous rocks and metamorphic rocks.

[0145] In one example, a rock thin section image is input into a preliminary rock type selection model, and a preliminary rock type result is output, which indicates the preliminary rock type structural characteristics corresponding to the rock thin section. The preliminary rock type structural selection model is a model built based on a preliminary rock type image sample set. After the rock thin section image is input into the preliminary rock type selection model, the preliminary rock type selection model will compare the rock thin section image with the sample images stored in the sample set, and then determine the preliminary rock type structural characteristics.

[0146] Step 502: Determine the primary rock type corresponding to the rock slice based on the primary rock structural characteristics.

[0147] In the embodiment of the present application, the preliminary rock type results include sedimentary rock, igneous rock and metamorphic rock, that is, the preliminary rock types are three types of rock obtained according to the genesis. Corresponding to different preliminary rock type structural features, the preliminary rock types can be obtained. In one example, the preliminary rock type structural features are implemented as similarities with the corresponding preliminary rock types.

[0148] Step 503 , according to the initially selected rock type, the rock thin section image is divided based on the positions of the component images to obtain a rock thin section segmentation image, wherein the rock thin section segmentation image includes at least two segmentation areas.

[0149] Optionally, after determining the preliminary rock type, the computer device can further determine the component area content corresponding to the preliminary rock type, and the common shape of the component area. In the embodiment of the present application, the computer device segments the rock thin section image based on the component area content and the common shape of the component area to obtain a rock thin section segmentation image, and the rock thin section segmentation image includes at least two segmentation areas. The segmentation area is used to indicate the segmentation of the rock thin section image based on the component area it has.

[0150] Step 504: determine geometric features corresponding to the rock slice based on the segmented region.

[0151] After the segmented area is determined, the geometric features corresponding to the rock slice can be determined based on the segmented area. In this process, the computer device regards the segmented area as a component area and identifies the geometric features. In this process, after the corresponding processing by the computer device, the geometric features can be used to characterize at least one of the area of ​​the component area, the maximum diameter of the component area, and the shape of the component area.

[0152] Step 505, determining mineral features corresponding to the rock slice based on the segmented region.

[0153] After determining the geometric features corresponding to the rock slice, the mineral features corresponding to the rock slice can be further determined based on the primary rock structure features. In the embodiment of the present application, the mineral features are used to indicate the mineral types of the components in each segmented area.

[0154] It should be noted that, in some embodiments of the present application, the geometric features corresponding to the rock slice are determined first, and then the mineral features corresponding to the rock slice are determined; in some other embodiments of the present application, the mineral features corresponding to the rock slice are determined first, and then the geometric features corresponding to the rock slice are determined; in some other embodiments of the present application, both the geometric features and the mineral features corresponding to the rock slice are determined simultaneously. The present application does not limit the order of determining the geometric features and the mineral features.

[0155] Step 506, determining the structural features of the subdivided rock types based on the geometric features and the mineral features.

[0156] In an embodiment of the present application, after determining the geometric characteristics and mineral characteristics, the subdivided rock structure characteristics can be further determined. The subdivided rock structure characteristics are used to indicate the subdivided rock type corresponding to the rock sample, that is, based on the genesis, combined with the rock component area and the mineral composition of the component area, the type is further determined.

[0157] Optionally, the geometric features and mineral features are input into the subdivided rock type selection model, and the subdivided rock type results are output, the subdivided rock type results indicate the subdivided rock type structural features corresponding to the rock thin section, and the subdivided rock type selection model is a model constructed based on the geometric-mineral feature interactive sample set, and the geometric-mineral feature interactive sample set indicates the corresponding relationship between the combination of geometric features and mineral features and the subdivided rock type structural features. In one example, the preliminary selected rock type structural features corresponding to the rock sample indicate that the rock sample is a sedimentary rock, and through the subdivided rock type selection model, according to the properties of the segmented regions in the rock sample thin section and the types of the segmented regions, the subdivided rock type structural features corresponding to the rock sample are obtained, indicating that the rock sample is a clastic rock in the sedimentary rock.

[0158] Optionally, both the preliminary rock types and the detailed rock types are included in the identification results.

[0159] To sum up, the method provided in the embodiment of the present application corresponds to the actual application content of structural features for rock identification, classifies structural features into preliminary rock type structural features and detailed rock type structural features, and after determining the preliminary rock type structural features, further determines the geometric features and structural features corresponding to the rock thin section image, and then determines the detailed rock type structural features, so that in the process of rock identification, the process of determining geometric features and structural features is more targeted, and the accuracy of rock identification is further improved.

[0160] Next, corresponding Figure 5 The detailed determination methods of the geometric features and mineral features involved in the illustrated embodiment are described one by one.

[0161] As described in the above embodiments, the geometric features are used to indicate the regional division of the rock slice based on the components. In this application, the geometric features are obtained through a neural network model based on machine learning. Figure 6 A schematic diagram of a method for segmenting a rock slice image provided by an exemplary embodiment of the present application is shown. The method can replace Figure 5 Step 503 to step 504 in the embodiment shown. Take the method applied to a computer device as an example for explanation, please refer to Figure 6 , the method comprising:

[0162] Step 601: Acquire a sample slice image, where the sample slice image is annotated with a sample segmentation result.

[0163] The embodiment of the present application is to explain the construction and application process of the rock thin section image segmentation model, which is a model for dividing component regions. In the embodiment of the present application, after selecting the basic network used by the rock thin section image segmentation model, the machine learning process is implemented by sample training. Step 601 is to obtain a sample thin section image, that is, the process of constructing a sample set.

[0164] In the embodiment of the present application, the sample thin section image is an image selected from the content of the corresponding rock thin section image. In one example, the sample thin section image is a rock thin section image corresponding to the primary rock type of the rock thin section; in another example, the sample thin section image is a rock thin section image corresponding to the subdivided rock type of the rock thin section. The present application does not limit the specific rock type corresponding to the sample thin section image.

[0165] Optionally, the sample slice image is annotated with a sample segmentation result, which is a division method of the sample slice image based on the segmentation area of ​​the component area determined after judgment by other methods.

[0166] Step 602: Input the sample thin section image into the rock thin section image segmentation model, and output the predicted segmentation result.

[0167] In the embodiment of the present application, the rock slice image segmentation model is a Mask-RCNN network model based on machine learning. In one example, please refer to Figure 7 The rock slice image segmentation model 700 includes a feature image generation network 701, a region proposal network 702 and a result generation network 703. The three sub-networks are interconnected in the rock slice image segmentation model. The feature image generation network 701 includes a feature extraction sub-network 7011 and a feature enhancement sub-network 7012, and the result generation network 703 includes a result region classification branch network 7031 and a frame regression branch network 7032.

[0168] When obtaining the segmentation result through the rock thin section image segmentation model. After the rock thin section image enters the rock thin section image segmentation model, it will pass through the feature image generation network. In the feature extraction subnetwork, after the surface feature extraction of the feature extraction subnetwork and the feature enhancement of the feature enhancement subnetwork, the feature image corresponding to the rock thin section image can be obtained. Subsequently, the feature image is input into the region proposal network, and the preliminary segmentation feature image corresponding to the thin section to be tested is output. After convenient refinement processing, the preliminary segmentation feature image is input into the region classification branch network, and the preliminary segmentation feature image is classified by the region classification branch network, and the preliminary segmentation feature image is corrected by the bounding box regression branch network, that is, the segmentation result is finally output.

[0169] In the embodiment of the present application, for the constructed rock slice image segmentation model, a predicted segmentation result is obtained by inputting a sample slice image, and the predicted segmentation result can be output in the form of a sample rock slice with a segmentation line superimposed and displayed.

[0170] Step 603: compare the predicted segmentation result with the sample segmentation result to obtain a segmentation difference.

[0171] This process is to compare the predicted segmentation result with the sample segmentation result to determine the difference between the two. The reason for the difference between the predicted segmentation result and the sample segmentation result is that the parameter adjustment of the rock thin section image segmentation model has not been completed. In one example, the segmentation difference is implemented as the length difference of the segmentation line in the rock thin section segmentation image; in another example, the segmentation difference is implemented as the shape difference of the segmentation line in the rock thin section segmentation image. This application does not limit the specific implementation form of the segmentation difference.

[0172] Step 604: adjusting the rock thin section image segmentation model based on the segmentation difference.

[0173] In an embodiment of the present application, it is determined whether the training of the rock thin section image segmentation model is completed based on the difference in the segmentation difference. If the training of the rock thin section image segmentation model is not completed, other predicted segmentation results are output by continuing to input the sample thin section image into the rock thin section image segmentation model, and a difference comparison is performed to execute the model training process. Optionally, in other embodiments of the present application, it is also possible to determine whether the training of the rock thin section image segmentation model is completed based on the number of times the rock thin section image segmentation model has been trained. In an embodiment of the present application, in response to the difference threshold between the predicted segmentation result and the sample segmentation result indicated by the segmentation difference, it is determined that the training of the rock thin section image segmentation model is completed; or, in response to the number of sample thin section images input to the rock thin section image segmentation model in a segmentation training cycle reaching the segmentation number threshold, it is determined that the training of the rock thin section image segmentation model is completed.

[0174] It should be noted that the various "thresholds" appearing in this application can be implemented as thresholds stored in a computer program, or as thresholds manually input when a computer device executes corresponding steps. This application does not limit the actual method of obtaining the thresholds.

[0175] After the training of the rock slice image segmentation model is completed, the rock slice image to be tested is inputted to segment the rock slice image so as to extract geometric features.

[0176] Step 605, in response to the completion of the rock thin section image segmentation model training, the rock thin section image is input into the rock thin section image segmentation model, and the rock thin section segmentation image is output.

[0177] The process corresponding to step 605 is the process of generating a rock slice segmentation image corresponding to the rock slice. The rock slice segmentation image includes at least two segmentation regions. As mentioned above, the segmentation region indicates the segmentation situation determined based on the component region. When the rock slice image segmentation model is trained, the segmentation region can directly represent the component region. Figure 8As shown in FIG. 8 , in the rock slice segmentation image 800 obtained after the rock slice image segmentation model, multiple component regions 802 are divided by segmentation lines 801. Figure 8 For example, the area completely surrounded by the dividing line 801 is the component area 802.

[0178] After obtaining the rock slice segmentation image, a segmentation result corresponding to the rock slice image can be generated. The segmentation result includes at least one of the number of component regions corresponding to the rock slice image, the distribution density of the component regions corresponding to the rock slice image, and the size of the component regions corresponding to the rock slice image. In one example, the segmentation result can directly indicate the geometric features corresponding to the rock slice, or the segmentation result can be processed to generate the geometric features corresponding to the rock slice.

[0179] In summary, the method provided in the embodiment of the present application, in the process of identifying rock thin section images, uses a rock thin section image segmentation model based on a Mask-RCNN network to segment rock thin section images to obtain geometric features corresponding to the rock thin sections. On the basis of using the image segmentation method in a computer instead of the manual segmentation method, a Mask-RCNN network model based on machine learning for image processing is introduced, which provides better division guidance for component areas in rock thin section images and improves the efficiency of obtaining geometric features.

[0180] Optionally, in this application, Figure 6 The embodiment shown is an embodiment of a computer device acquiring geometric features. In other embodiments of the present application, the geometric features also include other acquisition methods. The present application does not limit the specific acquisition method of the geometric features.

[0181] In the embodiment of the present application, since the geometric features indicate the division of component areas, and each component area corresponds to different components, after determining the geometric features, the mineral features are determined, that is, the mineral distribution in the rock slice is determined.

[0182] In one embodiment of the present application, the determination of mineral characteristics is performed by a neural network model based on machine learning. Fig. 9 A flowchart of a method for identifying mineral features provided by an exemplary embodiment of the present application is shown. The method can be implemented as follows: Figure 5 Step 505 in the embodiment shown is described by taking the method applied to a computer device as an example, the method includes:

[0183] Step 901, determining a component image based on the segmented region, wherein the component image includes a component region.

[0184] In the examples of this application, please refer to Fig.10 The segmentation region 1001 is a segmentation result determined based on the rock thin section segmentation image 1000, and the segmentation region 1001 can be used to determine the component image 1010, and the component image 1010 includes a component region. Since the rock thin section image segmentation model has been trained, the component region corresponds to the component region included in the rock thin section, and directly represents the components in the rock thin section.

[0185] Step 902: Acquire a sample component image, where the sample component image is annotated with sample component type features.

[0186] In an embodiment of the present application, the component type characteristics corresponding to the component region are determined by the component type recognition model, that is, the mineral composition corresponding to the component region is determined. Optionally, the component type recognition model is constructed in a similar manner to the rock thin section image segmentation model. Both are sample training based on machine learning. In an embodiment of the present application, the sample component image is selected from a sample component image data set, and each sample component image includes and only includes one component region, and the sample component region corresponds to the sample component type characteristics.

[0187] Step 903: Input the sample component image into the component type recognition model, and output the predicted component type feature.

[0188] In the embodiment of the present application, the predicted component type feature is the predicted component type feature obtained by the component type recognition model in the training process to recognize the sample component image. In the embodiment of the present application, the formation neural network that can be used by the component type recognition model includes at least one of the Mobilenet neural network, the Resnet neural network, and the VGG (Visual Geometry Group Network) neural network.

[0189] Step 904 , compare the predicted component type features with the sample component type features to obtain identification differences.

[0190] This process is a process of identifying the difference between the predicted component type feature and the sample component type feature. In one example, the component type feature is represented in numerical form, and different numerical values ​​represent different component types, as well as the probability that the component region belongs to the component type. In this case, the identification difference can indicate the difference between the predicted component type feature and the sample component type feature, or the ratio of the predicted component type feature to the sample component type feature. The embodiment of the present application does not limit the actual output form of the component type feature.

[0191] Step 905, adjusting the component type identification model based on the identification differences.

[0192] This process is to adjust the component type recognition model according to the component type recognition difference until the difference is less than the difference threshold, or the number of training times of the component type recognition model reaches the training number threshold. When the above conditions are met, or one of the above conditions is met, it can be determined that the training of the component type recognition model is completed.

[0193] Step 906, in response to the component type recognition model training being completed, the component image is input into the component type recognition model, and the component type features of the component image are output.

[0194] This process is the process of outputting the component type feature. It should be noted that the embodiment of the present application focuses on the identification of a single component in a component image, so when the component image inevitably includes two component areas, the component type recognition model will uniformly identify the above two component areas, or the component type recognition model will select an image area corresponding to any component area in the above two component areas for separate identification. For a component identification process, the component type recognition model will output and only output one component type feature.

[0195] Step 907, determining the mineral features corresponding to the rock slice based on the component type features corresponding to the component image.

[0196] Optionally, the computer device stores the correspondence between the component type characteristics and the mineral types. When the component type characteristics are output, the computer device can determine the mineral type corresponding to the component area from the correspondence between the component type characteristics and the mineral types.

[0197] After the mineral types of all component areas corresponding to the rock thin section image are determined, the mineral characteristics corresponding to the rock thin section can be obtained.

[0198] In summary, the method provided in the embodiment of the present application, corresponding to a single component region, inputs it into the component type recognition model, performs component recognition, and outputs a component type feature indicating the mineral characteristics of the component region. By inputting a component image with a single component region, the detailed type of component particles is recognized by using a recognition model based on machine learning, thereby improving the recognition efficiency of components in rock slices.

[0199] In other embodiments of the present application, the mineral characteristics are initially determined based on the classification recognition model, and then the preliminary determination results of the classification recognition model are verified by the mineral data acquisition device. In this case, the rock identification system also includes the mineral data acquisition device. Fig.11The structural block diagram of a rock identification system provided by an exemplary embodiment of the present application is shown, and the rock identification system includes an image acquisition device 1101, a computer device 1102 and a mineral data acquisition device 1103, wherein the image acquisition device 1101 and the mineral data acquisition device 1103 are connected to the computer device 1102 respectively. Correspondingly, Fig.12 A schematic diagram of a mineral data acquisition device provided by an exemplary embodiment of the present application is shown, the mineral data acquisition device includes a second light source 1201, a second stage 1202, a second objective lens 1203, a photoelectric signal converter 1204, a second base 1205 and a second bracket 1206; the second light source 1201 corresponds to a second optical path 1207, and the end point of the second optical path 1207 is projected onto the second stage 1202; the second stage 1202 is located in the center of the second bracket 1207, the second objective lens 1203 is located at the top of the second bracket 1207, and the bottom of the second objective lens 1203 is provided with a second mounted objective lens converter, and the second mounted objective lens converter is connected to the second objective lens 1203; the photoelectric signal converter 1204 is located on the top of the second objective lens 1203; when the mineral data acquisition device 1200 generates mineral data, the rock slice is located on the top of the second stage 1202, and the photoelectric signal converter 1204 is in the second working state to acquire spectral data. In the embodiment of the present application, the second working state is the state when the optoelectronic signal converter 1204 performs data acquisition.

[0200] It should be noted that the mineral data acquisition device and the image acquisition device can be implemented as the same device, that is, the mineral data acquisition device and the image acquisition device share the first stage, the first objective lens, the first base and the first bracket. The CCD camera and the photoelectric signal converter are both located on the first objective lens. The integrated device of the mineral data acquisition device and the image acquisition device can obtain rock thin section images and spectral data, and send the rock thin section images and spectral data to the computer device.

[0201] Based on the rock identification system described above, Fig.13 A schematic diagram of a flow chart of a method for determining mineral features corresponding to a rock slice provided by an exemplary embodiment of the present application is shown. The method can be implemented as follows Figure 5 In step 505 of the corresponding embodiment, the method is applied to a computer device as an example for explanation. The method includes:

[0202] Step 1301, input the rock thin section segmentation image into the classification recognition model, and output a single crystal component set image and a non-single crystal component set image.

[0203] In the embodiment of the present application, a classification recognition model is used to distinguish component regions belonging to different crystalline types in a rock thin section segmentation image.

[0204] Under normal circumstances, rocks are manifested as a multi-component mixture, in which the components can be divided into single crystalline components and non-single crystalline components according to the macroscopic morphology of the crystals in the component region. Optionally, the non-single crystalline components include polycrystalline components and amorphous components. In one example, the classification and recognition model includes a crystalline component classification model to distinguish the components into single crystalline components and non-single crystalline components, and the classification and recognition model also includes a non-single crystalline component classification model to distinguish the non-single crystalline components into polycrystalline components and amorphous components, and a single crystalline component recognition model, a polycrystalline component recognition model and an amorphous component recognition model are set accordingly to identify the specific mineral types corresponding to the component region.

[0205] In this case, for different rock types, the classification sub-model and the specific type of the recognition sub-model in the classification recognition model can be set to adapt to the different component areas contained in different rock types. Next, the rock type of the rock slice is clastic rock as an example for explanation.

[0206] Corresponding to the rock type of clastic rock, the components indicated by the component area in the clastic rock include clastic components and interstitial components. Among them, clastics are a component of sedimentary rocks or sediments, and are the product of mechanical weathering of parent rocks; interstitial materials are connecting materials located between individual clastics. Subsequently, when further subdivided, the clastic components are divided into terrigenous clastic components and non-terrigenous clastic components. Among them, terrigenous clastics refer to clastic materials formed by physical weathering or mechanical destruction of parent rocks in terrigenous areas, mainly including quartz, feldspar, rock fragments, etc., and non-terrigenous clastics refer to clastic materials formed by physical weathering or mechanical destruction of parent rocks in non-terrigenous areas, mainly including endogenous fragments, volcanic fragments, etc. At the same time, the interstitial components are divided into cement components and matrix components. Among them, the matrix indicates the fine mechanical admixtures filled between the clastic particles. In the embodiment of the present application, the matrix can be realized as at least one of silt, clay material, subdivision, and carbonate mortar. The cement indicates the chemical precipitation position in the interstitial material other than the matrix. Secondly, the cement component is further divided into siliceous calcareous component and mud component according to its material composition. And the terrigenous clastic component is divided into quartz feldspar component and rock fragment component according to its clastic morphology. That is, for the clastic rock type, the classification and recognition model needs to have the ability to identify the quartz feldspar clastic area, siliceous calcareous cement area, non-terrigenous clastic area, rock fragment area, matrix filling area and mud cement area. Fig.14 A schematic diagram of a rock slice corresponding to a clastic rock provided by an exemplary embodiment of the present application is shown. The rock slice of the clastic rock includes various components. Please refer to Fig.14 The rock slice 1400 of the clastic rock includes a quartz clastic region 1401 , a siliceous calcareous cement region 1402 , a non-terrigenous clastic region 1403 , a rock debris region 1404 , a matrix filling region 1405 and a mud cement region 1406 .

[0207] Please refer to Fig.15 , corresponding to the case where the rock type is clastic rock, the rock identification model 1500 includes a classification sub-model group and an identification sub-model group. The classification sub-model group includes a component classification sub-model 1511, a debris classification sub-model 1512, a terrigenous debris classification sub-model 1513, a filling material classification sub-model 1514 and a cementation classification sub-model 1515. The identification sub-model group includes a non-terrigenous debris identification sub-model 1521, a rock debris identification sub-model 1522, a quartz feldspar debris identification sub-model 1523, a siliceous calcareous cementation identification sub-model 1524, a mudstone cementation identification sub-model 1525 and a matrix filling material identification sub-model 1526. Among them, the component classification submodel 1511 is connected to the debris classification submodel 1512 and the filling material classification submodel 1514, the debris classification submodel 1512 is connected to the non-terrigenous debris identification submodel 1521 and the terrigenous debris classification submodel 1513, the terrigenous debris classification submodel 1513 is connected to the rock debris identification submodel 1522 and the quartz feldspar debris identification submodel 1523, the filling material classification submodel 1514 is connected to the cementation classification submodel 1515 and the matrix filling material identification submodel 1526, and the cementation classification submodel 1515 is connected to the siliceous and calcareous cementation identification submodel 1524 and the muddy cementation identification submodel 1525.

[0208] In this example, after the rock thin section segmentation image is input into the rock recognition model, each classification sub-model will generate a corresponding classification sub-image by dividing and reorganizing the rock thin section segmentation image to determine the corresponding area in the rock thin section segmentation image. Each recognition sub-model corresponds to each sub-image to determine the specific component type. In this process, it is also implemented as follows: the clastic rock thin section is input into the component classification sub-model, and a clastic sub-image and a filling material sub-image are output; the clastic sub-image is input into the clastic classification sub-model, and a non-terrestrial clastic sub-image and a terrestrial clastic sub-image are output; the terrestrial clastic sub-image is input into the terrestrial clastic classification sub-model, and a rock clastic sub-image and a quartz feldspar clastic sub-image are output; the filling material sub-image is input into the filling material classification sub-model, and an impurity filling material sub-image is output. and cementation sub-images; inputting the cementation sub-images into the cementation classification sub-model, and outputting the siliceous calcareous cementation sub-images and the argillaceous cementation sub-images; inputting the non-terrigenous debris sub-images into the non-terrigenous debris identification sub-model, and outputting the non-terrigenous debris identification results, and determining the non-terrigenous debris area based on the non-terrigenous debris identification results; inputting the rock debris sub-images into the rock debris identification sub-model, and outputting the rock debris identification results, and determining the rock debris area based on the rock debris identification results; The quartz feldspar sub-image is input into the quartz feldspar fragment identification sub-model, and the quartz feldspar fragment identification result is output, and the quartz feldspar fragment area is determined based on the quartz feldspar fragment identification result; the matrix interstitial material sub-image is input into the matrix interstitial material identification sub-model, and the matrix interstitial material identification result is output, and the matrix interstitial material area is determined based on the matrix interstitial material identification result; the siliceous calcareous cement sub-image is input into the siliceous calcareous cement identification sub-model, and the siliceous calcareous cement identification result is output. The results are obtained, and the siliceous and calcareous cement areas are determined based on the siliceous and calcareous cement identification results; the mud cement sub-image is input into the mud cement identification sub-model, and the mud cement identification results are output, and the mud cement area is determined based on the mud cement identification results; the single crystalline component area is determined based on the quartz feldspar debris area and the siliceous and calcareous cement area; the non-single crystalline component area is determined based on the non-terrigenous debris area, rock debris area, mud cement area and miscellaneous matrix filling area.

[0209] Corresponding to the actual situation of clastic rocks, the quartz-feldspar clastic area and the siliceous-calcium cement area belong to the single-crystalline component area; the non-terrigenous clastic area, the rock debris area, the mud cement area and the matrix filling area belong to the non-single-crystalline component area.

[0210] That is, in the embodiment of the present application, the single-crystalline component set image includes at least one single-crystalline component region, and a single-crystalline component identification result corresponding to the single-crystalline component region, and the non-single-crystalline component region includes at least one non-single-crystalline component region, and a non-single-crystalline component identification result corresponding to the non-single-crystalline component region.

[0211] It should be noted that the classification and recognition model in the embodiments of the present application, or the various sub-models that constitute the classification and recognition model, are all neural network models based on deep learning.

[0212] Step 1302: Send a scanning instruction to a mineral data acquisition device based on the single crystal component set image.

[0213] Since the image features corresponding to the single-crystalline component area are not obvious, it is necessary to collect scanning data through a mineral data acquisition device for the single-crystalline component set image. Optionally, the scanning data is the spectral data collected by the mineral data acquisition device. In an embodiment of the present application, the mineral data acquisition device can also be implemented as at least one device including but not limited to molecular spectra such as infrared spectra and Raman spectra, X-ray fluorescence spectra (X-ray Fluorescence, XRF), and X-ray energy spectrum analysis (Energy Dispersive X-ray Spectroscopy, EDS). The computer device sends a scanning instruction to the mineral data acquisition device, and the scanning instruction indicates to the mineral data acquisition device the position of the component area corresponding to the single-crystalline component set image.

[0214] Optionally, since impurities are inevitably introduced in the process of making rock slices, and the minerals corresponding to the single crystalline components are subject to alteration (referring to the phenomenon that the composition of minerals is changed by external influences to generate new minerals, in one example, feldspar is altered into clay minerals), twinning (referring to the regular intergrowth of two or more crystals of the same kind, and the connection between the crystals may interfere with data acquisition, in one example, plagioclase includes polysynthetic twinning), inclusion (the phenomenon that a mineral is composed of one or more phases of matter and has a closed system with a phase boundary with the main crystal mineral, in one example, quartz has rutile inclusions), and crack (referring to the phenomenon that a mineral is broken due to external stress, in one example, quartz has cracks), which affect the data acquisition results of the mineral data acquisition device, before generating a scanning instruction, the single crystalline component set image is first input into the regional screening model to obtain the regional screening result to determine the portion of the single crystalline component set image that is suitable for scanning, that is, the regional screening result indicates the clean area in the single crystalline component set image. In this case, the computer device generates a scanning instruction based on the regional screening result and sends the scanning instruction to the mineral data collection device.

[0215] Step 1303, receiving the scanning data fed back by the mineral data acquisition device based on the scanning instruction.

[0216] After the mineral data acquisition device receives the scanning instruction. In one example, the mineral data acquisition device moves the objective lens based on the scanning instruction, or moves the stage, so that the area corresponding to the rock slice placed on the stage is opposite to the objective lens, and collects the scanning data through the illumination of the light source and the signal conversion of the photoelectric signal converter, and feeds it back to the computer device. In the embodiment of the present application, the scanning data can be realized as spectral data.

[0217] Step 1304 , verifying the single crystal component identification result based on the scanning data to obtain a verification result.

[0218] The scanning data can be used to determine the component type in each component region in the single-crystalline component set image from the dimension of chemical composition or the dimension of crystal structure, wherein each single-crystalline component region in the single-crystalline component set image corresponds to scanning data indicating its chemical composition, so the scanning data can be used to determine the single-crystalline identification result of each single-crystalline component in the single-crystalline component set.

[0219] Step 1305, based on the non-single crystalline component identification result, the single crystalline component identification result and the verification result, determine the mineral characteristics corresponding to the rock thin section.

[0220] When the verification result is the same as the single crystal component identification result, determining that the verification result indicates that the verification is passed, and determining that the single crystal component identification result is unchanged;

[0221] When the verification result is different from the single crystal component identification result, it is determined that the verification result indicates that the verification failed. In the embodiment of the present application, the single crystal component identification result is determined based on the identification result generation rule. The identification result generation rule is a determination rule generated based on the influence of the mineral data acquisition device on the single crystal component identification result. The reason for the generation of this rule is that the scanning data generation result of the mineral data acquisition device is greatly affected by the specific type of the mineral data acquisition device.

[0222] After determining the mineral type corresponding to each component area, the mineral types corresponding to each component area can be summarized to determine the mineral characteristics corresponding to the rock thin section.

[0223] In summary, the method provided in the embodiment of the present application, before determining the components of the rock thin section segmentation image by surface scanning, uses a classification recognition model to preliminarily identify the rock thin section segmentation image based on the crystal state, and after selecting the area corresponding to the single crystal component area, the component recognition result is secondary verified, thereby improving the accuracy of the mineral characteristics finally generated.

[0224] In other embodiments of the present application, after the spectral data corresponding to the component region is collected by the mineral data collection device, the mineral characteristics are determined by searching and comparing in the spectral data database. Fig.16 A schematic flow chart of a method for determining mineral features corresponding to a component region provided by an exemplary embodiment of the present application is shown. The method can be implemented as follows: Figure 5 The corresponding step 505 in the embodiment is described by taking the method applied to a computer device as an example, the method includes:

[0225] Step 1601, determining a component image based on the segmented region, wherein the component image includes a component region.

[0226] This process is the process of segmenting and determining the component areas according to the segmentation areas in the rock thin section segmentation image.

[0227] Step 1602: Send a spectral data acquisition instruction to the mineral data acquisition device based on the component image.

[0228] For each independent component region, after selecting the spectral data collection point in the component region, a spectral data acquisition instruction is sent to the mineral data collection device.

[0229] Step 1603, receiving the spectral data sent by the mineral data acquisition device based on the spectral data acquisition instruction.

[0230] In the embodiment of the present application, the mineral data acquisition device obtains the corresponding spectral data and feeds back to the computer device.

[0231] Step 1604, based on the spectral data, determine the preliminary mineral type corresponding to the component region in the mineral spectral data database.

[0232] In the embodiment of the present application, a mineral spectral data database is stored in the computer device, and the mineral spectral data database stores the correspondence between the spectral data and the types of preliminary selected minerals.

[0233] In the embodiment of the present application, the preliminary selected mineral category is a category obtained by dividing based on the representative properties of the mineral. Optionally, the computer device classifies the spectral data sample set in the spectral data database, and determines the rarity corresponding to the mineral according to the number of classifications of different minerals. Optionally, under the division of a preliminary selected mineral category, at least two subdivided mineral categories are also included.

[0234] In the embodiment of the present application, the process of obtaining the primary mineral type is as follows:

[0235] First, based on the spectral data, in the mineral spectral data database, the primary category of the preliminary mineral type corresponding to the component area is determined. The primary category of the mineral type is a category divided based on the commonness of the mineral, and the primary category includes common mineral categories and uncommon mineral categories. In the embodiment of the present application, the primary category of the mineral is determined by comparing with the primary category frequency threshold. Optionally, when the proportion of the spectral data corresponding to the mineral in the spectral data in the component area is greater than the primary category frequency threshold, that is, the frequency of occurrence of the spectral data corresponding to the mineral is greater than the primary category frequency threshold, the primary category of the mineral is determined to be a common mineral, and when the frequency of occurrence of the spectral data corresponding to the mineral is less than or equal to the primary category frequency threshold, the mineral is determined to be an uncommon mineral. In the embodiment of the present application, based on the number of mineral samples in the spectral data database, the common category and the uncommon mineral category are determined, that is, when the number of mineral samples in the spectral data database is less than the sample number threshold, the primary category of the mineral type is determined to be an uncommon mineral category, and when the number of mineral samples in the spectral data database is greater than or equal to the sample number threshold, the primary category of the mineral type is determined to be an uncommon mineral category. Under this classification, the preliminary selected mineral types corresponding to the common mineral categories include mineral family types and common common mineral types, and the preliminary selected mineral types corresponding to the uncommon mineral categories include inclusion mineral types, highly sensitive mineral types, and alteration mineral types. Among them, the common common mineral types indicate the preliminary selected mineral types whose occurrence frequency is greater than a certain threshold of the frequency of the preliminary selected mineral types in the mineral spectrum data database; the inclusion mineral types are the mineral types corresponding to the minerals that appear as inclusions after the minerals grow; the highly sensitive mineral types indicate the mineral types corresponding to the minerals that affect the acquisition accuracy of the spectrum data acquisition equipment due to chemical properties during the acquisition of spectrum data, and make the spectrum data acquisition equipment pay attention to itself; the alteration mineral types are the mineral types corresponding to the minerals that produce identification phenomena during the mineral formation process.

[0236] It should be noted that the "primary category frequency threshold" and "mineral type frequency threshold" that appear in the above process can be thresholds pre-stored in the computer device, or can be thresholds determined by signals received when the computer device executes corresponding steps. This application does not limit the generation method of each threshold.

[0237] In the embodiment of the present application, the mineral spectrum data database also has multiple sub-databases. Fig.17 The mineral spectrum data database 1700 includes a common mineral category sub-database 1701 and an uncommon mineral category sub-database 1702. The common mineral category sub-database 1701 also includes a common mineral type sub-database 1711 and a mineral family database 1712. The uncommon mineral category sub-database 1702 also includes an inclusion mineral type database 1721, a highly sensitive mineral type database 1722, and an alteration mineral type database 1723. In addition, the mineral spectrum data database 1700 also includes an impurity category sub-database 1703 corresponding to impurities. Since there are many types of impurities, taking the rock as clastic rock as an example, the impurity component contained in the clastic rock is silicon carbide, and the impurity category sub-database 1703 is an identification database corresponding to silicon carbide.

[0238] Step 1605, determining the mineral type verification rule corresponding to the initially selected mineral type.

[0239] In an embodiment of the present application, corresponding to the above-mentioned preliminary selected mineral types, the mineral type verification rules include classification verification rules, direct verification rules and point selection re-verification rules. In response to the preliminary selected mineral type being a mineral family type, the mineral type verification rule is determined to be a classification verification rule; in response to the preliminary selected mineral type being an inclusion mineral type, or, a highly sensitive mineral type, or, an alteration mineral type, the mineral type verification rule is determined to be a point selection re-verification rule; in response to the preliminary selected mineral type being an ordinary common mineral type, or, the primary category corresponding to the preliminary selected mineral type is an uncommon mineral category, and the preliminary selected mineral type is not any one of the inclusion mineral type, the highly sensitive mineral type and the alteration mineral type, the mineral type verification rule is determined to be a direct verification rule. In an embodiment of the present application, the point selection re-verification rule indicates a rule for a computer device to reacquire scan data from a spectral data acquisition device; the classification verification rule indicates a rule for selecting a sub-database corresponding to the preliminary selected mineral type and performing a secondary verification; the direct verification rule indicates a rule for directly using the result of the preliminary selected mineral type as the verified mineral type result.

[0240] Among them, when the initially selected mineral type is an inclusion mineral type, or a highly sensitive mineral type, or an altered mineral type, the corresponding verification rule is the point selection re-verification rule, that is, due to the special properties of the crystals in the component area, it is necessary to re-select points in the component area. The number of re-selections corresponds to the number of verifications pre-stored in the computer device.

[0241] In the embodiment of the present application, the classification verification rule is a rule for further subdividing the mineral type based on the sub-database in the mineral spectrum data database. Optionally, when the initially selected mineral type is a mineral family type, the mineral type of the component area is subdivided based on the sub-database in the mineral family database, and a verification result is obtained. That is, the classification verification rule is applicable to minerals of the mineral family type.

[0242] In the embodiment of the present application, when the primary selected mineral type is a common mineral type, or the primary category of the corresponding mineral type is an uncommon mineral type, and the subdivided mineral type is not any of the inclusion mineral type, the highly sensitive mineral type, and the altered mineral type, a direct verification rule is determined. The direct verification rule is a rule for determining the primary selected mineral type as a subdivided mineral category.

[0243] Step 1606, determining the mineral type corresponding to the component region based on the mineral type verification rule and the preliminary selected mineral type.

[0244] This process is the process of determining the mineral types corresponding to the component areas.

[0245] Step 1607, based on the mineral types corresponding to the component images, determine the mineral features corresponding to the rock slices.

[0246] This process is the process of generating mineral characteristics corresponding to rock thin sections based on the mineral types corresponding to each component area.

[0247] In summary, the method provided in the embodiment of the present application, after obtaining the spectral data corresponding to the component region, performs a preliminary determination of the mineral type corresponding to the component region in the corresponding database, and determines the verification rules accordingly, and finally determines the mineral type corresponding to the component region on the basis of the verification rules. In the process of determining the mineral type corresponding to the component region, on the basis of determining its preliminary selected type based on the database, corresponding to each preliminary selected type, the relevant verification rules are determined, and the subdivided mineral types are further obtained, thereby improving the accuracy of mineral type identification, and thus improving the efficiency and accuracy of determining the mineral characteristics of rock thin sections.

[0248] It should be noted that Fig.16 The embodiment shown, Fig.13 The embodiment shown and Fig. 9The embodiment shown is a parallel identification method of three mineral features. In the process of rock identification, any superposition and combination can be performed to achieve the effect of identifying and verifying the mineral features, and further improve the accuracy of determining the mineral features of rock slices. That is, Fig. 9 The embodiment shown, Fig.13 The embodiment shown and Figure 6 The embodiments shown are three embodiments for obtaining mineral characteristics during rock identification.

[0249] It should be noted that, in some embodiments of the present application, the mineral features and geometric features corresponding to the rock slices can be determined simultaneously. Fig.18 FIG. 1 shows a schematic diagram of a process for determining geometric features and mineral features provided by an exemplary embodiment of the present application. Please refer to FIG. Fig.18 , the method comprising:

[0250] Step 1801, obtaining a sample slice image.

[0251] In the embodiment of the present application, the rock slice image recognition model is a neural network model based on machine learning. The neural network model can be composed of Figure 6 The rock slice image segmentation model in the embodiment shown and Fig. 9 The component type recognition model in the illustrated embodiment is obtained by combining, or the neural network model can be constructed according to the parameters of the rock thin section image segmentation model and the component type recognition model. The sample corresponding to the training of the rock thin section image recognition model is a sample thin section image, and the sample thin section image is annotated with the sample segmentation result and the sample component type feature.

[0252] Step 1802: Input the sample thin section image into the rock thin section image segmentation and recognition model, and output the predicted segmentation area and predicted component type features.

[0253] This process is the process of using the rock thin section image that has not yet been trained to predict the sample thin section image by the model to obtain the predicted segmentation area and the predicted component type characteristics.

[0254] Step 1803, compare the predicted segmentation region with the sample segmentation region to obtain the sample segmentation difference, and compare the predicted component type feature with the sample component type feature to obtain the sample component difference.

[0255] In the embodiment of the present application, since the predicted segmentation region and the predicted component type features are output simultaneously, the sample segmentation difference and the sample component difference are determined simultaneously.

[0256] Step 1804: adjusting the rock thin section image segmentation and recognition model based on the sample segmentation differences and sample component differences.

[0257] This process is the process of adjusting the rock thin section image segmentation and recognition model.

[0258] Step 1805: input the rock thin section image into the rock thin section image segmentation and recognition model, and output at least two segmented regions corresponding to the rock thin section image and component type features corresponding to the segmented regions.

[0259] This process is the process of inputting the rock thin section image to be tested into the rock thin section image segmentation and recognition model, and outputting the segmented area and the component type characteristics corresponding to the area.

[0260] Step 1806, combining the segmented area and component type characteristics, determine the geometric characteristics and mineral characteristics corresponding to the rock slice.

[0261] In the embodiment of the present application, the segmented areas correspond to the geometric features, and the component type features correspond to the mineral features.

[0262] In summary, the method provided in the embodiment of the present application corresponds to the sample thin section image as the input of the rock thin section image segmentation and recognition model, and on the premise that the model training is completed, the geometric features and mineral features of the rock thin section image are synchronously extracted to further improve the efficiency of feature extraction in the rock identification process.

[0263] Optionally, in this application, Fig.18 The embodiment shown is an embodiment in which a computer device simultaneously obtains geometric characteristics and mineralogical characteristics.

[0264] After the mineralogical characteristics and geometric characteristics are determined in the manner described in the above embodiments, in the embodiment of the present application, the identification results of the corresponding rock slices are generated through the geometric characteristics, mineralogical characteristics and structural characteristics, that is, a textual description of the characteristics of the rock sample. Fig.19 A schematic diagram showing a method for generating identification results of a rock slice provided by an exemplary embodiment of the present application is shown. The method can be implemented as follows: Figure 3 Step 303 shown is described by taking the method applied to a computer device as an example. The method includes:

[0265] Step 1901, based on the primary rock type characteristics and the detailed rock type characteristics, determine the type of at least one rock thin section identification result corresponding to the identification result.

[0266] In the embodiment of the present application, the rock thin section identification result is an identification result obtained based on the shape of the component region and the mineral properties of the component region, and has a characterization function for the specific type of the rock thin section. Optionally, the rock thin section identification result includes at least one of a particle size range identification result, a maximum particle size identification result, a particle sorting result, a particle roundness identification result, a mineral eumorphism result, and a particle contact mode identification result.

[0267] Step 1902, generating rock thin section identification results based on geometric features and mineral features.

[0268] Optionally, examples are given of methods for determining the above-mentioned particle size range identification results, maximum particle size identification results, particle sorting results, particle roundness identification results, mineral automorphic degree results and particle contact mode identification results to illustrate the generation process of rock thin section identification results.

[0269] In one example, the rock thin section identification result includes a grain size interval recognition result, which indicates the grain size interval of a component region corresponding to a type of component in the rock thin section. First, a grain size interval component region set corresponding to the grain size interval recognition result is determined based on the mineral characteristics, and the grain size interval component region set includes at least two component regions. Secondly, based on the geometric characteristics corresponding to the component regions in the grain size interval component region set, the grain size corresponding to the component region in the grain size interval component region is determined. Finally, based on the grain size corresponding to the component region, the grain size interval recognition result is determined.

[0270] In one example, the rock slice identification result includes a maximum grain size identification result, which indicates the grain size of a component region with the largest grain size among component regions corresponding to a type of component in the rock slice. In this case, the maximum grain component region corresponding to the maximum grain size identification result is determined based on the mineral characteristics, the maximum grain size corresponding to the component region is determined based on the geometric characteristics, and the maximum grain size identification result is generated based on the maximum grain size.

[0271] In one example, the rock thin section identification result includes a grain sorting result, which indicates the distribution of the particle size corresponding to a type of particle in the rock thin section. In this case, first, a set of sorting component regions corresponding to the grain sorting identification result is determined based on the mineral characteristics, and the sorting component region set includes at least two component regions; second, based on the geometric characteristics corresponding to the component regions in the sorting component region set, a grain size classification correspondence is determined, and the grain size classification correspondence includes at least two levels of grain size levels and the number of component regions in the grain size level; finally, based on the grain size classification correspondence, the grain sorting result is determined.

[0272] In one example, the rock thin section identification result includes a particle roundness recognition result, which indicates the roundness of a component region, and the roundness corresponds to a roundness level. In this case, first, a roundness recognition component region set corresponding to the particle roundness recognition result is determined based on the mineral characteristics, and the roundness recognition component region set includes at least two component regions; second, based on the geometric features corresponding to the component regions in the roundness component recognition region set, a roundness classification correspondence is determined, and the roundness classification correspondence includes at least two roundness levels and the number of component regions in the roundness level; finally, based on the roundness classification correspondence, the particle roundness recognition result is determined.

[0273] In one example, the rock thin section identification result includes the result of mineral eumorphism, which indicates the degree of development of minerals in the component region based on their own crystallization habits. In this case, firstly, a set of eumorphic component regions corresponding to the result of mineral eumorphism is determined based on the mineral characteristics, and the set of eumorphic component regions includes at least two eumorphic component regions; secondly, the shape of the component region is determined based on the geometric features corresponding to the component regions in the set of eumorphic component regions; finally, the result of mineral eumorphism is determined based on the shape of the component region.

[0274] In one example, the rock thin section identification result includes the particle contact mode identification result. The particle contact mode identification result indicates the contact state of two adjacent component regions. In this case, firstly, at least two groups of component region pairs are determined based on the mineral characteristics, and the component region pairs include two adjacent component regions; secondly, based on the geometric characteristics, the intersection and union ratio of the two component regions in the component region pair is determined; finally, based on the intersection and union ratio of at least two component regions, the particle contact mode identification result is determined.

[0275] To summarize the above process, after determining the type of rock thin section identification result, the component area corresponding to the rock thin section identification result in the rock thin section image is determined based on the mineral type, and based on the geometric features, the rock thin section identification result is determined from the shape, form, size and other aspects of the component area.

[0276] Step 1903, based on the subdivided rock characteristics, determine the naming rules corresponding to the rock thin section names.

[0277] In the embodiment of the present application, different rock types have different naming rules, so it is necessary to determine the name of the rock slice based on the characteristics of the detailed rock types.

[0278] Step 1904, generating a rock thin section name based on the naming rules and the rock thin section identification results.

[0279] This process is the process of finally determining the name of the rock thin section. Optionally, the computer device can generate corresponding keywords based on the rock thin section identification results, and finally determine the name corresponding to the rock thin section in combination with the subdivided rock characteristics.

[0280] Optionally, for the identification results of the rock slices, in some embodiments of the present application, a rock sample identification report is also generated. The rock sample identification report includes the name of the rock slice and the identification results of each rock slice. In the embodiment of the present application, the computer device pre-stores at least two rock sample identification reports according to different subdivided rock characteristics.

[0281] In summary, the method provided in the embodiment of the present application, after obtaining the preliminary rock type characteristics and the detailed rock type characteristics, selects the rock thin section identification sub-results corresponding to the rock type, and selects the obtained rock thin section identification sub-results accordingly, so that the content of the finally generated rock thin section identification results corresponds to the type of the rock thin section, thereby further improving the accuracy of the rock thin section identification results.

[0282] As described in the above examples, Fig.19 The corresponding embodiment illustrates the process of obtaining rock thin section identification sub-results through geometric features, mineral features and structural features, and then obtaining the final rock thin section identification results.

[0283] In one example, corresponding to the identification process of clastic rock thin sections, Fig. 20 A schematic diagram of a rock identification method provided by an exemplary embodiment of the present application is shown, the process comprising:

[0284] Step 2001, collecting thin section images of clastic rocks.

[0285] In the embodiment of the present application, the clastic rock thin section sample image includes a single polarization image and an orthogonal polarization image. Optionally, the images of the clastic rock thin section sample are continuously acquired under an image acquisition device, and then spliced ​​to finally obtain a single polarization image and an orthogonal polarization image corresponding to the clastic rock thin section sample.

[0286] Step 2002, determining the primary rock type of the clastic rock thin section image.

[0287] In the embodiment of the present application, the primary rock type of the clastic rock thin section image is determined to be sedimentary rock through a rock classification model.

[0288] Step 2003, determine the subdivided rock types of the clastic rock thin section.

[0289] In the embodiment of the present application, the rock structure classification model is used to determine that the subdivided rock type of the clastic rock slice is clastic rock, and has a sandy structure.

[0290] Step 2004: input the clastic rock thin section image into the clastic rock segmentation model to obtain the segmented clastic rock thin section image.

[0291] This process is to segment the clastic rock thin section image through the rock thin section segmentation model to obtain the segmented image. After segmentation, the computer device can determine the clastic rock geometric features corresponding to the clastic rock thin section image.

[0292] Step 2005: extract component regions based on the segmented clastic rock thin section image, and obtain component type features through a component type recognition model.

[0293] This process is the process of obtaining the mineral characteristics corresponding to the clastic rock thin section image.

[0294] Step 2006, mapping the component type features on the clastic rock thin section image to obtain an intermediate image of the clastic rock thin section sample.

[0295] The process is to annotate the clastic rock thin section image based on the component type characteristics and segmentation image to obtain an intermediate image that can characterize the geometric and mineral characteristics of the clastic rock thin section.

[0296] In other embodiments of the present application, the component type recognition model will directly output the intermediate image of the clastic rock thin section sample.

[0297] Step 2007: Count the areas of corresponding component regions in the intermediate image of the clastic rock thin section sample.

[0298] In the embodiment of the present application, based on the results of mineral type identification for each component area, the relative content of different components is determined by counting the area of ​​the component area. In one example, the mineral types of the components are divided, and the proportion of siliceous cement is 2.0%, the proportion of calcareous cement is 10.5%, the proportion of miscellaneous interstitial materials is 1%, the proportion of rock fragments is 13.84%, the proportion of feldspar fragments is 20.76%, and the proportion of quartz fragments is 51.9%. Since quartz fragments, feldspar fragments and rock fragments are all terrigenous debris, siliceous, calcareous cements, and miscellaneous interstitial materials are all interstitial materials, it can be further determined that the proportion of terrigenous debris is 86.5%, and the proportion of interstitial materials is 13.5%. It can be determined that the clastic rock thin section image does not contain the proportion of non-terrigenous debris, and the relative content of particles and interstitial materials is 86.5% for particles and 13.5% for interstitial materials.

[0299] Step 2008, respectively calculate the sub-region images corresponding to the quartz debris region, the feldspar debris region, the siliceous cement region and the calcareous cement region, and perform clean region screening.

[0300] Since the component characteristics corresponding to the quartz debris region, the feldspar debris region, the siliceous cement region and the calcareous cement region are single crystal, it is necessary to screen the clean regions available for scanning. Optionally, the process is implemented by a model based on machine learning.

[0301] Step 2009, determining the surface scanning coordinate diagram.

[0302] This process is the process of determining the position of a specific scanning area. In the embodiment of the present application, the scanning method is surface scanning.

[0303] Step 2010, calculate the surface scanning coordinates, plan the optimal path, and generate scanning instructions.

[0304] Optionally, the process is a process in which a computer device determines a scanning instruction. Since the areas that need to be scanned are not adjacent in actual situations, it is necessary to plan an optimal path.

[0305] Step 2011, controlling the mineral data acquisition device to collect scanning data through scanning instructions.

[0306] This process is a process of sending the scan data to the mineral data acquisition device, so that the mineral data acquisition device generates corresponding scan data. Optionally, the scan data is spectral data.

[0307] Step 2012, generating mineral information based on the scanning data.

[0308] This process is a process of generating scanning data sent by a mineral data acquisition device based on the scanning data, and determining the mineral composition of the component area represented by the scanning data.

[0309] Step 2013, compare the mineral information with the classification and identification results to obtain the final result in each component area.

[0310] This process is to compare the mineral composition with the scanned data and finally obtain the result. Optionally, in this process, the final verification result is determined according to the type of the mineral data acquisition device.

[0311] Step 2014: replace the recognition result of the corresponding component area on the intermediate image of the clastic rock thin section sample based on the final result.

[0312] This process is to replace the component area mineral identification results obtained by the machine learning model on the intermediate image of the clastic rock thin section sample with the verification results of the surface scanning. Since spectral data is more accurate in identifying mineral components, in this step, the verification results of the surface scanning are used to determine the mineral characteristics of the component area.

[0313] Step 2015, determining the updated final image of the clastic rock thin section sample.

[0314] Combined with the example in step 2007, after the update, the proportion of siliceous cement is 2.0%, the actual form of calcareous cement is calcite, accounting for 10.5%, the proportion of matrix filling is 1%, the proportion of rock debris is 13.84%, the proportion of quartz debris is 69.2%, and the proportion of feldspar debris is 3.46%. That is, the proportion of terrigenous debris is still 86.5%, and the proportion of non-terrigenous debris is still 0%. Correspondingly, the proportion of particles is 86.5%, and the proportion of filling is 13.5%.

[0315] Step 2016, determining the thin section identification result corresponding to the clastic rock thin section.

[0316] In the embodiment of the present application, the thin slice identification results include the particle size range identification results, the particle sorting results, the maximum particle size identification results, the relative component content results, the particle roundness identification results and the particle contact mode identification results. After selecting the component area based on the mineral characteristics and generating the results based on the geometric characteristics, it can be determined that in one example, the main particle size range of the particles is fine sand (0.125-0.25mm) accounts for 10%, medium sand (0.25-0.5mm) accounts for 30%, and coarse sand (0.5-2mm) accounts for 60%; the maximum particle size is 1.8mm; the particle sorting is medium; 85% of the component areas have a particle roundness of sub-edge; 15% of the component areas have a particle roundness of sub-circle; 75% of the component areas are in line contact, and 25% of the component areas are in point contact.

[0317] Step 2017, determining the sheet name and the sheet identification report based on the sheet identification result.

[0318] Please refer to Fig.21 In the thin section identification report 2100, there is included a clastic rock naming area 2110, which includes the naming of the clastic rock thin section based on the thin section identification results. In the embodiment of the present application, the clastic rock is named "medium-coarse-grained lithic quartz sandstone". Optionally, the basis for the naming is the industry standard stored in the computer device, that is, the industry standard SY / T 5368-2016 "Rock Thin Section Identification". In the thin section identification report 2100, there is also a thin section sample information area 2120 indicating the location and method of obtaining the clastic rock sample, a component feature area 2130 indicating the mineral composition of the clastic rock sample, a structural feature area 2140 indicating the structural features of the clastic rock sample, and a rock thin section image display area 2150.

[0319] In summary, the method provided in the embodiment of the present application corresponds to the case where the rock thin section sample is clastic rock. After determining that the type is clastic rock based on structural characteristics, image segmentation and image recognition based on machine learning are used, and the recognition result obtained by image recognition is verified by acquiring spectral data. The method finally generates an identification result, which combines the chemical characteristics of the component area represented in the spectral data with the image characteristics of the component area represented in the image, to identify the clastic rock thin section, thereby improving the efficiency and accuracy of identifying the clastic rock thin section.

[0320] As described in the above examples, Fig. 20 That is, for the clastic rock type, the geometric characteristics, mineral characteristics and structural characteristics are determined, and based on the three types of characteristics, the sub-results of clastic rock thin section identification are determined, and finally the results of clastic rock thin section identification are obtained.

[0321] In one example, for the thin section identification process of igneous rocks, Fig. 22 A schematic diagram of a rock identification method provided by an exemplary embodiment of the present application is shown, the process comprising:

[0322] Step 2201, obtaining a thin section sample image of igneous rock.

[0323] Fig.23 A schematic diagram of an exemplary igneous rock thin section image provided by the present application is shown. The igneous rock thin section 2300 includes at least one olivine component region 2301, a common pyroxene component region 2302, a basic plagioclase component region 2303, a pyroxene component region 2304 and a magnetite component region 2305.

[0324] In the embodiment of the present application, Fig. 20 Similar to the illustrated embodiment, the igneous rock thin section image also includes a single polarization image and an orthogonal polarization image corresponding to the igneous rock thin section sample.

[0325] Step 2202, determine the primary rock type corresponding to the igneous rock thin section image.

[0326] In the embodiment of the present application, the igneous rock thin section image is determined to be igneous rock based on its appearance, and accordingly, the geometric features and mineral features are also features corresponding to the igneous rock.

[0327] In other embodiments of the present application, when the rock slice image is determined to be a magmatic rock slice image, the subdivided structure of the magmatic rock structure can be further determined to be a gabbro structure by model recognition. That is, after determining the primary rock type, there is a situation where the subdivided rock type is directly further determined. Accordingly, the geometric features and mineral features are also features corresponding to the magmatic rock with a gabbro structure.

[0328] Step 2203, input the igneous rock thin section sample image into the igneous rock segmentation model to obtain the segmented igneous rock thin section image.

[0329] This process is to segment the igneous rock thin section image and obtain the segmentation result.

[0330] Step 2204, extracting component regions based on the segmented igneous rock thin section image, and obtaining component type features through a component type recognition model.

[0331] This process is to extract the component area based on the segmentation results and further identify the mineral composition corresponding to the component area.

[0332] Step 2205, clean area screening is performed on the corresponding component areas in the segmented igneous rock thin section image to determine the surface scanning area.

[0333] In the embodiment of the present application, for each component area, it is necessary to verify the mineral composition corresponding to the component area by acquiring scanning data, so step 2205 is the process of determining the surface scanning area by screening the clean area.

[0334] Step 2206, calculate the surface scanning coordinates, plan the optimal path, and generate scanning instructions.

[0335] This process is the process of planning the optimal path when the corresponding surface scanning areas are not adjacent to each other, so as to generate a scanning instruction to be sent to the mineral data acquisition device.

[0336] Step 2207, controlling the mineral data acquisition device to acquire scan data through the scan instruction.

[0337] This process is the process of computer equipment controlling mineral data acquisition equipment to collect scanning data.

[0338] Step 2208, generate mineral information based on the scan data.

[0339] This process is to receive scanning data, that is, spectral data, from the mineral data acquisition equipment, and then determine the mineral type corresponding to the component area based on the spectral data.

[0340] Step 2209, compare the mineral information with the classification and identification results to obtain the final result on each component.

[0341] This process is to verify and finally determine the mineral types corresponding to each component area.

[0342] Step 2210, determine the mineral types contained in the igneous rock thin section.

[0343] In the embodiment of the present application, the mineral types corresponding to each component region can be determined to determine the mineral types contained in the igneous rock thin section. In one example, the igneous rock thin section contains mineral types including plagioclase, pyroxene, pyroxene, olivine, biotite, magnetite and apatite.

[0344] Step 2211, based on the mineral types contained, determine the subdivided rock type corresponding to the igneous rock thin section.

[0345] Based on the mineral types obtained in step 2210, it can be determined that the subdivided rock type corresponding to the igneous rock thin section is plagioclase, and the subfamily is labradorite, which belongs to basic plagioclase.

[0346] It should be noted that in the embodiment of the present application, for the type of igneous rock, the number of subdivided structural features corresponding to the igneous rock is at least two. In one example, in addition to the structural features indicating the specific type of rock, the subdivided structural features corresponding to the igneous rock also include a reaction edge structural feature, which is used to indicate the type of reaction edge contained in the igneous rock.

[0347] Step 2212, based on the rock subdivision category, determine the thin section identification result corresponding to the igneous rock thin section.

[0348] In one example of the embodiment of the present application, based on the rock subdivision category, the thin section identification results are determined to include the particle size recognition results, particle size shape results, mineral automorphic degree results, particle interaction results and mineral content results. Based on the mineral characteristics indicated by the igneous rock thin section image, in the igneous rock thin section, the plagioclase content is 52.4%, the common pyroxene content is 34.9%, the pyroxene content is 3.2%, the olivine content is 7.6%, the biotite content is 1.6%, the magnetite content is 0.4%, and the apatite content is 0.1%.

[0349] Based on the interaction between the particles corresponding to olivine and pyroxene, and the particles corresponding to pyroxene and biotite, computer equipment can determine that the olivine has a rim of pyroxene on the outside and a reaction edge structure, which indicates that the igneous rock is a basic intrusive rock or an ultrabasic intrusive rock.

[0350] In addition, through the particle size analysis of the component area, it can be obtained that the particle size of plagioclase is (1.6-2.1)x(1.8-3.5)mm, the particle size of common pyroxene is 1.2-2.2mm, the particle size of pyroxene is 1.1-1.7mm, the particle size of olivine is 0.4-2.5mm, the particle size of biotite is 1.2-2.2mm, the particle size of magnetite is 0.05-0.2mm, and the particle size of apatite is 0.05-0.1mm. After statistics, the corresponding relationship between particle size and shape can be judged by computer equipment. Plagioclase is granular-short plate, common pyroxene is irregular granular and short columnar, pyroxene is granular, olivine is granular, biotite is plate-shaped, magnetite is granular, and apatite is short columnar. And after corresponding analysis with the euhedral degree, it can be determined that plagioclase is semi-euhedral-xenomorphic, common pyroxene is xenomorphic-semi-euhedral, pyroxene is xenomorphic, olivine is xenomorphic, biotite is euhedral-semi-euhedral, magnetite is euhedral, and apatite is semi-euhedral-euhedral.

[0351] It should be noted that for the type of igneous rock, the determination and generation process of the thin section identification results can be regarded as a verification process for the structural features. In the above-mentioned embodiment, by calculating the relationship between olivine and pyroxene, or pyroxene and biotite, the computer device can determine that there is a reaction edge structure in the igneous rock thin section, and the thin section identification sub-result corresponding to the igneous rock thin section includes a reaction edge structure indicating the existence of a reaction edge. In the aforementioned steps, in the process of determining the subdivided rock type, it has been determined that the rock thin section may have a reaction edge structure, and the reaction edge structure determination process here is the verification process for the subdivided rock type.

[0352] Step 2213, determining the sheet name and the sheet identification report based on the sheet identification result.

[0353] As described in the above embodiment, in the process of determining the thin section name and the thin section identification report based on the thin section identification results, the following elements for naming igneous rocks can be extracted:

[0354] Factor 1: In the thin section image of igneous rocks, the component area corresponding to plagioclase indicates that the particle size of plagioclase is medium.

[0355] Element 2: In the thin section image of igneous rock, the mineral characteristics indicate that the main minerals are plagioclase and pyroxene.

[0356] Element 3: When the main minerals are plagioclase and pyroxene, based on geometric characteristics, it is determined that both plagioclase and pyroxene are present in granular form.

[0357] Element 4: When the main minerals are plagioclase and pyroxene, based on geometric characteristics, it is determined that the euhedral degree of plagioclase and pyroxene is similar, both showing sub-euhedral-heteromorphic shapes.

[0358] Based on the above factors, it is possible to determine the existence of olivine gabbro in the naming of igneous rock thin sections.

[0359] In the embodiment of the present application, after the computer device determines the possibility of the existence of olive gabbro in the naming, the naming of the igneous rock thin section can be verified by the method in the following example. It should be noted that the following verification process is performed when the relevant national standard data and industry standard data are stored in the computer device.

[0360] Example method: Based on the combination of mineral characteristics and geometric characteristics, the content is indicated by the area of ​​the component region. At the same time, Q is defined as the sum of the contents of quartz, tridymite and cristobalite, A is the content of alkali feldspar (i.e., the collection of orthoclase, microcline, striated feldspar, plagioclase, sansholtz and albite); P is the sum of the contents of plagioclase and scapolite; F is the sum of the contents of parafeldspar and parafeldspar (i.e., the collection of nepheline, leucite, calciticite, pseudoleucite, sodalite, pyroxene, blue spherule, cancrinite and analcime); M is the sum of the contents of mafic minerals and related minerals. Since in igneous rocks, Q group minerals and F group minerals repel each other, or if Q exists, F must be missing, and vice versa. In the case where the corresponding igneous rock is an ultrabasic intrusive rock or a basic intrusive rock, and Q = 0, A = 0, P = 52.4, F = 0, M = 47.6, please refer to Fig.24 On the intrusive rock classification phase diagram 2400, according to the relative content of P being 100%, it can be determined that the classification area corresponding to the igneous rock thin section is classification area 2401, and the classification area 2401 indicates that the rock name corresponding to the igneous rock thin section may include at least one of diorite, gabbro and anorthite. In this case, it can be verified that the type of the igneous rock thin section is gabbro.

[0361] Optionally, in other embodiments of the present application, the verification method also includes verification by subdividing the gabbro structure indicated by the structural features, etc., which will not be described in detail in the embodiments of the present application.

[0362] In this case, since the mineral characteristics indicate that the olivine content in the igneous rock thin section is 7.6%, which is greater than 5%, the olivine characteristics need to be reflected in the naming. Correspondingly, after verification, it can be finally determined that the name of the igneous rock thin section is "olivine gabbro".

[0363] It should be noted that the naming standards are based on industry standards and national standards stored in computer devices, namely, industry standard SY / T5368-2016 "Rock Thin Section Identification" and national standard GB / T 17412.1-1998 "Rock Classification and Naming Scheme - Igneous Rock Classification and Naming Scheme". Optionally, the present application embodiment also provides a thin section identification report corresponding to the igneous rock thin section.

[0364] In summary, the method provided in the embodiment of the present application corresponds to the case where the rock thin section sample is igneous rock. After determining that the type is igneous rock based on structural characteristics, image segmentation and image recognition based on machine learning are used, and the recognition result obtained by image recognition is verified by acquiring spectral data. The method finally generates an identification result, which combines the chemical characteristics of the component area represented in the spectral data and the image characteristics of the component area represented in the image to determine the subdivided type of igneous rock, and then further identifies the structural characteristics and component characteristics represented by the igneous rock thin section, thereby improving the efficiency and accuracy of identifying clastic rock thin sections.

[0365] As described in the above examples, Fig. 22 This is an embodiment of a method for identifying igneous rocks, which is a process of obtaining sub-results of thin-section identification of igneous rocks through geometric features, mineral features and structural features, and then obtaining the final identification results.

[0366] Fig.25 The following is a structural block diagram of a rock identification device provided by an exemplary embodiment of the present application, wherein the device comprises:

[0367] The receiving module 2501 is used to receive a rock thin section image sent by an image acquisition device. The rock thin section image is an image obtained by photographing a rock thin section. The rock thin section is a thin section obtained by cutting a rock sample. The rock thin section image includes at least one component region.

[0368] A generating module 2502 is used to generate geometric features, mineral features and structural features corresponding to the rock slice based on the rock slice image, wherein the geometric features are used to indicate the regional division of the rock slice based on components, the mineral features are used to indicate the distribution of minerals in the rock slice, and the structural features are used to indicate the spatial structure of the rock sample;

[0369] Based on geometric, mineral and structural characteristics, the identification results of rock thin sections are generated, including a textual description of the characteristics of the rock samples.

[0370] In an optional embodiment, the structural feature includes a preliminary rock type structural feature, and the preliminary rock type structural feature is used to indicate the preliminary rock type corresponding to the rock sample, and the preliminary rock type includes sedimentary rock, igneous rock and metamorphic rock;

[0371] The generation module 2502 is also used to generate preliminary rock structure features based on the rock thin section image;

[0372] Please refer to Fig.26 , the device also includes a determination module 2503, which is used to determine the primary rock type corresponding to the rock slice based on the structural characteristics of the primary rock type;

[0373] The generation module 2502 is also used to generate geometric features based on the rock slice image according to the initially selected rock type;

[0374] According to the preliminary rock type selection, mineral characteristics are generated based on rock thin section images.

[0375] In an optional embodiment, the structural feature further includes a subdivided rock type structural feature, and the subdivided rock type structural feature is used to indicate the subdivided rock type corresponding to the rock sample;

[0376] The determination module 2503 is also used to determine the structural characteristics of the subdivided rock types based on the geometric characteristics and mineral characteristics.

[0377] In an optional embodiment, the rock slice image includes at least two component images representing component regions;

[0378] The device further includes a segmentation module 2504, which is used to segment the rock slice image based on the positions of the component images to obtain a rock slice segmentation image, wherein the rock slice segmentation image includes at least two segmentation regions, and the segmentation regions are used to indicate the segmentation of the rock slice image based on the component regions;

[0379] The determination module 2503 is further used to determine the geometric features corresponding to the rock slice based on the segmented area;

[0380] The mineral features corresponding to the rock thin sections are determined based on the segmented regions.

[0381] In an optional embodiment, the device further includes an input module 2505, which is used to input the rock thin section image into a rock thin section image segmentation model, and output a rock thin section segmentation image, and the rock thin section image segmentation model is a Mask-RCNN network model based on machine learning.

[0382] In an optional embodiment, the determination module 2503 is further configured to determine a component image based on the segmented region, wherein the component image includes a component region;

[0383] The input module 2505 is also used to input the component image into the component type recognition model, and output the component type features of the component image. The component type features are used to characterize the mineral features of the component area. The component type recognition model is a neural network model based on machine learning.

[0384] The determination module 2503 is further used to determine the mineral features corresponding to the rock slice based on the component type features corresponding to the component image.

[0385] In an optional embodiment, the rock identification system further includes a mineral data acquisition device, which is connected to the computer device;

[0386] The input module 2505 is also used to input the rock thin section segmentation image into the classification recognition model, and output a single crystalline component set image and a non-single crystalline component set image, wherein the single crystalline component set image includes at least one single crystalline component region and a single crystalline component recognition result corresponding to the single crystalline component region, and the non-single crystalline component region includes at least one non-single crystalline component region and a non-single crystalline component recognition result corresponding to the non-single crystalline component region, and the classification recognition model is a neural network model based on deep learning;

[0387] The device also includes a sending module 2506 for sending a scanning instruction to a mineral data acquisition device based on the single crystal component set image;

[0388] The receiving module 2501 is also used to receive the scanning data fed back by the mineral data acquisition device based on the scanning instruction;

[0389] The device further includes a verification module 2507 for verifying the single crystal component identification result based on the scanning data to obtain a verification result;

[0390] The determination module 2503 is further used to determine the mineral characteristics corresponding to the rock thin section based on the non-single crystalline component identification results, the single crystalline component identification results and the verification results.

[0391] In an optional embodiment, the input module 2505 is further used to input the single crystal component set image into the regional screening model, and output the regional screening result, which is used to indicate the portion of the single crystal component set image used for surface scanning data;

[0392] The generating module 2502 is further used to generate a scanning instruction based on the area screening result;

[0393] The sending module 2506 is also used to send a scanning instruction to the mineral data acquisition device.

[0394] In an optional embodiment, the determination module 2503 is further configured to determine a scan recognition result based on the scan data;

[0395] In response to the scanning identification result being the same as the single-crystalline component identification result, determining that the verification result indicates that the verification is passed, and determining that the single-crystalline component identification result is unchanged;

[0396] In response to the surface scanning recognition result being different from the single crystal component recognition result, it is determined that the verification result indicates that the verification failed, and the single crystal component recognition result is determined based on the recognition result generation rule, and the recognition result generation rule includes a rule based on the type of mineral collection equipment.

[0397] In an optional embodiment, the determination module 2503 is further configured to determine a component image based on the segmented region, wherein the component image includes a component region;

[0398] The sending module 2506 is also used to send a spectral data acquisition instruction to the mineral data acquisition device based on the component image;

[0399] The receiving module 2501 is also used to receive the spectral data sent by the mineral data acquisition device based on the spectral data acquisition instruction;

[0400] The determination module 2503 is further used to determine the preliminary selected mineral type corresponding to the component region in the mineral spectrum data database based on the spectrum data;

[0401] Determine the mineral type verification rules corresponding to the primary selected mineral types;

[0402] Determine the mineral type corresponding to the component area based on the mineral type verification rules and the preliminary selected mineral types;

[0403] Based on the mineral types corresponding to the component images, the mineral characteristics corresponding to the rock thin sections are determined.

[0404] In an optional embodiment, the mineral spectral data database includes a mineral type sub-database;

[0405] The determination module 2503 is further used to determine the primary categories of the preliminary selected mineral types corresponding to the component region in the mineral spectrum data database based on the spectral data, the primary categories of the mineral types are categories divided based on the commonness of the minerals, the primary categories include common mineral categories and uncommon mineral categories, the preliminary selected mineral types corresponding to the common mineral categories include mineral family categories and common common mineral types, and the preliminary selected mineral types corresponding to the uncommon mineral categories include inclusion mineral types, highly sensitive mineral types and alteration mineral types;

[0406] Based on the mineral type sub-database corresponding to the primary category of mineral types, the preliminary mineral types corresponding to the component areas are determined.

[0407] In an optional embodiment, the mineral type verification rules include classification verification rules, direct verification rules, and point selection re-verification rules;

[0408] The determination module 2503 is further used to determine the mineral type verification rule as a classification verification rule in response to the primary selected mineral type being a mineral family type;

[0409] In response to the primary selected mineral type being an inclusion mineral type, or a highly sensitive mineral type, or an altered mineral type, determining the mineral type verification rule to be a point selection re-verification rule;

[0410] In response to the preliminary selected mineral type being a common mineral type, or the primary category corresponding to the preliminary selected mineral type being an uncommon mineral type, and the preliminary selected mineral type is not any of the inclusion mineral type, the highly sensitive mineral type and the alteration mineral type, the mineral type verification rule is determined to be a direct verification rule.

[0411] In an optional embodiment, the rock slice image includes at least two component images representing component regions;

[0412] The input module 2505 is further used to input the rock thin section image into the rock thin section image segmentation and recognition model, and output at least two segmentation regions corresponding to the rock thin section image and component type features corresponding to the segmentation regions, wherein the segmentation regions are used to indicate the segmentation of the rock thin section image based on the component regions, and the component type features are used to characterize the mineral features of the component regions in the segmentation regions. The rock thin section image segmentation and recognition model is a neural network model based on machine learning;

[0413] The determination module 2503 is also used to determine the geometric features and mineral features corresponding to the rock slices in combination with the segmented regions and component type features.

[0414] In an optional embodiment, the input module 2505 is also used to input the rock thin section image into the preliminary rock type selection model, and output the preliminary rock type result, which indicates the preliminary rock type structural characteristics corresponding to the rock thin section. The preliminary rock type structural selection model is a model constructed based on the preliminary rock type image sample set.

[0415] In an optional embodiment, the identification results include subdivided rock types. The input module 2505 is also used to input geometric features and mineral features into a subdivided rock type selection model, and output a subdivided rock type result. The subdivided rock type result indicates the subdivided rock structural features corresponding to the rock thin section. The subdivided rock type selection model is a model constructed based on a geometric-mineral feature interactive sample set. The geometric-mineral feature interactive sample set indicates the combination of geometric features and mineral features, and the corresponding relationship between the subdivided rock type structural features.

[0416] In an optional embodiment, the identification result includes at least one of a rock thin section sub-identification result and a rock thin section name;

[0417] The determination module 2503 is further used to determine the type of at least one rock thin section identification result corresponding to the identification result based on the primary rock type characteristics and the detailed rock type characteristics, and the rock thin section identification result includes at least one of a particle size range identification result, a maximum particle size identification result, a particle sorting result, a particle roundness identification result, a mineral eumorphism result, and a particle contact mode identification result;

[0418] The generation module 2502 is also used to generate rock thin section identification results based on the geometric features and the mineral features;

[0419] The determination module 2503 is further used to determine the name of the rock thin section based on the rock thin section identification result.

[0420] In summary, the device provided in the embodiment of the present application, after acquiring the rock thin section image, performs feature extraction based on three dimensions of geometric features, mineral features, and structural features on the rock thin section image, determines the properties of the rock from the perspective of microscopic composition and macroscopic performance, and finally generates an identification result including a textual description. In the process of identifying the rock, after obtaining the microscopic visualization image corresponding to the rock thin section, the image is subjected to feature extraction in multiple dimensions, and the rock thin section is identified with reference to the features in multiple dimensions, thereby improving the accuracy of rock identification.

[0421] It should be noted that the rock identification device provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the equipment can be divided into different functional modules to complete all or part of the functions described above.

[0422] Fig. 27 A schematic diagram of the structure of a computer device for executing a rock identification method provided by an exemplary embodiment of the present application is shown, and the computer device includes:

[0423] The processor 2701 includes one or more processing cores. The processor 2701 executes various functional applications and data processing by running software programs and modules.

[0424] The receiver 2702 and the transmitter 2703 can be implemented as a communication component, and the communication component can be a communication chip. Optionally, the communication component can be implemented to include a signal transmission function. That is, the transmitter 2703 can be used to transmit a control signal to the image acquisition device and the scanning device, and the receiver 2702 can be used to receive the corresponding feedback instruction.

[0425] The memory 2704 is connected to the processor 2701 via a bus 2705 .

[0426] The memory 2704 may be used to store at least one instruction, and the processor 2701 may be used to execute the at least one instruction to implement each step in the above method embodiment.

[0427] An embodiment of the present application also provides a computer-readable storage medium, which stores at least one instruction, at least one program, code set or instruction set, which is loaded and executed by a processor to implement the above-mentioned rock identification method.

[0428] The present application also provides a computer program product or a computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the rock identification method described in any of the above embodiments.

[0429] Optionally, the computer readable storage medium may include: a read-only memory (ROM), a random access memory (RAM), a solid state drive (SSD), or an optical disk. Among them, the random access memory may include a resistance random access memory (ReRAM) and a dynamic random access memory (DRAM). The serial numbers of the above embodiments of the present application are only for description and do not represent the advantages and disadvantages of the embodiments.

[0430] A person skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware, or by a program to instruct the relevant hardware, and the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk or an optical disk, etc.

[0431] The above are only optional embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A rock identification method, characterized in that: The method is applied to a computer device of a rock identification system, wherein the rock identification system comprises an image acquisition device and a computer device; The image acquisition device is connected to the computer device; The method comprises: receiving a rock thin section image sent by the image acquisition device, wherein the rock thin section image is an image obtained by photographing a rock thin section, wherein the rock thin section is a thin section obtained by cutting a rock sample, and wherein the rock thin section image includes at least one component region; Generate geometric features, mineral features and structural features corresponding to the rock slice based on the rock slice image, wherein the geometric features are used to indicate the regional division of the rock slice based on components, the mineral features are used to indicate the mineral identification results obtained by performing mineral identification on the rock slice, and the structural features are used to indicate the spatial structure of the rock sample; Based on the geometric features, the mineral features and the structural features, generating an identification result of the rock slice, wherein the identification result includes a textual description of the features of the rock sample; The structural features include primary rock type structural features and detailed rock type structural features, wherein the primary rock type structural features are used to indicate the primary rock type corresponding to the rock sample, the primary rock type includes sedimentary rock, igneous rock and metamorphic rock, and the detailed rock type structural features are used to indicate the detailed rock type corresponding to the rock sample; The generating of geometric features, mineral features and structural features corresponding to the rock slice based on the rock slice image comprises: Generating the primary selected rock structure features based on the rock thin section image; Determining the initially selected rock type corresponding to the rock slice based on the structural characteristics of the initially selected rock type; According to the initially selected rock type, the rock thin section image is divided based on the positions of the component images to obtain a rock thin section segmentation image, wherein the rock thin section segmentation image includes at least two segmentation areas; determining geometric features corresponding to the rock slice based on the segmented region; determining a mineral feature corresponding to the rock slice based on the segmented region; Based on the geometric features and the mineral features, the structural features of the subdivided rock types are determined.

2. The method according to claim 1, characterized in that The step of dividing the rock slice image based on the positions of the component images to obtain the rock slice segmentation image comprises: The rock thin section image is input into a rock thin section image segmentation model, and the rock thin section segmentation image is output. The rock thin section image segmentation model is a Mask-RCNN network model based on machine learning.

3. The method according to claim 1, characterized in that The determining of the mineral features corresponding to the rock slice based on the segmented area comprises: Determine a component image based on the segmented area, wherein the component image includes one of the component areas; Input the component image into a component type recognition model, and output component type features of the component image, wherein the component type features are used to characterize the mineral features of the component region, and the component type recognition model is a neural network model based on machine learning; The mineral features corresponding to the rock slice are determined based on the component type features corresponding to the component image.

4. The method according to claim 1, characterized in that: The rock identification system also includes a mineral data acquisition device, which is connected to the computer device; The determining of the mineral features corresponding to the rock slice based on the segmented area comprises: Inputting the rock thin section segmentation image into a classification and recognition model, outputting a single crystalline component set image and a non-single crystalline component set image, wherein the single crystalline component set image includes at least one single crystalline component region and a single crystalline component recognition result corresponding to the single crystalline component region, and the non-single crystalline component region includes at least one non-single crystalline component region and a non-single crystalline component recognition result corresponding to the non-single crystalline component region, and the classification and recognition model is a neural network model based on deep learning; Sending a scanning instruction to the mineral data acquisition device based on the single crystal component set image; Receiving scanning data fed back by the mineral data acquisition device based on the scanning instruction; Verifying the single crystal component identification result based on the scanning data to obtain a verification result; Based on the non-single crystalline component identification result, the single crystalline component identification result and the verification result, the mineral characteristics corresponding to the rock thin section are determined.

5. The method according to claim 4, characterized in that The sending of a scanning instruction to the mineral data acquisition device based on the single crystal component set image comprises: Inputting the single crystal component set image into a regional screening model, and outputting a regional screening result, wherein the regional screening result is used to indicate a portion of the single crystal component set image used for performing surface scanning data; generating the scanning instruction based on the area screening result; The scanning instruction is sent to the mineral data acquisition device.

6. The method according to claim 5, characterized in that The verifying the single crystal component identification result based on the scanning data to obtain a verification result includes: Determine a scan recognition result based on the scan data; In response to the scanning identification result being the same as the single-crystalline component identification result, determining that the verification result indicates that the verification is passed, and determining that the single-crystalline component identification result is unchanged; In response to the surface scanning identification result being different from the single crystal component identification result, it is determined that the verification result indicates that the verification has failed, and the single crystal component identification result is determined based on an identification result generation rule, wherein the identification result generation rule includes a rule based on the type of the mineral data acquisition device.

7. The method according to claim 4, characterized in that The step of determining the mineral features corresponding to the rock slice based on the segmented area further comprises: Determine a component image based on the segmented area, wherein the component image includes a component area; Sending a spectral data acquisition instruction to the mineral data acquisition device based on the component image; receiving spectral data sent by the mineral data acquisition device based on the spectral data acquisition instruction; Based on the spectral data, determining the primary mineral type corresponding to the component region in a mineral spectral data database; Determining a mineral type verification rule corresponding to the initially selected mineral type; Determine the mineral type corresponding to the component area based on the mineral type verification rule and the preliminary selected mineral type; The mineral features corresponding to the rock slice are determined based on the mineral types corresponding to the component images constituting the rock slice segmentation image.

8. The method according to claim 7, characterized in that The mineral spectrum data database includes a mineral type sub-database; The step of determining the type of preliminary selected minerals corresponding to the component region in a mineral spectral data database based on the spectral data comprises: Based on the spectral data, the primary categories of the preliminary selected mineral types corresponding to the component area are determined in the mineral spectral data database, the primary categories of the mineral types are categories divided based on the commonness of the minerals, the primary categories include common mineral categories and uncommon mineral categories, the preliminary selected mineral types corresponding to the common mineral categories include mineral family categories and common common mineral types, and the preliminary selected mineral types corresponding to the uncommon mineral categories include inclusion mineral types, highly sensitive mineral types and alteration mineral types; Based on the mineral type sub-database corresponding to the primary category of the mineral type, the preliminary selected mineral type corresponding to the component area is determined.

9. The method according to claim 8, characterized in that The mineral type verification rules include classification verification rules, direct verification rules and point selection re-verification rules; The step of determining a mineral type verification rule corresponding to the initially selected mineral type includes: In response to the initially selected mineral type being the mineral family type, determining the mineral type verification rule to be the classification verification rule; In response to the initially selected mineral type being the inclusion mineral type, or the strongly sensitive mineral type, or the altered mineral type, determining the mineral type verification rule to be the point selection re-verification rule; In response to the preliminary selected mineral type being the common mineral type, or the primary category corresponding to the preliminary selected mineral type being the uncommon mineral category, and the preliminary selected mineral type is not any one of the inclusion mineral type, the highly sensitive mineral type and the alteration mineral type, the mineral type verification rule is determined to be the direct verification rule.

10. The method according to claim 1, characterized in that The rock slice image includes at least two component images representing component areas; The generating of geometric features, mineral features and structural features corresponding to the rock slice based on the rock slice image comprises: Input the rock thin section image into a rock thin section image segmentation and recognition model, and output at least two segmentation regions corresponding to the rock thin section image and component type features corresponding to the segmentation regions, wherein the segmentation regions are used to indicate the segmentation of the rock thin section image based on component regions, and the component type features are used to characterize the mineral features of the component regions in the segmentation regions, and the rock thin section image recognition model is a neural network model based on machine learning; The geometric features and mineral features corresponding to the rock slice are determined in combination with the segmented areas and the component type features.

11. The method according to claim 1, characterized in that: The identification result includes a preliminary rock type, and the determining of the preliminary rock type corresponding to the rock thin section based on the structural characteristics of the preliminary rock type includes: The rock thin section image is input into a preliminary rock type selection model, and a preliminary rock type result is output, wherein the preliminary rock type result indicates the structural characteristics of the preliminary rock type corresponding to the rock thin section, and the preliminary rock type structural selection model is a model constructed based on a sample set of preliminary rock type images.

12. The method according to claim 11, characterized in that The identification result includes the subdivided rock type, and the determination of the subdivided rock type structural features based on the geometric features and the mineral features includes: The geometric features and the mineral features are input into a subdivided rock type selection model, and a subdivided rock type result is output, wherein the subdivided rock type result indicates the subdivided rock type structural feature corresponding to the rock thin section. The subdivided rock type selection model is a model constructed based on a geometric-mineral feature interactive sample set, and the geometric-mineral feature interactive sample set indicates the combination of the geometric features and the mineral features, and the corresponding relationship between the subdivided rock type structural feature.

13. The method according to claim 12, characterized in that The identification result includes at least one of a rock thin section identification result and a rock thin section name; The step of generating the identification result of the rock slice based on the geometric features, the mineral features and the structural features comprises: Based on the primary rock type characteristics and the detailed rock type characteristics, determining the type of at least one rock thin section identification result corresponding to the identification result, wherein the rock thin section identification result includes at least one of a particle size range identification result, a maximum particle size identification result, a particle sorting result, a particle roundness identification result, a mineral automorphic degree result, and a particle contact mode identification result; generating the rock thin section identification result based on the geometric features and the mineral features; The name of the rock thin section is determined based on the identification result of the rock thin section.

14. A rock identification system, characterized in that: The rock identification system includes an image acquisition device and a computer device, wherein the image acquisition device is connected to the computer device; The image acquisition device is used to generate a rock thin section image; and send the rock thin section image to the computer device; The computer device is used to receive a rock slice image sent by the image acquisition device, wherein the rock slice image is an image obtained by photographing the rock slice, and the rock slice is a slice obtained by cutting a rock sample, and the rock slice image includes at least one component region; based on the rock slice image, a geometric feature, a mineral feature, and a structural feature corresponding to the rock slice are generated, wherein the geometric feature is used to indicate the regional division of the rock slice based on the component, the mineral feature is used to indicate the mineral distribution in the rock slice, and the structural feature is used to indicate the spatial structure of the rock sample; based on the geometric feature, the mineral feature, and the structural feature, an identification result of the rock slice is generated, and the identification result includes a textual description of the characteristics of the rock sample; the structural feature includes a primary rock type structure. The preliminary selected rock type structural features and detailed rock type structural features, wherein the preliminary selected rock type structural features are used to indicate the preliminary selected rock type corresponding to the rock sample, the preliminary selected rock types include sedimentary rocks, igneous rocks and metamorphic rocks, and the detailed rock type structural features are used to indicate the detailed rock type corresponding to the rock sample; the generating of geometric features, mineral features and structural features corresponding to the rock thin section based on the rock thin section image comprises: generating the preliminary selected rock type structural features based on the rock thin section image; determining the preliminary selected rock type corresponding to the rock thin section based on the preliminary selected rock type structural features; generating the geometric features based on the rock thin section image according to the preliminary selected rock type; generating the mineral features based on the rock thin section image according to the preliminary selected rock type; and determining the detailed rock type structural features based on the geometric features and the mineral features.

15. The system according to claim 14, characterized in that The image acquisition device comprises a first light source, a first stage, a first objective lens, a polarization device, a charge coupled device (CCD) camera, a first base and a first bracket; The first stage is located at the center of the first bracket, the first objective lens is located at the top of the first bracket, the first objective lens has an analyzer, the bottom of the first objective lens has a first mounted objective lens converter, and the first mounted objective lens converter is connected to the first objective lens; The CCD camera is located on the top of the first objective lens; The polarization device includes a polarizer and an analyzer, the polarizer and the first light source are located on the first base, and the first base is connected to the bottom of the first bracket; When the image acquisition device generates the rock slice image, the rock slice is located on the top of the first stage, and the CCD camera is in a first working state.

16. The system according to claim 15, characterized in that The rock identification system also includes mineral data acquisition equipment; The mineral data acquisition device is communicatively connected with the computer device; The mineral data acquisition device includes a second light source, a second stage, a second objective lens, a measuring device, a photoelectric signal converter, a second base and a second bracket; The second light source corresponds to a second light path, and the end point of the second light path is projected onto the second stage; The second stage is located at the center of the second bracket, the second objective lens is located at the top of the second bracket, the bottom of the second objective lens has a second mounted objective lens converter, and the second mounted objective lens converter is connected to the second objective lens; The photoelectric signal converter is located on the top of the second objective lens and is communicatively connected with the computer device; When the mineral data acquisition device generates mineral data, the rock slice is located on the top of the second stage, and the photoelectric signal converter is in a second working state.

17. The system according to claim 16, characterized in that The mineral data acquisition device and the image acquisition device share the first stage, the first objective lens, the first base and the first bracket.

18. A rock identification device, characterized in that: The device comprises: A receiving module, used for receiving a rock thin section image sent by an image acquisition device, wherein the rock thin section image is an image obtained by photographing a rock thin section, wherein the rock thin section is a thin section obtained by cutting a rock sample, and wherein the rock thin section image includes at least one component region; A generation module is used to generate geometric features, mineral features and structural features corresponding to the rock slice based on the rock slice image, wherein the geometric features are used to indicate the regional division of the rock slice based on components, the mineral features are used to indicate the distribution of minerals in the rock slice, and the structural features are used to indicate the spatial structure of the rock sample; the structural features include primary rock type structural features and detailed rock type structural features, wherein the primary rock type structural features are used to indicate the primary rock type corresponding to the rock sample, the primary rock type includes sedimentary rock, igneous rock and metamorphic rock, and the detailed rock type The class structure feature is used to indicate the subdivided rock type corresponding to the rock sample; the generating of the geometric features, mineral features and structural features corresponding to the rock thin section based on the rock thin section image includes: generating the preliminary rock type structural features based on the rock thin section image; determining the preliminary rock type corresponding to the rock thin section based on the preliminary rock type structural features; generating the geometric features based on the rock thin section image according to the preliminary rock type; generating the mineral features based on the rock thin section image according to the preliminary rock type; determining the subdivided rock type structural features based on the geometric features and the mineral features; The generation module is used to generate an identification result of the rock slice based on the geometric features, the mineral features and the structural features, wherein the identification result includes a textual description of the features of the rock sample.

19. A computer device, characterized in that: The computer device includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, at least one program, a code set or an instruction set is loaded and executed by the processor to implement the rock identification method as described in any one of claims 1 to 13.

20. A computer-readable storage medium, characterized in that: The readable storage medium stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by the processor to implement the rock identification method as described in any one of claims 1 to 13.

Images