Mine integrated modeling method and system driven by data and knowledge fusion

Abstract

The invention relates to the technical field of environment model construction, in particular to a data and knowledge fusion driven mine integrated modeling method and system. The method comprises the following steps: firstly, preprocessing multi-source heterogeneous data to establish a unified data set, and constructing an object-attribute-environment three-domain associated mine knowledge graph based on physical space information; then, initial geometric models of the earth surface, the ore body, the roadway and the like are constructed respectively, and mapping is established between the models and the atlas; performing forced geometric constraint and logic optimization on the initial model by emphatically utilizing semantic rules in the atlas, and automatically correcting space conflicts; and finally, realizing multi-element space integration under a unified coordinate system. According to the method, deep fusion of geometric reconstruction and geological semantics is realized, the problem of model logic deficiency is effectively solved, a high-precision three-dimensional model with dynamic evolution ability and logic self-consistency is constructed, and reliable support is provided for mine intelligent management.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of environment model construction, in particular to a data and knowledge fusion driven integrated modeling method and system for a mine. BACKGROUND

[0002] With the extension of mineral resources exploitation to deep and complex environments, constructing a high-fidelity 3D model of underground mines has become a key foundation for realizing safety monitoring, risk assessment and intelligent management of the whole life cycle of mines. Virtual geographic environment technology can display the 3D spatial distribution relationship of ore bodies, roadways and the surface by mapping the physical world to the information space, and is the mainstream means of current mine digital construction.

[0003] However, existing 3D modeling techniques for mines mainly rely on geometric fitting of multi-source data such as drill holes, geological profiles and point clouds. This data-driven method generates a model that is essentially a "geometric shell" with only appearance characteristics, lacking expression of the internal logical relationship between geological objects (such as ore bodies and faults) and engineering objects (such as roadways and goaf). In areas with sparse data or occlusions, simple geometric modeling often cannot guarantee the correctness of spatial topological relationships, and may result in phenomena such as model breakage, penetration or suspension that violate geological common sense, making it difficult to meet the needs of detailed analysis.

[0004] On the other hand, although some existing techniques attempt to introduce geological rules to assist modeling, these rules are usually independent of the 3D modeling process and cannot be deeply integrated with the geometric construction process. There is still a lack of an effective mechanism to convert abstract geological domain knowledge into quantifiable constraints and directly inject them into the geometric modeling process. Therefore, existing methods cannot ensure that the model has complete semantic information and rigorous geological logical consistency while constructing a high-precision geometric model, and cannot support dynamic evolution simulation and intelligent decision-making in complex mine environments. SUMMARY

[0005] The main purpose of the present application is to provide a data and knowledge fusion driven integrated modeling method for a mine, which aims to solve the problem that the model generated by the existing data-driven modeling method lacks geological semantic constraints, resulting in spatial topological errors in areas with sparse or conflicting data, and failing to balance the geometric high precision and geological logical consistency of the model.

[0006] To achieve the above purpose, the present application provides a data and knowledge fusion driven integrated modeling method for a mine, which comprises the following steps: acquiring multi-source heterogeneous spatial data of the mine, preprocessing the multi-source heterogeneous spatial data to establish a unified data set; acquiring physical space information of the mine, and constructing a mine knowledge graph according to the acquired physical space information; constructing initial geometric models of the ground surface, the geological body, the ore body, and the tunnel respectively according to the unified data set; mapping and associating the initial geometric models with the mine knowledge graph, performing geometric constraint and logical optimization on the initial geometric models according to semantic rules in the mine knowledge graph, to construct a revised model with semantic information; spatially integrating the revised ground surface model, the geological body model, the ore body model, and the tunnel model in a unified coordinate system, to obtain a mine integrated three-dimensional model with data and knowledge fusion.

[0007] Optionally, the preprocessing of the multi-source heterogeneous spatial data comprises the following steps: performing coordinate unification, format conversion, redundant data cleaning, and attribute field standardization processing on the unmanned aerial vehicle image, the drilling data, the geological profile, the three-dimensional laser point cloud, and the CAD data; extracting core geological and engineering element information of the ground surface, the geological body, the ore body, and the tunnel, and establishing the unified data set with consistent data format and spatial reference.

[0008] Optionally, the constructing of the mine knowledge graph according to the obtained physical space information comprises the following steps: constructing a three-domain association structure including an object domain, an attribute domain, and an environment domain; obtaining association rules between the object domain, the attribute domain, and the environment domain, to establish an entity and relationship sample set; wherein, the object domain comprises an underground ore body, a geological body, a tunnel structure, and a ground surface element; the attribute domain comprises spatial position, lithology, structural characteristics, grade, geometric size, and topographic form; the environment domain comprises ground subsidence, collapse, and tailings accumulation environment elements.

[0009] Optionally, the constructing of the mine knowledge graph according to the obtained physical space information further comprises the following steps: obtaining semantic logic between objects according to the three-domain association structure of the mine; mapping the entity and relationship sample set into an RDF triple form according to the three-domain association structure of the mine, and importing into a graph database to generate a node and relationship network; wherein, the semantic logic at least includes that the ore body contains the tunnel, the tunnel is located inside the ore body, and the ground surface is affected by mining.

[0010] Optionally, the constructing of the initial geometric models of the ground surface, the geological body, the ore body, and the tunnel respectively according to the unified data set comprises the following steps: fusing the three-dimensional point cloud and the CAD data in the unified data set to construct a tunnel model; According to the drilling lithology data in the unified data set, the stratum and structure information interpreted from the geological profile map is fused to construct a three-dimensional geological body model; According to the unmanned aerial vehicle image in the unified data set, a ground surface model is generated through aerial triangulation, image matching and color calibration; According to the drilling data in the unified data set, the ore body boundary is determined through geological interpretation, the triangular net modeling method is used to construct the ore body roof and floor curved surface, and a three-dimensional ore body entity model is constructed.

[0011] Optionally, when constructing the ground surface model, the distance power inverse ratio method is used to grid estimate the ground surface sampling points to generate a smooth surface model.

[0012] Optionally, the geometric constraint and logical optimization of the initial geometric model according to the semantic rules in the mine knowledge graph comprises the following steps: According to the spatial topology, spatial distribution and geometric shape rules defined in the mine knowledge graph, the spatial discontinuity of the data sparse area in the initial geometric model is checked; When the topological relationship of the initial geometric model conflicts with the semantic rules, the geometric shape or spatial position of the initial geometric model is adjusted according to the semantic rules until the geological semantic consistency is met.

[0013] Optionally, the method further comprises the following steps: Obtaining the unique identifier mapping of the geometric object in the mine integrated three-dimensional model and the semantic entity node in the mine knowledge graph; When the object attribute or environment state in the mine knowledge graph changes is monitored, the geometric shape of the corresponding object in the mine integrated three-dimensional model is synchronously adjusted through the unique identifier mapping; when the data of the mine integrated three-dimensional model is updated, the new object or new attribute is connected to the mine knowledge graph through the incremental update mechanism.

[0014] Optionally, the semantic rules are specifically configured as: For the tunnel object, the spatial semantic constraint is set, which includes limiting the tunnel model to be located within the internal boundary of the ore body model or the geological body model, and the extension direction of the tunnel model matches the structure feature trend of the ore body; For the ground surface object, the environmental dynamic constraint is set, and the deformation parameters of the ground surface model are associated and constrained with the ground subsidence environmental elements in the mine knowledge graph; For the geological body object, the geological semantic constraint is set, the geological body model matches the drilling data, and is consistent with the geological map and the interpreted profile in space; For the ore body object, set geological and engineering semantic constraints, and match the ore body model with drilling data and lithology data.

[0015] To achieve the above object, the application further provides a modeling system, which comprises: A data preprocessing module is used to acquire multi-source heterogeneous spatial data of a mine, and the multi-source heterogeneous spatial data is preprocessed to establish a unified data set. A knowledge graph construction module is used to acquire physical space information of the mine, and a mine knowledge graph is constructed according to the acquired physical space information. A geometric modeling module is used to construct initial geometric models of the ground surface, geological body, ore body and roadway respectively according to the unified data set. A semantic fusion and optimization module is used to map and correlate the initial geometric models with the mine knowledge graph, and the initial geometric models are geometrically constrained and logically optimized according to the semantic rules in the mine knowledge graph to construct revised models with semantic information. An integrated expression module is used to spatially integrate the revised ground surface revised model, geological body revised model, ore body revised model and roadway revised model in a unified coordinate system to acquire a mine integrated three-dimensional model with data and knowledge fusion.

[0016] The application has the following beneficial effects: The application constructs a mine knowledge graph based on the correlation of object-attribute-environment three domains, and establishes deep mapping correlation between the initial geometric models generated by multi-source heterogeneous data and the graph, and uses the semantic rules in the graph to perform forced geometric constraint and logical optimization on the models; effectively solves the technical problems of spatial topology errors such as suspended roadway and inserted ore body caused by lack of geological semantic guidance under the condition of sparse or blind area of existing single data driven modeling, and the technical problem of missing internal logical correlation due to only having appearance geometric features of the model; realizes the leap of the mine three-dimensional scene from geometric reconstruction to semantic and logical dual high-fidelity expression; specifically, the application not only fills the geometric defects of the data blind area by using geological knowledge, ensures the absolute logical self-consistency of the ore body, roadway and ground surface in spatial relationship, and eliminates the modeling errors contrary to the geological common sense; but also establishes a bidirectional dynamic index of geometric objects and semantic entities, and gives the static model the ability of real-time updating with environmental evolution (such as subsidence and mining progress); this deep coupling of data and knowledge builds a digital twin foundation with explainability, logical completeness and dynamic evolution ability, significantly improves the spatial analysis precision under complex mine environment, and provides reliable support with geometric precision and semantic depth for safety monitoring, risk early warning and intelligent decision-making of the whole life cycle of the mine. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced. In all the drawings, similar elements or parts are generally identified by similar reference signs. In the drawings, each element or part is not necessarily drawn according to the actual scale.

[0018] Figure 1 a flowchart of the method in embodiment 1 of the present application; Figure 2 a structural block diagram of the system in embodiment 2 of the present application; The implementation, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all the other embodiments obtained by those skilled in the art without any creative work under the premise that the present application falls within the scope of protection.

[0020] It should be noted that all the directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present application are only used to explain the relative position relationship, movement condition, etc. between the components in a certain specific posture, and if the specific posture changes, the directional indications will also change accordingly.

[0021] In the present application, unless otherwise explicitly specified and limited, the terms "connection", "fixation", etc. should be understood in a broad sense, for example, "connection" can be fixed connection, or detachable connection, or integral; can be mechanical connection, or electrical connection; can be direct connection, or indirect connection through an intermediate medium; can be the internal connection of two elements or the interaction relationship between two elements, unless otherwise explicitly limited. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0022] In addition, if the description of "first", "second" and the like is involved in the embodiments of the present application, the description of "first", "second" and the like is only for the purpose of description, and cannot be understood as indicating or implying the relative importance of the indicated technical features or implicitly indicating the number of the indicated technical features. Therefore, the features limited by "first", "second" can be explicitly or implicitly included at least one of the features. In addition, the meaning of "and / or" appearing throughout the text includes three parallel schemes. For example, "A and / or B" includes A scheme, or B scheme, or A and B scheme. In addition, the technical solutions of each embodiment can be combined with each other, but it must be based on the realization of ordinary skilled in the art. When the combination of technical solutions contradicts each other or cannot be realized, it should be considered that the combination of technical solutions does not exist, and is not within the protection scope required by the present application. Embodiments

[0023] Referring to Figure 1 The present embodiment provides a data and knowledge fusion driven mine integrated modeling method, the method comprising the following steps: Obtaining multi-source heterogeneous spatial data of the mine, preprocessing the multi-source heterogeneous spatial data to establish a unified data set; Obtaining physical space information of the mine, and constructing a mine knowledge graph according to the obtained physical space information; According to the unified data set, an initial geometric model of the ground surface, geological body, ore body and tunnel is constructed respectively; Mapping and correlating the initial geometric model with the mine knowledge graph, and performing geometric constraint and logical optimization on the initial geometric model according to the semantic rules in the mine knowledge graph to construct a corrected model with semantic information; Integrating the corrected ground surface model, the corrected geological body model, the corrected ore body model and the corrected tunnel model in a unified coordinate system to obtain a data and knowledge fusion mine integrated three-dimensional model.

[0024] In the traditional existing mine three-dimensional modeling technology, the existing method mainly relies on the geometric fitting process of multi-source data such as drilling, geological profile and three-dimensional laser point cloud, and the generated model lacks the expression of the internal logical association between mine geological objects and engineering objects. Among them, the method takes data as a single driving mechanism, and in the area where the data is sparse or there is occlusion, it cannot ensure the correctness of the spatial topological relationship, resulting in phenomena such as model fracture, insertion or suspension, which violates the geological common sense. Further, the geological field knowledge and the geometric modeling process are independent of each other, and the abstract geological rules cannot be converted into quantifiable constraint conditions and directly injected into the modeling process, thereby causing the model to only have appearance characteristics and lack of geological semantic constraints, making it difficult to support fine analysis and dynamic evolution simulation.

[0025] Based on the above problems, the embodiment provides a data and knowledge fusion driven integrated modeling method for a mine, which comprises the following steps: acquiring multi-source heterogeneous spatial data of the mine, and preprocessing the multi-source heterogeneous spatial data to establish a unified data set; in actual operation, the data sources of the mine area are diverse, for example, point position data can be collected by manual measurement, traditional surveying and mapping equipment, or digitized input through historical archives and paper maps. These original data may have problems such as incompatible formats, inconsistent coordinate systems, redundant or missing data, etc. Therefore, the data needs to be preliminarily cleaned and converted to unify the data of different formats into a standard format, or project the data to the same reference system through simple coordinate transformation, thereby forming a preliminarily integrated data set to provide a basis for subsequent modeling.

[0026] Secondly, the method comprises acquiring physical space information of the mine, and constructing a mine knowledge graph according to the acquired physical space information. The physical space information can be acquired through various ways, such as by consulting text data such as geological reports, exploration data, engineering design drawings, or by acquiring domain knowledge through expert interviews, experience summaries, etc. These information can be organized into structured data, and the attribute information of objects such as ore bodies, geological bodies and roadways, and their simple relationships are recorded in table form; based on these structured information, a basic knowledge graph can be constructed, which contains entities and part of their attributes and relationships of mine objects.

[0027] Further, the method comprises constructing initial geometric models of the ground surface, geological bodies, ore bodies and roadways according to the unified data set. In the construction process, various geometric modeling techniques can be used. For the ground surface model, it can be discretized based on ground elevation point data, and a grid regular or triangular irregular network (TIN) model can be generated through interpolation algorithm. For the geological body and ore body model, the stratigraphic boundary points and mineralization information in the drilling data can be combined with the manually drawn profile lines to generate a preliminary three-dimensional body model by using simple cross-section contour line connection or stretching method. For the roadway model, the model can be generated by simple sweeping or extrusion operation based on the measured center line data to generate a pipe-shaped model. These initial models mainly rely on geometric data and may not fully reflect the complex spatial topological relationship.

[0028] On this basis, the method comprises mapping and associating the initial geometric model with the mine knowledge graph, and performing geometric constraint and logical optimization on the initial geometric model according to the semantic rules in the mine knowledge graph, to construct a revised model with semantic information. Each geometric object in the initial geometric model can be assigned an identifier, which is associated with the corresponding semantic entity in the mine knowledge graph. For example, a geometric body is identified as "ore body A", and the semantic information of "ore body A" is found in the knowledge graph. When it is found that the initial geometric model has conditions that do not conform to geological common sense, for example, a roadway model penetrates the ore body boundary, or there is a gap between the geological body models, the model can be adjusted according to the pre-set semantic rules in the knowledge graph. For example, the boundary of the roadway model can be manually adjusted to be inside the ore body model, or the gap between the geological body models can be filled by simple geometric operations. In this way, the initial model is assigned with preliminary semantic information and is logically revised.

[0029] Finally, the method comprises spatial integration of the revised surface correction model, the geological body correction model, the ore body correction model and the roadway correction model in a unified coordinate system, to obtain a mine integrated three-dimensional model with data and knowledge fusion. After the revision of each independent model, it is necessary to ensure that all models are in the same spatial reference framework. All models can be imported into a three-dimensional geographic information system platform, and they can be combined together by simple superposition operation. The integration process integrates each independent revised model into a whole, forming a preliminary mine integrated three-dimensional model. The model fuses geometric data and part of semantic information, and can provide an overall view of the mine area.

[0030] It can be understood that the existing data-driven modeling method, such as geometric fitting relying only on point cloud or borehole data, often generates "models with only geometric shape" lacking geological semantic constraints in the condition of complex geological conditions or sparse data distribution in the deep mining area of the mine area A. These models may have spatial topological errors such as roadway penetrating ore body and geological body fracture, which are contrary to geological common sense, and are difficult to support detailed analysis and intelligent decision-making.

[0031] In contrast, the method of the embodiment formalizes the abstract expression of geological field knowledge into quantifiable semantic rules by introducing the mine knowledge graph. In the above example, even if the initial geometric model has defects due to insufficient data, such as the roadway model penetrating the ore body boundary, the method can use semantic rules such as "roadway must be inside the ore body" in the knowledge graph to perform geometric constraint and logical optimization on the initial model. This knowledge-driven revision mechanism makes the revised model not only accurate in geometry, but also consistent with geological semantics in logic.

[0032] In the present embodiment, the preprocessing of the multi-source heterogeneous spatial data comprises the following steps: The unmanned aerial vehicle images, drilling data, geological profile maps, three-dimensional laser point clouds and CAD data are subjected to coordinate unification, format conversion, redundant data cleaning and attribute field standardization processing. The core geological and engineering element information of the surface, geological body, ore body and roadway is extracted, and the unified data set with consistent data format and spatial reference is established.

[0033] It should be noted that the unmanned aerial vehicle images, drilling data, geological profile maps, three-dimensional laser point clouds and CAD data are subjected to coordinate unification, format conversion, redundant data cleaning and attribute field standardization processing, which aims to solve the inherent inconsistency problem of multi-source heterogeneous data, ensure the unity and standardization of all data on the spatial, format and semantic levels, and lay a solid foundation for subsequent modeling work.

[0034] It should also be noted that the coordinate unification can be achieved by defining a globally unified coordinate system, and using professional geographic information system (GIS) software or measurement software to project the coordinates of data from different sources, perform seven-parameter conversion or four-parameter conversion, and convert all data to the unified coordinate system. Format conversion can use data conversion tools or programming interfaces to convert data of different formats (such as TIFF, JPEG for images, LAS, TXT for point clouds, DWG, DXF for CAD, CSV, Excel for drilling) into a unified and easy-to-process format (such as GeoJSON, Shapefile, database table structure). Redundant data cleaning can identify and delete duplicate or overlapping data points, lines and surfaces through spatial analysis algorithms (such as buffer analysis, overlap detection), or identify and eliminate outliers and noise data through statistical analysis methods. Attribute field standardization processing can rename, merge, split or value domain mapping of attribute fields of different data sources by defining a unified data dictionary and attribute field naming rules, to ensure the semantic consistency of attribute information.

[0035] It can be understood that the core geological and engineering element information of the ground, geological body, ore body and tunnel is extracted, which aims to identify and extract the key information directly related to the integrated modeling of the mine from the preprocessed original data, remove irrelevant or secondary information, and thus focus on the core elements to improve the modeling efficiency and accuracy. The extraction process can be realized by manual interpretation combined with semi-automatic or automatic feature extraction algorithm, for example, for unmanned aerial vehicle image, the ground contour is extracted by using image segmentation, edge detection and other technologies; for drilling data, the geological body and ore body boundary is identified according to lithology and mineralization information; for three-dimensional laser point cloud, the tunnel structure is identified by using point cloud classification algorithm. It can also be realized by establishing a rule-based expert system or a machine learning model, for example, a deep learning model is trained to identify the ground features in the image, or the rules are defined by using the geological expert knowledge to extract the ore body information from the drilling log. Finally, the establishment of the unified data set with consistent data format and spatial reference is the final goal of the foregoing preprocessing and element extraction, which aims to integrate all the standardized processed and core element extracted data into a structured and consistent data set, and provide reliable and high quality input for subsequent geometric modeling. The unified data set can be stored in a relational database (such as PostgreSQL / PostGIS, Oracle Spatial), the unified table structure and spatial index are defined to manage different types of geological and engineering elements, or a file geographic database (such as Esri File Geodatabase) or a cloud storage based geographic spatial data platform is used to organize and store the data of different elements in a unified format (such as Feature Class, Raster Dataset), and ensure that all data share the same spatial reference system.

[0036] It can also be understood that the above-mentioned content is systematically pre-processed by multiple data sources such as unmanned aerial vehicle images, drilling data, geological profiles, three-dimensional laser point clouds and CAD data, firstly, basic standardization processing is carried out, including coordinate unification, ensuring that all data are in the same spatial reference system; format conversion, eliminating compatibility problems caused by different data formats; redundant data cleaning, removing duplicate or incorrect information to improve data quality; attribute field standardization processing, unified semantic expression of attribute information. On this basis, the core geological and engineering element information of the surface, geological body, ore body and tunnel is accurately extracted from these standardized data, and only the information important for subsequent modeling is retained. Finally, through the above series of processing and extraction, a unified data set with completely consistent data format and spatial reference is established. This systematic preprocessing process makes the originally disorganized and difficult to directly use multi-source heterogeneous data into a high-quality, standardized and unified data set. This not only solves the problem of subsequent modeling difficulty caused by data inconsistency, but also provides a reliable and consistent data basis for subsequent construction of the initial geometric model of the surface, geological body, ore body and tunnel based on the unified data set, greatly improves the efficiency and accuracy of modeling, and provides a solid data support for subsequent knowledge graph mapping and semantic optimization.

[0037] In the embodiment, the constructing the mine knowledge graph according to the acquired physical space information comprises the following steps: constructing a three-domain association structure comprising an object domain, an attribute domain and an environment domain; acquiring association rules between the object domain, the attribute domain and the environment domain to establish an entity and relationship sample set; wherein, the object domain comprises underground ore bodies, geological bodies, tunnel structures and surface elements; the attribute domain comprises spatial position, lithology, structural characteristics, grade, geometric size and topographic form; the environment domain comprises ground subsidence, collapse and tailing accumulation environment elements.

[0038] It should be noted that the above steps construct a three-domain association structure including the object domain, the attribute domain and the environment domain, and obtain the association rules between these domains to establish the entity and relationship sample set, thereby constructing a structured and semantically rich mine knowledge graph. This three-domain structure can comprehensively cover the entities, their inherent attributes and external environmental influences in the mine, making the expression of mine knowledge more fine and complete. By clearly defining the underground ore body, geological body, roadway structure and surface elements in the object domain, the spatial position, lithology, structural characteristics, grade, geometric size and landform shape in the attribute domain, and the ground subsidence, collapse and tailings accumulation environment elements in the environment domain, and establishing the association rules between them, the mine knowledge graph can more accurately reflect the complex interactions of mine geology, engineering and environment. This fine knowledge graph provides a solid foundation for subsequent mapping and association of the initial geometric model with the mine knowledge graph, and geometric constraint and logical optimization of the initial geometric model according to the semantic rules in the mine knowledge graph, ensuring the accuracy and consistency of the semantic rules, and thereby improving the reliability of the revised model with semantic information.

[0039] In the present embodiment, the constructing a mine knowledge graph according to the obtained physical space information further comprises the following steps: According to the three-domain association structure of the mine, obtaining the semantic logic between objects; According to the three-domain association structure of the mine, mapping the entity and relationship sample set into the form of RDF triples, and importing into the graph database to generate node and relationship network; The semantic logic at least includes that the ore body contains the roadway, the roadway is located inside the ore body, and the surface is affected by mining.

[0040] It should be noted that the above steps, when constructing the mine knowledge graph, not only establish a sample set of entities and relationships, but also further acquire and formally express the semantic logic between mine elements, and import this semantic information into the graph database in the form of RDF triples, thereby constructing a node and relationship network with deep semantic understanding capabilities. This process upgrades the knowledge graph from a simple data association to an intelligent knowledge reasoning platform. For example, when constructing the initial geometric model of the surface, geological body, ore body, and tunnels based on a unified dataset, this semantically rich knowledge graph can provide accurate semantic rules. These rules, such as "ore body contains tunnels" or "tunnels are located inside ore bodies," can be directly used to geometrically constrain and logically optimize the initial geometric model. The efficient query capabilities of the graph database enable the system to quickly retrieve and apply relevant semantic rules, verifying and adjusting any potential topological conflicts or logical inconsistencies in the initial model until the model meets geological semantic consistency. This deep integration of data and knowledge ensures that the final generated integrated 3D mine model is not only geometrically accurate, but also semantically consistent with the professional knowledge and objective laws of the mining field.

[0041] It's also understandable that the mine knowledge graph is no longer just a simple collection of entities and relationships, but an intelligent knowledge base with deep semantic understanding and reasoning capabilities. This enables precise geometric constraints and logical optimization of the initial geometric model based on explicit semantic rules during the subsequent construction of the integrated 3D mine model, effectively avoiding problems caused by data inconsistencies or model topology errors. Ultimately, this significantly improves the accuracy, consistency, and reliability of the integrated 3D mine model, providing a more solid data foundation and decision support for mine planning, production, and safety management.

[0042] In this embodiment, the initial geometric models of the surface, geological bodies, ore bodies, and tunnels constructed based on the unified dataset include the following steps: A tunnel model is constructed by integrating the 3D point cloud and CAD data from the unified dataset. Based on borehole lithology data in a unified dataset, and by integrating stratigraphic and structural information interpreted from geological profiles, a three-dimensional geological model is constructed. Based on the UAV imagery in the unified dataset, a surface model is generated through aerial triangulation, image matching, and color calibration. Based on borehole data in a unified dataset, the orebody boundary is determined through geological interpretation, and the top and bottom surfaces of the orebody are constructed using a triangular mesh modeling method to build a three-dimensional solid model of the orebody.

[0043] It can be understood that in the initial geometric model stage of building the integrated three-dimensional model of the mine, differentiated modeling strategies are adopted according to the characteristics and data sources of different geological elements. For the roadway model, due to its clear engineering design and complex spatial form, by fusing high-precision three-dimensional point cloud data and structured CAD design data, the authenticity of point cloud and the accuracy of CAD can be fully utilized to avoid the limitations of a single data source, thereby building a roadway model that is consistent with reality and has good geometric quality. For the ore body model, it mainly relies on discrete drilling data and two-dimensional geological profile obtained by underground exploration. These data are essentially discrete, linear or two-dimensional, so the triangular mesh modeling method is adopted to effectively convert these discrete information into continuous three-dimensional surfaces, and the irregular triangular network is formed by connecting discrete points, thereby realistically expressing the three-dimensional form and spatial distribution of the ore body. For the surface model, considering its large range and high detail, aerial triangulation, image matching and color calibration are performed using unmanned aerial images to efficiently obtain high-precision three-dimensional geometry and texture information of the surface. The three-dimensional information is recovered from the two-dimensional images through photogrammetry principles, and realistic visual effects are given. This targeted modeling method enables each initial geometric model to achieve high accuracy and realism under its own data constraints, laying a solid foundation for subsequent semantic fusion and optimization, effectively solving the problem that a single modeling method cannot adapt to multi-source heterogeneous data and complex geological elements, ensuring the quality of the initial model, and further improving the accuracy and reliability of the entire integrated three-dimensional model of the mine.

[0044] In the present embodiment, when constructing the surface model, the distance power inverse ratio method is used to grid estimate the surface sampling points to generate a smooth surface model.

[0045] When constructing the surface model, after obtaining the surface sampling points, the method estimates the grid of these surface sampling points by using the distance power inverse ratio method. Specifically, for each grid point in the smooth surface model to be generated, the system calculates the contribution weight of each sampling point to the grid point elevation value according to the distance between the grid point and the surrounding known surface sampling points, and combines the preset power parameter. The closer the sampling point, the greater the weight, and the greater the influence on the grid point elevation value. By weighting and averaging the elevation values of all related sampling points, the elevation value of the grid point can be estimated. Repeat the process to assign values to all grid points, and finally form a continuous and smooth grid elevation data, thereby constructing a smooth surface model. This processing method ensures the continuity and visual realism of the surface model, so that it can be more accurately integrated into the integrated three-dimensional model of the mine, and the corrected geological body correction model, the ore body correction model and the roadway correction model are integrated in the unified coordinate system to form the data and knowledge integrated three-dimensional model of the mine.

[0046] whose expression satisfies: ; ; wherein, represents the elevation of the prediction point, represents the elevation measured at the sample point, represents the spatial distance between the interpolation point and the sample point, is the power value (usually assumed to be 2), representing the number of interpolation points, represents the weight of each elevation. In this embodiment, the step of performing geometric constraint and logical optimization on the initial geometric model according to the semantic rules in the mine knowledge graph comprises the following steps: checking the spatial discontinuity of the data sparse area in the initial geometric model according to the spatial topology, spatial distribution and geometric shape rules defined in the mine knowledge graph; when a conflict between the topological relationship of the initial geometric model and the semantic rules is detected, adjusting the geometric shape or spatial position of the initial geometric model according to the semantic rules until the geological semantic consistency is met.

[0047] The system uses the spatial topology, spatial distribution and geometric shape rules defined in the knowledge graph to comprehensively check the initial geometric model, and pays special attention to the spatial discontinuity of the data sparse area to ensure that the geometric representation conforms to the basic geological principles. At the same time, the system detects the conflict between the topological relationship of the initial geometric model and the semantic rules in the knowledge graph. For example, when a part of an initial tunnel model extends out of the ore body it belongs to in geometry, while the knowledge graph clearly stipulates that "tunnel must be contained within the ore body", a conflict will be identified. Once such a conflict is detected, the system will adjust the geometric shape or spatial position of the initial geometric model according to the semantic rules in the knowledge graph. This adjustment process is iterative until the model meets the geological semantic consistency, that is, the geometric model not only becomes reasonable in geometry, but also accurately reflects the geological and engineering reality encoded in the knowledge graph. This data and knowledge fusion modeling method overcomes the limitations of purely data-driven modeling in complex geological environments, ensuring the accuracy and reliability of the final revised model.

[0048] In this embodiment, the method further comprises the following steps: obtaining the unique identifier mapping between the geometric objects in the mine integrated three-dimensional model and the semantic entity nodes in the mine knowledge graph; When the change of object attribute or environment state in the mine knowledge graph is monitored, the geometric morphology of the corresponding object in the mine integrated three-dimensional model is synchronously adjusted through the unique identification mapping; when the data of the mine integrated three-dimensional model is updated, the new object or new attribute is connected to the mine knowledge graph through the incremental updating mechanism.

[0049] The above method establishes a unique identification mapping between the geometric object in the mine integrated three-dimensional model and the semantic entity in the mine knowledge graph, and constructs a bidirectional synchronization mechanism. When the object attribute or environment state in the mine knowledge graph changes, for example, the lithology of the geological body is updated or the ground subsidence data changes, the system can quickly locate the corresponding geometric object in the three-dimensional model through the pre-established unique identification mapping, and drive the geometric morphology to be synchronously adjusted, so as to ensure that the three-dimensional model can reflect the latest semantic information in real time. Conversely, when the data of the mine integrated three-dimensional model is updated, for example, a new tunnel model is added or the boundary of the ore body model is adjusted, the incremental updating mechanism can identify these changes and connect the new geometric object or attribute information to the mine knowledge graph, so as to maintain the integrity and timeliness of the knowledge graph. This bidirectional and dynamic synchronization mechanism makes the mine integrated three-dimensional model no longer a static geometric representation, but a digital twin closely coupled with the mine knowledge graph and evolving in real time, greatly improving the practical value and decision support capability of the model.

[0050] In the embodiment, the semantic rules are specifically configured as: For the tunnel object, a spatial semantic constraint is set, which includes limiting the tunnel model to be located within the internal boundary of the ore body model or the geological body model, and the extension direction of the tunnel model matches the structural feature trend of the ore body; For the surface object, an environment dynamic constraint is set, which associates and constrains the deformation parameters of the surface model with the ground subsidence environment element in the mine knowledge graph; For the geological body object, a geological semantic constraint is set, the geological body model matches the drilling data, and is consistent with the geological map and the interpreted profile in space; For the ore body object, a geological and engineering semantic constraint is set, the ore body model matches the drilling data and the lithology data.

[0051] The unified dataset is established by acquiring and preprocessing multi-source heterogeneous spatial data of the mine, and the physical space information of the mine is acquired and the mine knowledge graph is constructed. On this basis, the initial geometric models of the ground, geological body, ore body and tunnel are constructed according to the unified dataset. Then, the initial geometric models are mapped and associated with the mine knowledge graph, and the initial geometric models are geometrically constrained and logically optimized according to the specific semantic rules configured in the mine knowledge graph. Specifically, for the tunnel object, the system will apply the preset spatial semantic constraints, check whether the initial tunnel model is completely located within the internal boundary of the associated ore body model or geological body model, and verify whether the extension direction of the tunnel model matches the structure feature trend of the ore body. If there is inconsistency, the system will automatically adjust the geometric shape or spatial position of the tunnel model until the strict spatial topological and geological consistency requirements are met. At the same time, for the ground object, the system will apply the environmental dynamic constraints to associate the deformation parameters of the ground model with the ground subsidence environmental elements recorded in the mine knowledge graph. This means that once the ground subsidence environmental elements change (for example, the subsidence value in the knowledge graph is updated through external monitoring data), the deformation parameters of the ground model will be automatically updated, thereby driving the ground model to make corresponding geometric adjustments to reflect the ground deformation in real time. Finally, the corrected ground model, the corrected geological body model, the corrected ore body model and the corrected tunnel model are spatially integrated in the unified coordinate system to obtain the integrated three-dimensional model of the mine with data and knowledge fusion. The fine semantic rule configuration enables the integrated three-dimensional model of the mine to have higher geometric precision, geological rationality and environmental dynamic response capability in key elements. Embodiment

[0052] As shown in the accompanying drawings, Figure 2 The embodiment provides a modeling system, which comprises: A data preprocessing module is configured to acquire multi-source heterogeneous spatial data of a mine, and preprocess the multi-source heterogeneous spatial data to establish a unified dataset. A knowledge graph construction module is configured to acquire physical space information of the mine, and construct a mine knowledge graph according to the acquired physical space information. A geometric modeling module is configured to construct initial geometric models of the ground, geological body, ore body and tunnel according to the unified dataset. A semantic fusion and optimization module is configured to map and associate the initial geometric models with the mine knowledge graph, and perform geometric constraint and logical optimization on the initial geometric models according to the semantic rules in the mine knowledge graph to construct corrected models with semantic information. An integrated expression module is used to spatially integrate the corrected surface correction model, the geological body correction model, the ore body correction model and the roadway correction model under a unified coordinate system to obtain a data and knowledge fused mine integrated three-dimensional model.

[0053] By obtaining multi-source heterogeneous spatial data of the mine and preprocessing to establish a unified data set, physical space information of the mine is obtained and a mine knowledge graph is constructed. On this basis, initial geometric models of the surface, geological body, ore body and roadway are constructed according to the unified data set. Then, the initial geometric models are mapped and associated with the mine knowledge graph, and the initial geometric models are geometrically constrained and logically optimized according to the specific semantic rules configured in the mine knowledge graph. Specifically, for the roadway object, the system will apply the preset spatial semantic constraints, check whether the initial roadway model is completely located within the internal boundary of the associated ore body model or geological body model, and verify whether the extension direction matches the structure feature trend of the ore body. If there is inconsistency, the system will automatically adjust the geometric shape or spatial position of the roadway model until the strict spatial topological and geological consistency requirements are met. At the same time, for the surface object, the system will apply the environmental dynamic constraints to associate the deformation parameters of the surface model with the ground subsidence environmental elements recorded in the mine knowledge graph. This means that once the ground subsidence environmental elements change (for example, the subsidence value in the knowledge graph is updated through external monitoring data), the deformation parameters of the surface model will be automatically updated, thereby driving the surface model to make corresponding geometric adjustments to reflect the surface deformation in real time. Finally, the corrected surface correction model, the geological body correction model, the ore body correction model and the roadway correction model are spatially integrated under a unified coordinate system to obtain a data and knowledge fused mine integrated three-dimensional model. The fine semantic rule configuration enables the mine integrated three-dimensional model to have higher geometric accuracy, geological rationality and environmental dynamic response capability on key elements.

[0054] It should be noted that in this document, the terms "comprising", "including", or any other variant thereof are intended to cover non-exclusive inclusions, such that processes, methods, articles or systems that include a series of elements not only include those elements, but also include other elements not explicitly listed, or inherent to such processes, methods, articles or systems. Without more limitations, the element defined by the statement "comprising a" does not exclude the presence of another identical element in the process, method, article or system that includes the element.

[0055] The above-mentioned embodiment numbers of the present application are only for description, and do not represent the advantages and disadvantages of the embodiments.

[0056] Through the above description of the embodiments, those skilled in the art can clearly understand that the above-mentioned example method can be realized by means of software and a necessary general hardware platform, and of course, it can also be realized by hardware, but in many cases, the former is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory / random access memory, a magnetic disk, or an optical disk) and includes a plurality of instructions for causing a multimedia terminal device (which can be a mobile phone, a computer, a television receiver, or a network device, etc.) to execute the method described in each embodiment of the present application.

[0057] The above is only a preferred embodiment of the present application, and does not limit the patent scope of the present application, and any equivalent structure or equivalent flow transformation made by using the content of the specification and drawings of the present application, or directly or indirectly applied to other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A data and knowledge-driven integrated mine modeling method, characterized in that, The method includes the following steps: Acquire multi-source heterogeneous spatial data of the mine, preprocess the multi-source heterogeneous spatial data to establish a unified dataset; Obtain the physical spatial information of the mine, and construct a mine knowledge graph based on the obtained physical spatial information; Initial geometric models of the surface, geological bodies, ore bodies, and tunnels are constructed based on the unified dataset. The initial geometric model is mapped and associated with the mine knowledge graph. Based on the semantic rules in the mine knowledge graph, the initial geometric model is subjected to geometric constraints and logical optimization to construct a modified model with semantic information. The corrected surface model, geological body model, ore body model, and tunnel model are spatially integrated in a unified coordinate system to obtain an integrated 3D mine model that combines data and knowledge.

2. The data and knowledge fusion-driven integrated mine modeling method as described in claim 1, characterized in that, The preprocessing of multi-source heterogeneous spatial data includes the following steps: Coordinate unification, format conversion, redundant data cleaning, and attribute field standardization are performed on UAV imagery, borehole data, geological profile maps, 3D laser point clouds, and CAD data. Extract core geological and engineering element information from the surface, geological bodies, ore bodies, and tunnels, and establish a unified dataset with a data format consistent with the spatial benchmark.

3. The data and knowledge fusion-driven integrated mine modeling method as described in claim 1, characterized in that, The construction of the mine knowledge graph based on the acquired physical spatial information includes the following steps: Construct a three-domain association structure that includes object domain, attribute domain, and environment domain; Obtain the association rules between the object domain, attribute domain, and environment domain to establish a sample set of entities and relationships; in, The object domain includes elements such as underground ore bodies, geological bodies, tunnel structures, and the surface. The attribute domain includes spatial location, lithology, structural features, grade, geometric dimensions, and landform morphology, etc. The environmental domain includes environmental elements such as ground subsidence, collapse, and tailings accumulation.

4. The data and knowledge fusion-driven integrated mine modeling method as described in claim 3, characterized in that, The construction of the mine knowledge graph based on the acquired physical spatial information also includes the following steps: Based on the three-domain association structure of the mine, obtain the semantic logic between objects; Based on the three-domain association structure of the mine, the entity and relation sample set is mapped into RDF triple form and imported into the graph database to generate a node and relation network; The semantic logic includes at least the following: the ore body contains tunnels, the tunnels are located inside the ore body, and the surface is affected by mining.

5. The data and knowledge fusion-driven integrated mine modeling method as described in claim 1, characterized in that, The initial geometric models of the surface, geological bodies, ore bodies, and tunnels constructed based on the unified dataset include the following steps: A tunnel model is constructed by integrating the 3D point cloud and CAD data from the unified dataset. Based on borehole lithology data in a unified dataset, and by integrating stratigraphic and structural information interpreted from geological profiles, a three-dimensional geological model is constructed. Based on the UAV imagery in the unified dataset, a surface model is generated through aerial triangulation, image matching, and color calibration. Based on borehole data in a unified dataset, the orebody boundary is determined through geological interpretation, and the top and bottom surfaces of the orebody are constructed using a triangular mesh modeling method to build a three-dimensional solid model of the orebody.

6. The data and knowledge fusion-driven integrated mine modeling method as described in claim 5, characterized in that, When constructing the surface model, the inverse power distance method is used to perform grid estimation of the surface sampling points to generate a smooth surface model.

7. The data and knowledge fusion-driven integrated mine modeling method as described in claim 1, characterized in that, The geometric constraint and logical optimization of the initial geometric model based on the semantic rules in the mine knowledge graph includes the following steps: Based on the spatial topology, spatial distribution, and geometric shape rules defined in the mine knowledge graph, the spatial discontinuity of the sparse data region in the initial geometric model is verified. When a conflict is detected between the topological relationship of the initial geometric model and the semantic rules, the geometric shape or spatial position of the initial geometric model is adjusted according to the semantic rules until the geological semantic consistency is satisfied.

8. The data and knowledge fusion-driven integrated mine modeling method as described in claim 1, characterized in that, The method further includes the following steps: Obtain the unique identifier mapping between geometric objects in the integrated 3D model of the mine and semantic entity nodes in the mine knowledge graph; When changes are detected in the object attributes or environmental state of the mine knowledge graph, the geometry of the corresponding object in the integrated 3D mine model is adjusted synchronously through unique identifier mapping; when the data of the integrated 3D mine model is updated, the new object or new attribute is added to the mine knowledge graph through an incremental update mechanism.

9. The data and knowledge fusion-driven integrated mine modeling method as described in claim 1, characterized in that, The semantic rules are specifically configured as follows: For the tunnel object, spatial semantic constraints are set, including restricting the tunnel model to be located within the internal boundary of the ore body model or geological body model, and the extension direction of the tunnel model to match the structural features of the ore body. For surface objects, set dynamic environmental constraints and associate the deformation parameters of the surface model with the ground subsidence environmental elements in the mine knowledge graph. For geological bodies, geological semantic constraints are set, the geological body model is matched with borehole data, and it is spatially consistent with geological maps and interpretation profiles; For orebody objects, geological and engineering semantic constraints are set, and the orebody model is matched with borehole data and lithological data.

10. A modeling system, characterized in that, The system is based on the data and knowledge fusion-driven integrated mine modeling method as described in any one of claims 1 to 9, and the system includes: The data preprocessing module is used to acquire multi-source heterogeneous spatial data of the mine and preprocess the multi-source heterogeneous spatial data to establish a unified dataset. A knowledge graph construction module is used to acquire the physical spatial information of the mine and construct a mine knowledge graph based on the acquired physical spatial information. A geometric modeling module is used to construct initial geometric models of the surface, geological bodies, ore bodies, and tunnels based on the unified dataset. The semantic fusion and optimization module is used to map and associate the initial geometric model with the mine knowledge graph, and to perform geometric constraints and logical optimization on the initial geometric model according to the semantic rules in the mine knowledge graph, so as to construct a modified model with semantic information. An integrated expression module is used to spatially integrate the corrected surface correction model, geological body correction model, ore body correction model and roadway correction model in a unified coordinate system to obtain an integrated three-dimensional mine model that combines data and knowledge.

Images