A multi-modal data-driven open-pit mine true three-dimensional model construction method

Abstract

This invention discloses a method and apparatus for constructing a true 3D model of an open-pit mine driven by multimodal data. The method and apparatus involve processing and fusing multi-source data from the open-pit mine to obtain a multimodal information database. Based on the multimodal data in the database, a single 3D model of the open-pit mine is constructed. A quadtree-based spatial partitioning method and preset mesh coding rules are used to obtain the mesh code corresponding to the open-pit mine area. A mapping relationship is established between the mesh code and the model entities in the single 3D model of the open-pit mine, resulting in a mesh-coded true 3D model of the open-pit mine. This invention, based on the mapping relationship between the mesh code corresponding to the open-pit mine area and the model entities in the single 3D model of the open-pit mine, obtains a mesh-coded true 3D model of the open-pit mine, thereby improving the detection accuracy and precision of key information in open-pit mine disasters.

Description

Technical Field

[0001] This invention relates to the field of model building technology, and in particular to a method, apparatus and storage medium for building a true 3D model of an open-pit mine driven by multimodal data. Background Technology

[0002] The rapid development of geological exploration and monitoring technologies has laid a solid data foundation for the digital and intelligent management of open-pit mines. This includes the acquisition of massive amounts of multi-source data through various technologies such as satellite remote sensing, UAV aerial surveying, and ground sensor networks.

[0003] However, in practical applications, it is difficult to conduct effective correlation analysis and collaborative utilization among different modalities in the massive multi-source data mentioned above. This makes it difficult to fully explore the inherent connections between the data and to construct a true 3D model of an open-pit mine that includes multi-physics coupling, thereby reducing the detection accuracy and precision of key information in open-pit mine disasters. Summary of the Invention

[0004] The present invention aims to at least partially solve one of the technical problems in the related art.

[0005] To address this, this invention proposes a method and apparatus for constructing a true 3D model of an open-pit mine driven by multimodal data. This method processes and fuses multi-source data from the open-pit mine to obtain a multimodal information database. Based on the multimodal data in this database, a single 3D model of the open-pit mine is constructed. Using a quadtree-based spatial partitioning method and preset mesh coding rules, the mesh code corresponding to the open-pit mine area is obtained. A mapping relationship is established between the mesh code and the model entities in the single 3D model of the open-pit mine, resulting in a mesh-coded true 3D model of the open-pit mine. Therefore, this invention can obtain the inherent connections within multi-source data through a multimodal information database and, based on the mapping relationship between the mesh code corresponding to the open-pit mine area and the model entities in the single 3D model of the open-pit mine, obtain a mesh-coded true 3D model of the open-pit mine, thereby improving the detection accuracy and precision of key information in open-pit mine disasters.

[0006] To achieve the above objectives, this invention proposes a method for constructing a true 3D model of an open-pit mine driven by multimodal data, the method comprising:

[0007] Collect and gather multi-source data from open-pit mines, and process and fuse the multi-source data to obtain a multimodal information database; Based on the multimodal data in the multimodal information database, a three-dimensional model of an open-pit mine is constructed. The grid code corresponding to the open-pit mine area is obtained by using a quadtree-based spatial partitioning method and a preset grid coding rule; Establish the mapping relationship between the mesh code and the model entities in the single 3D model of the open-pit mine to obtain the true 3D model of the open-pit mine with mesh code.

[0008] The multimodal data-driven true 3D model construction method for open-pit mines according to embodiments of the present invention may also have the following additional technical features: In one embodiment of the present invention, the collection and acquisition of multi-source data from open-pit mines, and the processing and fusion of the multi-source data to obtain a multimodal information database, includes: By collecting and acquiring multi-source data from open-pit mines, long-term multimodal data and monitoring data can be obtained. The long-term multimodal data and the monitoring data are subjected to data cleaning, missing data interpolation, format conversion and standardization to obtain the first multimodal data; Feature extraction is performed on the first multimodal data, and the extracted features are fused using a data fusion model to obtain the second multimodal data; Based on the second multimodal data, a hybrid database is constructed through a distributed architecture, a data warehouse architecture, and a cloud database architecture.

[0009] In one embodiment of the present invention, constructing a single three-dimensional model of an open-pit mine based on the multimodal data in the multimodal information database includes: The multimodal information database is used to acquire satellite and UAV remote sensing images; Three-dimensional reconstruction and texture mapping are performed on the satellite and UAV remote sensing images to generate digital elevation models and digital surface models; Based on the digital elevation model and the digital surface model, the stress field distribution inside the open-pit mine is deduced; Based on the stress field distribution inside the open-pit mine, a three-dimensional model of the open-pit mine is generated using parametric modeling technology.

[0010] In one embodiment of the present invention, the quadtree-based space partitioning method includes: Based on the actual extent and required accuracy of the slope in the open-pit mine area, determine the initial parameters and maximum number of levels for quadtree partitioning; Based on the initial parameters and the maximum number of levels, the slope area is divided into grids using a quadtree algorithm, and a location index is generated for each grid.

[0011] In one embodiment of the present invention, the preset grid encoding rule includes: The grid coding structure includes a first fixed-length code, a second fixed-length code, and a variable-length code. The first fixed-length code is located at the beginning of the grid coding and is used to represent grid location information. The variable-length code is located in the middle of the grid coding and is used to represent geology. The second fixed-length code is located at the end of the grid coding and is used to represent disaster risk level.

[0012] In one embodiment of the present invention, the method further includes: Based on the true 3D model of the open-pit mine and the multimodal data in the multimodal information database, a fusion analysis is performed to obtain key indicator data in the disaster scenario through a disaster scenario simulation algorithm; The true 3D model of the open-pit mine is verified by comparing and correcting historical data and key indicator data. Based on the verification results, the true 3D model of the open-pit mine is optimized to obtain the target true 3D model of the open-pit mine.

[0013] To achieve the above objectives, another aspect of the present invention proposes a multimodal data-driven true 3D model construction device for open-pit mines, the device comprising: The first construction module is used to collect and gather multi-source data from open-pit mines, and to process and fuse the multi-source data to obtain a multimodal information database. The second construction module is used to construct a single three-dimensional model of the open-pit mine based on the multimodal data in the multimodal information database. The encoding module is used to obtain the grid code corresponding to the open-pit mine area by using a quadtree-based spatial partitioning method and preset grid encoding rules; The third construction module is used to establish the mapping relationship between the mesh code and the model entities in the single 3D model of the open-pit mine, so as to obtain the mesh-coded true 3D model of the open-pit mine.

[0014] The multimodal data-driven method and apparatus for constructing true 3D models of open-pit mines, as described in this invention, can process and fuse multi-source data from open-pit mines to obtain a multimodal information database. Based on the multimodal data in the database, a single 3D model of the open-pit mine is constructed. Using a quadtree-based spatial partitioning method and preset mesh coding rules, the mesh code corresponding to the open-pit mine area is obtained. A mapping relationship is established between the mesh code and the model entities in the single 3D model of the open-pit mine, resulting in a mesh-coded true 3D model of the open-pit mine. Therefore, this invention can obtain the inherent connections within multi-source data through a multimodal information database and, based on the mapping relationship between the mesh code corresponding to the open-pit mine area and the model entities in the single 3D model of the open-pit mine, obtain a mesh-coded true 3D model of the open-pit mine, thereby improving the detection accuracy and precision of key information in open-pit mine disasters.

[0015] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0016] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart illustrating a method for constructing a true 3D model of an open-pit mine driven by multimodal data according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the structure of a true three-dimensional model of an open-pit mine according to an embodiment of the present invention; Figure 3 This is a schematic diagram of a multimodal data-driven true 3D model construction device for open-pit mines according to an embodiment of the present invention. Detailed Implementation

[0017] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0018] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0019] The following description, with reference to the accompanying drawings, describes a method and apparatus for constructing a true 3D model of an open-pit mine driven by multimodal data, according to an embodiment of the present invention.

[0020] Figure 1 This is a flowchart illustrating the multimodal data-driven method for constructing a true 3D model of an open-pit mine, according to an embodiment of the present invention.

[0021] like Figure 1 As shown, the method may include the following steps: Step 101: Collect and gather multi-source data from the open-pit mine, and process and fuse the multi-source data to obtain a multimodal information database.

[0022] In one embodiment of the present invention, the method for collecting and acquiring multi-source data from open-pit mines, and processing and fusing the multi-source data to obtain a multimodal information database may include the following steps: Step 1011: Collect and gather multi-source data from open-pit mines to obtain long-term multimodal data and monitoring data.

[0023] In one embodiment of the present invention, long-term time-series multimodal data and monitoring data can be acquired using both collection and acquisition methods, based on differences in data source, timing, and format. Table 2 is a data acquisition relationship table proposed in an embodiment of the present invention.

[0024] Table 2

[0025] As shown in Table 2, long-term time-series multimodal data and monitoring data can be obtained through collection and acquisition.

[0026] Step 1012: Perform data cleaning, missing data interpolation, format conversion and standardization on the long-term multimodal data and monitoring data to obtain the first multimodal data.

[0027] In one embodiment of the present invention, after obtaining long-term multimodal data and monitoring data through the above steps, the characteristics of various types of data can be identified according to the source, type and format of the obtained long-term multimodal data and monitoring data. Corresponding data cleaning techniques are then used to remove obvious errors and abnormal data, ensuring that the accuracy rate after data cleaning reaches more than 95%.

[0028] In one embodiment of the present invention, the K-nearest neighbor interpolation method is used to intelligently fill in missing data based on the distribution characteristics of existing data, ensuring that the interpolation error range is controlled within 5%.

[0029] In one embodiment of the present invention, different data are converted according to a unified format requirement (such as SQL format) to ensure the accuracy and integrity of the data.

[0030] In one embodiment of the present invention, a unified core semantic model and data dictionary are constructed based on international / national standards to standardize the definition, concepts, units and precision of data, ensuring that all data standards are unified.

[0031] In one embodiment of the present invention, after performing data cleaning, missing data interpolation, format conversion and standardization on long-time multimodal data and monitoring data through the above steps, the first multimodal data can be obtained.

[0032] Step 1013: Extract features from the first multimodal data and fuse the extracted features using a data fusion model to obtain the second multimodal data.

[0033] In one embodiment of the present invention, after obtaining the first multimodal data through the above steps, feature extraction can be performed on the first multimodal data, and the extracted features can be fused through a data fusion model to obtain the second multimodal data.

[0034] In one embodiment of the present invention, the first multimodal data is subjected to feature analysis of its dimensionality, distribution patterns, and correlations. Feature selection algorithms, combined with domain expert experience, are used to extract feature data closely related to open-pit mine disasters. In one embodiment of the present invention, the feature selection algorithm may include random forest feature importance and recursive feature elimination algorithms.

[0035] In one embodiment of the present invention, a multimodal data fusion model constructed using a Bayesian network is used to fuse the extracted feature data to obtain a second multimodal data.

[0036] In one embodiment of the present invention, the extracted feature data can be input into a multimodal data fusion model employing a Bayesian neural network (BNN) to obtain second multimodal data, wherein a dynamic attention mechanism is embedded in the multimodal data fusion model. Furthermore, the adjustment rule for the attention weights in the dynamic attention mechanism can dynamically allocate the importance ratio of each modality based on real-time operating condition features. In one embodiment of the present invention, the aforementioned real-time operating condition features may include equipment load rate and mining progress.

[0037] Step 1014: Based on the second multimodal data, construct a hybrid database through a distributed architecture, a data warehouse architecture, and a cloud database architecture.

[0038] In one embodiment of the present invention, after obtaining the second multimodal data through the above steps, a hybrid database can be constructed based on the second multimodal data through a distributed architecture, a data warehouse architecture, and a cloud database architecture.

[0039] In one embodiment of the present invention, a hybrid database architecture combining distributed architecture, data warehouse architecture, and cloud database architecture can be constructed by comprehensively considering factors such as the scale, read / write requirements, storage costs, and security of multimodal data from open-pit mines. Furthermore, table structures, field types, indexes, etc., in the database are defined, and data storage strategies are formulated to store the processed and integrated data in the database, achieving efficient data storage, querying, and daily management. In one embodiment of the present invention, the aforementioned data storage strategy can be: adopting a hybrid storage architecture design principle, using a distributed architecture (big data processing layer) for massive storage and batch processing of raw multimodal data; using a data warehouse architecture (analysis and query layer) to conduct OLAP analysis of integrated feature data and historical data mining; and using a cloud database architecture (business service layer) to realize real-time data services, transaction processing, metadata management, etc.

[0040] Step 102: Construct a single three-dimensional model of the open-pit mine based on the multimodal data in the multimodal information database.

[0041] In one embodiment of the present invention, after obtaining the multimodal information database through the above steps, a single three-dimensional model of an open-pit mine can be constructed based on the multimodal data in the multimodal information database.

[0042] Specifically, in one embodiment of the present invention, the method for constructing a single three-dimensional model of an open-pit mine based on multimodal data in a multimodal information database may include the following steps: Step 1021: Obtain satellite and UAV remote sensing images through a multimodal information database.

[0043] In one embodiment of the present invention, satellite remote sensing images with a resolution better than 1m are acquired through a multimodal information database for macroscopic analysis of terrain, landforms and land features; and UAV remote sensing images with a resolution better than 0.2m are acquired for extraction of local detailed features of open-pit mines.

[0044] Step 1022: Perform 3D reconstruction and texture mapping on satellite and UAV remote sensing images to generate digital elevation models and digital surface models.

[0045] In one embodiment of the present invention, after obtaining satellite and UAV remote sensing images through the above steps, image correction, image fusion and image cropping can be performed on the satellite and UAV remote sensing images in sequence to obtain processed satellite and UAV remote sensing images.

[0046] In one embodiment of the present invention, after obtaining the processed satellite and UAV remote sensing images through the above steps, realistic mining area DEM (Digital Elevation Model) and DSM (Digital Surface Model) can be generated through aerial triangulation, point cloud generation, point cloud filtering, 3D reconstruction and texture mapping.

[0047] In one embodiment of the present invention, after obtaining the DEM and DSM through the above steps, the terrain and geological data can be processed and spatially analyzed. Through coordinate transformation and projection settings, the coordinate systems of satellite remote sensing images, UAV remote sensing images, DEM, DSM and GIS data are consistent, avoiding model misalignment.

[0048] Step 1023: Based on the digital elevation model and digital surface model, the stress field distribution inside the open-pit mine is deduced.

[0049] In one embodiment of the present invention, after obtaining the digital elevation model and digital surface model through the above steps, the stress field distribution inside the open-pit mine can be deduced based on the digital elevation model and digital surface model.

[0050] In one embodiment of the present invention, the method for deriving the internal stress field distribution of an open-pit mine based on a digital elevation model and a digital surface model may include the following steps: Step 1: Based on the digital elevation model and digital surface model, obtain the surface structure trace information of the slope through image recognition technology, deduce the distribution information of the internal structural surface of the open-pit mine, and obtain supplementary geological structure data.

[0051] In one embodiment of the present invention, based on a digital elevation model (DEM) and a digital surface model (DSM), image recognition algorithms corresponding to edge detection or linear feature extraction can be used to automatically identify the trace information of slope surface structure surfaces. The trace information of slope surface structure surfaces may include trace distribution, attitude, and spacing.

[0052] Furthermore, in one embodiment of the present invention, by using the information of the surface structural traces of the slope, combined with the extensibility and conjugate relationship of the structural surface, and using three-dimensional interpolation and spatial statistical analysis, the extension trend and spatial distribution of the structural surface inside the rock mass can be deduced, a supplementary three-dimensional network model of the geological structural surface can be generated, and supplementary geological structural data can be obtained based on the supplementary three-dimensional network model of the geological structural surface.

[0053] Step 2: Based on supplementary geological structure data, combined with exploration data, measurement data and monitoring data, a similar physical model and numerical model of the open-pit mine are established through engineering layout and big data analysis technology.

[0054] In one embodiment of the present invention, based on the supplementary geological structure data obtained in step 1 above, combined with borehole exploration data, surface measurement data (such as GNSS, laser scanning) and monitoring data (such as slope displacement, stress monitoring), a three-dimensional solid model of the rock mass structure is constructed through geological statistical analysis and data fusion technology.

[0055] In one embodiment of the present invention, a physical model can be established based on a three-dimensional solid model of the rock mass structure and similarity theories (such as scaled-down geometric and mechanical parameters). Furthermore, a corresponding numerical model (such as a finite element model or discrete element model) is generated using a rock mass constitutive model (such as Mohr-Coulomb or Hawke-Brown) and the discrete fracture network (DFN) method, and the model parameters are calibrated.

[0056] Step 3: Based on the similar physical model and numerical model of the open-pit mine, the stress field distribution inside the open-pit mine is deduced through model experiments and numerical experiments.

[0057] In one embodiment of the present invention, a simulated load (such as gravity or excavation unloading) is applied to a physical model similar to an open-pit mine, and the stress distribution is measured by strain gauges and stress sensors to obtain physical test results; simultaneously, the same boundary conditions and load conditions are set in a numerical model to perform numerical simulation calculations to obtain numerical test results.

[0058] In one embodiment of the present invention, the physical test results are compared with the numerical test results, and the distribution of the internal stress field of the rock mass (such as the magnitude and direction of the principal stress) before and after the mine excavation is deduced through parameter inversion and iterative optimization.

[0059] Step 1024: Based on the stress field distribution inside the open-pit mine, a three-dimensional model of the open-pit mine is generated using parametric modeling technology.

[0060] In one embodiment of the present invention, after obtaining the internal stress field distribution of the open-pit mine through the above steps, a three-dimensional model of the open-pit mine can be generated based on the internal stress field distribution using parametric modeling technology.

[0061] In one embodiment of the present invention, a single three-dimensional model of the open-pit mine can be generated based on the stress field distribution inside the mine and with reference to the generation method of BIM models, using parametric modeling technology. Specifically, in one embodiment of the present invention, the stress field distribution inside the open-pit mine can be superimposed onto the model using remote sensing imagery and GIS data. Through model association and data binding operations, deep integration and visualization of multimodal data can be achieved, intuitively displaying the three-dimensional structure, topography, geological structure, stress and strain information of the open-pit mine.

[0062] Furthermore, in one embodiment of the present invention, detailed attribute information can be assigned to each element in the model, such as geological attributes (rock type, stratigraphic age, etc.) and mechanical parameters (elastic modulus, Poisson's ratio, etc.).

[0063] Furthermore, in one embodiment of the present invention, a first data access interface can be developed, which utilizes Internet of Things (IoT) technology and data transmission protocols to transmit "air-space-ground" monitoring data and meteorological data in real time to the single three-dimensional model of the open-pit mine through a wireless sensor network, thereby realizing real-time updates of slope status information.

[0064] Step 103: Using a quadtree-based spatial partitioning method and preset grid coding rules, obtain the grid code corresponding to the open-pit mine area.

[0065] In one embodiment of the present invention, the above-described quadtree-based space partitioning method may include the following steps: Step 1031: Determine the initial parameters and maximum number of levels for quadtree partitioning based on the actual extent and required accuracy of the slope in the open-pit mine area. In one embodiment of the present invention, the extreme values ​​of the three-dimensional coordinates of the slope area are obtained through GNSS measurement, laser scanning, or a three-dimensional geological model, namely the maximum and minimum values ​​of the X-axis (strike), Y-axis (dip), and Z-axis (elevation), to determine the actual area to be covered as [X...]. min X max ]×[Y min Y max ]×[Z min Z max ].

[0066] In one embodiment of the present invention, a minimum grid size threshold is obtained for different regions. Also, in one embodiment of the present invention, using the extreme values ​​of the slope spatial coordinates obtained from the actual range as boundaries, a minimum rectangular (two-dimensional) bounding box that can completely cover the entire slope region is generated as the root node spatial domain of the quadtree; and the initial grid size is calculated, for example, for a two-dimensional quadtree, the initial grid size D0 = max(X... max -X min Y max -Y min In one embodiment of the present invention, the initial parameters may include an initial grid size and a minimum grid size threshold for different regions.

[0067] Furthermore, in one embodiment of the present invention, the splitting trigger condition of the root node is set to trigger the quadtree split (one parent grid is divided into four child grids, and the size of the child grid is 1 / 2 of the parent grid) when the accuracy requirement of the grid coverage area is higher than the current grid size.

[0068] In one embodiment of the present invention, the minimum grid size threshold for different regions can be converted into a grid size threshold table, and the maximum allowed grid size can be marked for each partition (i.e., the grid size of the partition must not exceed this value). A unified rule is set, if the threshold for a certain region is D... th Then the quadtree mesh size in this region must satisfy D≤D th Otherwise, it will need to continue splitting.

[0069] In one embodiment of the present invention, the relationship between the quadtree levels and the grid size can be as follows: each time the quadtree splits, the level number k increases by 1, and the grid size becomes half of the previous level, that is, the grid size of the k-th level is... Furthermore, in one embodiment of the present invention, the maximum number of levels is calculated as follows: using the minimum grid size d min Constraints, let D k =d min Derive the maximum number of levels ,in This is a rounding function (ensuring the grid size is no larger than the minimum precision threshold).

[0070] Step 1032: Based on the initial parameters and the maximum number of levels, the slope area is divided into grids using the quadtree algorithm, and a location index is generated for each grid.

[0071] In one embodiment of the present invention, each of the above-mentioned grid levels has a clear spatial range and location identifier. The lower-level grids inherit the location information of the upper-level grids, and the grids do not overlap with each other, thus comprehensively and orderly covering the entire slope area.

[0072] In one embodiment of the present invention, the aforementioned preset grid encoding rule may include: The grid coding structure includes a first fixed-length code, a second fixed-length code, and a variable-length code. The first fixed-length code is located at the beginning of the grid coding and is used to represent grid location information. The variable-length code is located in the middle of the grid coding and is used to represent geology. The second fixed-length code is located at the end of the grid coding and is used to represent disaster risk level.

[0073] Furthermore, in one embodiment of the present invention, each of the above-mentioned grids has a unique code. In another embodiment of the present invention, the number of bits in the first fixed-length code, the second fixed-length code, and the variable-length code can be set as needed.

[0074] For example, in one embodiment of the present invention, a unit grid is coded as A03-B2-024-GS-3, where A03-B2-024 is the first fixed-length code. The entire coded area is divided into large areas such as A, B, and C, and A03 represents the "3rd sub-area" under "large area A". Within the scope of the first-level grid A03, further subdivision is performed, and B2 may represent the grid in the Bth row and 2nd column within this sub-area, divided according to a certain rule (such as 10m×10m), which inherits the spatial range of A03. Within the scope of the second-level grid B2, a more refined division is performed, and 024 is a sequence number representing a specific small grid unit (such as 1m×1m), which inherits the positional information of B2 and A03. Additionally, GS is a variable-length code, where G represents "gravelly soil" and S represents "strongly weathered sandstone". GS is defined as a composite stratum "mainly gravelly soil with interspersed strong weathered sandstone". 3 is a second fixed-length code, indicating that the potential risk level of landslides, collapses and other disasters in this grid unit is level 3 (medium risk).

[0075] Furthermore, in one embodiment of the present invention, after obtaining the grid code corresponding to the open-pit mine area through the above steps, a storage structure combining hash table and B+ tree can be adopted. The hash table is used to quickly locate the approximate storage location of the code, and the B+ tree is used to store the detailed information of the code and its corresponding data.

[0076] Specifically, in one embodiment of the present invention, a storage structure framework combining hash tables and B+ trees is constructed, and the node structure and parameters of the hash function and B+ tree are determined; furthermore, the generated grid code and its related data are gradually inserted into the storage structure to ensure the correctness and orderliness of data storage; and further, based on a parallel computing framework, parallel query and processing algorithms are written and debugged, and through experimental testing and optimization, it is ensured that when processing large-scale data, the data retrieval time is controlled at the second level, and the data retrieval accuracy reaches more than 98%.

[0077] Step 104: Establish the mapping relationship between the mesh code and the model entities in the single 3D model of the open-pit mine to obtain the true 3D model of the open-pit mine with mesh code.

[0078] In one embodiment of the present invention, after determining the mesh code and the three-dimensional model of the open-pit mine through the above steps, a mapping relationship between the mesh code and the model entities in the three-dimensional model of the open-pit mine can be established to obtain the true three-dimensional model of the open-pit mine with mesh code.

[0079] Furthermore, in one embodiment of the present invention, the method for establishing the mapping relationship between the mesh code and the model entities in the single 3D model of the open-pit mine to obtain the true 3D model of the open-pit mine with mesh code may include: establishing mapping tables in the single 3D model of the open-pit mine and the mesh code system respectively, and recording the correspondence between the code and the model entities.

[0080] Specifically, in one embodiment of the present invention, in the three-dimensional model, each entity is associated with a corresponding mesh code; the mesh code system also records model entity information associated with each code. Based on this, a mapping relationship between codes and model entities can be established, allowing for quick location and querying of detailed information about related entities through the codes, and enabling analysis and manipulation of the model based on the codes. Figure 2 This invention provides a grid-encoded true 3D model of an open-pit mine.

[0081] In one embodiment of the present invention, a compatible interface between the multimodal database and the single true 3D model of the open-pit mine can also be developed, which closely links the data in the multimodal database with the single true 3D model of the open-pit mine, realizes data-model interaction, and enables the data to drive the dynamic update of the model in real time, so as to truly reflect the dynamic changes of the open-pit mine.

[0082] In one embodiment of the present invention, after obtaining the true three-dimensional model of the open-pit mine through the above steps, the above method may further include the following steps: Step 105: Based on the true 3D model of the open-pit mine and the multimodal data in the multimodal information database, perform fusion analysis and obtain key indicator data in the disaster scenario through disaster prediction algorithm.

[0083] In one embodiment of the present invention, for different disaster scenarios (such as slope instability or water damage), based on in-depth analysis of multimodal data, the dynamic change trend of the true 3D model of the open-pit mine is predicted, and key indicator data in the corresponding disaster scenario are deduced through a disaster prediction algorithm. In another embodiment of the present invention, the above-mentioned disaster prediction algorithm can be a dynamic simulation performed through limit equilibrium analysis and a physics engine (such as the material point method) to deduce key indicator data in the corresponding disaster scenario.

[0084] In one embodiment of the present invention, the aforementioned key indicator data may include the location of the sliding surface, the volume of the landslide body, the boundary of the collapse area, the depth of water accumulation, and the range of influence.

[0085] For example, in one embodiment of the present invention, for a landslide disaster scenario, the geological structure in the true 3D model of the open-pit mine, the real-time displacement / stress data of the slope monitoring points, and the rainfall data are integrated, and dynamic simulation is performed through limit equilibrium analysis and a physics engine (such as the material point method) to deduce key indicator data such as the location of the sliding surface, the volume of the landslide body, and the range of influence.

[0086] Step 106: Verify the true 3D model of the open-pit mine by comparing and correcting historical data and key indicator data, and optimize the true 3D model of the open-pit mine based on the verification results to obtain the target true 3D model of the open-pit mine.

[0087] The multimodal data-driven true 3D model construction method for open-pit mines in this invention can process and fuse multi-source data from open-pit mines to obtain a multimodal information database. Based on the multimodal data in the multimodal information database, a single 3D model of the open-pit mine is constructed. Using a quadtree-based spatial partitioning method and preset mesh coding rules, the mesh code corresponding to the open-pit mine area is obtained. A mapping relationship is established between the mesh code and the model entities in the single 3D model of the open-pit mine, resulting in a mesh-coded true 3D model of the open-pit mine. Therefore, this invention can obtain the inherent connections in multi-source data through a multimodal information database, and based on the mapping relationship between the mesh code corresponding to the open-pit mine area and the model entities in the single 3D model of the open-pit mine, a mesh-coded true 3D model of the open-pit mine is obtained, thereby improving the detection accuracy and precision of key information in open-pit mine disasters.

[0088] Figure 3This is a schematic diagram of the structure of a multimodal data-driven true 3D model construction device for open-pit mines according to an embodiment of the present invention.

[0089] like Figure 3 As shown, the device may include: The first construction module 301 is used to collect and gather multi-source data from open-pit mines, and to process and fuse the multi-source data to obtain a multimodal information database. The second construction module 302 is used to construct a single three-dimensional model of an open-pit mine based on multimodal data in a multimodal information database. The encoding module 303 is used to obtain the grid code corresponding to the open-pit mine area by using a quadtree-based spatial partitioning method and a preset grid encoding rule; The third construction module 304 is used to establish the mapping relationship between the mesh code and the model entities in the single 3D model of the open-pit mine, so as to obtain the mesh-coded true 3D model of the open-pit mine.

[0090] In one embodiment of the present invention, the first construction module 301 is specifically used for: By collecting and acquiring multi-source data from open-pit mines, long-term multimodal data and monitoring data can be obtained. Data cleaning, missing data interpolation, format conversion, and standardization are performed on long-term multimodal data and monitoring data to obtain the first multimodal data. Feature extraction is performed on the first multimodal data, and the extracted features are fused using a data fusion model to obtain the second multimodal data; Based on second-modal data, a hybrid database is built through a distributed architecture, a data warehouse architecture, and a cloud database architecture.

[0091] In one embodiment of the present invention, the second building module 302 is specifically used for: Acquire satellite and UAV remote sensing images through a multimodal information database; Perform 3D reconstruction and texture mapping on satellite and UAV remote sensing images to generate digital elevation models and digital surface models; Based on digital elevation models and digital surface models, the stress field distribution inside the open-pit mine was deduced; Based on the stress field distribution inside the open-pit mine, a three-dimensional model of the open-pit mine is generated using parametric modeling technology.

[0092] In one embodiment of the present invention, the encoding module 303 is specifically used for: Based on the actual extent and required accuracy of the slope in the open-pit mine area, determine the initial parameters and maximum number of levels for quadtree partitioning; Based on the initial parameters and the maximum number of levels, the quadtree algorithm is used to divide the slope area into grids, and a location index is generated for each grid.

[0093] In one embodiment of the present invention, the aforementioned preset grid encoding rule includes: The grid coding structure includes a first fixed-length code, a second fixed-length code, and a variable-length code. The first fixed-length code is located at the beginning of the grid coding and is used to represent grid location information. The variable-length code is located in the middle of the grid coding and is used to represent geology. The second fixed-length code is located at the end of the grid coding and is used to represent disaster risk level.

[0094] In one embodiment of the present invention, the above-described apparatus is further configured to: Based on the fusion analysis of the true 3D model of the open-pit mine and the multimodal data in the multimodal information database, key indicator data in the disaster scenario are obtained through disaster prediction algorithms; The true 3D model of the open-pit mine is verified by comparing and correcting historical data and key indicator data. Based on the verification results, the true 3D model of the open-pit mine is optimized to obtain the target true 3D model of the open-pit mine.

[0095] The multimodal data-driven true 3D model construction device for open-pit mines, as described in this invention, can process and fuse multi-source data from open-pit mines to obtain a multimodal information database. Based on the multimodal data in the database, a single 3D model of the open-pit mine is constructed. Using a quadtree-based spatial partitioning method and preset mesh coding rules, the mesh code corresponding to the open-pit mine area is obtained. A mapping relationship is established between the mesh code and the model entities in the single 3D model of the open-pit mine, resulting in a mesh-coded true 3D model of the open-pit mine. Therefore, this invention can obtain the inherent connections within multi-source data through a multimodal information database and, based on the mapping relationship between the mesh code corresponding to the open-pit mine area and the model entities in the single 3D model of the open-pit mine, obtain a mesh-coded true 3D model of the open-pit mine, thereby improving the detection accuracy and precision of key information in open-pit mine disasters.

[0096] In this specification, the use of terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refers to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0097] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

Claims

1. A method for constructing a true 3D model of an open-pit mine driven by multimodal data, characterized in that, The method includes: Collect and gather multi-source data from open-pit mines, and process and fuse the multi-source data to obtain a multimodal information database; Based on the multimodal data in the multimodal information database, a three-dimensional model of an open-pit mine is constructed. The grid code corresponding to the open-pit mine area is obtained by using a quadtree-based spatial partitioning method and a preset grid coding rule; Establish the mapping relationship between the mesh code and the model entities in the single 3D model of the open-pit mine to obtain the true 3D model of the open-pit mine with mesh code.

2. The method according to claim 1, characterized in that, The collection and acquisition of multi-source data from open-pit mines, followed by data processing and fusion of the multi-source data to obtain a multimodal information database, includes: By collecting and acquiring multi-source data from open-pit mines, long-term multimodal data and monitoring data can be obtained. The long-term multimodal data and the monitoring data are subjected to data cleaning, missing data interpolation, format conversion and standardization to obtain the first multimodal data; Feature extraction is performed on the first multimodal data, and the extracted features are fused using a data fusion model to obtain the second multimodal data; Based on the second multimodal data, a hybrid database is constructed through a distributed architecture, a data warehouse architecture, and a cloud database architecture.

3. The method according to claim 1, characterized in that, The construction of a single 3D model of an open-pit mine based on the multimodal data in the multimodal information database includes: The multimodal information database is used to acquire satellite and UAV remote sensing images; Three-dimensional reconstruction and texture mapping are performed on the satellite and UAV remote sensing images to generate digital elevation models and digital surface models; Based on the digital elevation model and the digital surface model, the stress field distribution inside the open-pit mine is deduced; Based on the stress field distribution inside the open-pit mine, a three-dimensional model of the open-pit mine is generated using parametric modeling technology.

4. The method according to claim 1, characterized in that, The quadtree-based space partitioning method includes: Based on the actual extent and required accuracy of the slope in the open-pit mine area, determine the initial parameters and maximum number of levels for quadtree partitioning; Based on the initial parameters and the maximum number of levels, the slope area is divided into grids using a quadtree algorithm, and a location index is generated for each grid.

5. The method according to claim 1, characterized in that, The preset grid encoding rules include: The grid coding structure includes a first fixed-length code, a second fixed-length code, and a variable-length code. The first fixed-length code is located at the beginning of the grid coding and is used to represent grid location information. The variable-length code is located in the middle of the grid coding and is used to represent geology. The second fixed-length code is located at the end of the grid coding and is used to represent disaster risk level.

6. The method according to claim 1, characterized in that, The method further includes: Based on the true 3D model of the open-pit mine and the multimodal data in the multimodal information database, a fusion analysis is performed to obtain key indicator data in the disaster scenario through a disaster scenario simulation algorithm; The true 3D model of the open-pit mine is verified by comparing and correcting historical data and key indicator data. Based on the verification results, the true 3D model of the open-pit mine is optimized to obtain the target true 3D model of the open-pit mine.

7. A multimodal data-driven true 3D model construction device for open-pit mines, characterized in that, The device includes: The first construction module is used to collect and gather multi-source data from open-pit mines, and to process and fuse the multi-source data to obtain a multimodal information database. The second construction module is used to construct a single three-dimensional model of the open-pit mine based on the multimodal data in the multimodal information database. The encoding module is used to obtain the grid code corresponding to the open-pit mine area by using a quadtree-based spatial partitioning method and preset grid encoding rules; The third construction module is used to establish the mapping relationship between the mesh code and the model entities in the single 3D model of the open-pit mine, so as to obtain the mesh-coded true 3D model of the open-pit mine.

8. The method according to claim 7, characterized in that, The first building module is specifically used for: By collecting and acquiring multi-source data from open-pit mines, long-term multimodal data and monitoring data can be obtained. The long-term multimodal data and the monitoring data are subjected to data cleaning, missing data interpolation, format conversion and standardization to obtain the first multimodal data; Feature extraction is performed on the first multimodal data, and the extracted features are fused using a data fusion model to obtain the second multimodal data; Based on the second multimodal data, a hybrid database is constructed through a distributed architecture, a data warehouse architecture, and a cloud database architecture.

9. An electronic device, comprising: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-6.

10. A computer storage medium, wherein, The computer storage medium stores computer-executable instructions; when executed by a processor, the computer-executable instructions can implement the method as described in any one of claims 1-6.

Images