A bauxite resource prediction and analysis method and system

Abstract

The present application provides a bauxite resource prediction and analysis method and system, which relates to the field of resource prediction technology. The method comprises collecting a surface remote sensing image of a target bauxite; performing distortion correction on the surface remote sensing image according to digital elevation data of the target bauxite to obtain a surface remote sensing correction image; performing image masking on the surface remote sensing correction image to obtain a surface remote sensing mask image, and extracting surface resource distribution characteristics from the surface remote sensing mask image; obtaining geological exploration data of the target bauxite, generating a three-dimensional distribution model of the ore body according to the surface resource distribution characteristics and the geological exploration data, obtaining a bauxite database, and constructing a mineralization knowledge map of bauxite according to the bauxite database; predicting the ore body grade according to the three-dimensional distribution model of the ore body and the mineralization knowledge map of bauxite to obtain ore body grade prediction information. The present application can improve the quality of remote sensing images and enhance the complementarity of data to improve the reliability of bauxite resource prediction.

Description

Technical Field

[0001] The present application relates to the technical field of resource prediction, and more specifically, to a bauxite resource prediction and analysis method and system. Background Art

[0002] Bauxite resource prediction and analysis is a complex technical process that integrates multidisciplinary knowledge. It mainly integrates remote sensing image analysis, geological exploration data, mineralization theory and modern computing technology to scientifically predict the reserves, distribution and grade of bauxite resources.

[0003] In the existing technology, although remote sensing images can provide a wide range of surface information, their resolution and spectral characteristics may be affected by factors such as weather and vegetation coverage, resulting in a decrease in the accuracy of identifying mineral-bearing areas. In addition, geological exploration data are usually only obtained in a local area, and there are problems such as insufficient samples and uneven spatial distribution. Especially under complex geological conditions, the sampling representativeness of drilling data may not fully reflect the actual distribution characteristics of the ore body. In addition, there may be contradictions between multi-source data, and data mining technology is needed to eliminate redundant information and enhance the complementarity of data to improve the stability and reliability of the prediction model. Therefore, how to improve the quality of remote sensing images and enhance the complementarity of data to improve the reliability of bauxite resource prediction is a difficult problem faced by the industry. Summary of the invention

[0004] The present application provides a bauxite resource prediction and analysis method and system, which can improve the quality of remote sensing images and enhance the complementarity of data to improve the reliability of bauxite resource prediction.

[0005] In a first aspect, the present application provides a bauxite resource prediction and analysis method, the prediction and analysis method comprising the following steps:

[0006] Collect surface remote sensing images of the target bauxite mine;

[0007] According to the digital elevation data of the target bauxite mine, the surface remote sensing image is subjected to distortion correction to obtain a surface remote sensing corrected image of the target bauxite mine, the surface remote sensing corrected image is subjected to image masking to obtain a surface remote sensing masked image of the target bauxite mine, and the surface resource distribution characteristics of the target bauxite mine are extracted from the surface remote sensing masked image;

[0008] Acquire geological exploration data of the target bauxite mine, generate a three-dimensional distribution model of the ore body of the target bauxite mine through the surface resource distribution characteristics and the geological exploration data, acquire a bauxite database, and construct a mineralization knowledge map of bauxite based on the bauxite database;

[0009] The ore body grade of the target bauxite is predicted based on the three-dimensional distribution model of the ore body and the mineralization knowledge map of the bauxite to obtain ore body grade prediction information.

[0010] In this embodiment, the surface remote sensing image of the target bauxite mine is collected by a sensor carried by a drone.

[0011] In this embodiment, the surface remote sensing image is subjected to distortion correction according to the digital elevation data of the target bauxite mine to obtain the surface remote sensing corrected image of the target bauxite mine, specifically comprising:

[0012] Get pre-built image correction models;

[0013] Selecting a calibration base point based on the image calibration model and digital elevation data of the target bauxite mine;

[0014] The surface remote sensing image is subjected to distortion correction through the correction base points to obtain a surface remote sensing corrected image of the target bauxite mine.

[0015] In this embodiment, performing image masking on the surface remote sensing corrected image includes performing vegetation masking, water body masking and town masking on the surface remote sensing corrected image respectively.

[0016] In this embodiment, the surface resource distribution characteristics of the target bauxite are extracted from the surface remote sensing mask image, specifically including:

[0017] Obtaining the spectral characteristics of bauxite;

[0018] Extracting the sensitive band of the target bauxite in the surface remote sensing mask image based on the spectral characteristics of the bauxite;

[0019] The sensitive bands are used to construct the surface resource distribution characteristics of the target bauxite.

[0020] In this embodiment, generating a three-dimensional distribution model of a target bauxite ore body by using the surface resource distribution characteristics and the geological exploration data specifically includes:

[0021] Generating a surface resource distribution map of the target bauxite based on the surface resource distribution characteristics and the digital elevation data of the target bauxite;

[0022] The surface resource distribution map and the geological exploration data are input into a three-dimensional modeling tool to generate a three-dimensional distribution model of the ore body of the target bauxite mine.

[0023] In this embodiment, constructing a mineralization knowledge graph of bauxite based on the bauxite database specifically includes:

[0024] Extracting ore body information from the bauxite database to obtain mineralization information of the bauxite;

[0025] Extracting semantic features from the mineralization information to obtain multiple semantic feature vectors;

[0026] Relationship extraction is performed on all semantic feature vectors to obtain the mineralization entity relationship structure of bauxite;

[0027] The mineralization entity relationship structure is used to construct a mineralization knowledge graph of bauxite.

[0028] In this embodiment, all semantic feature vectors are subjected to relationship extraction to obtain the mineralization entity relationship structure of bauxite, which specifically includes:

[0029] Get all entities in mineralization information;

[0030] Input each entity and all semantic feature vectors into the relation extraction model to extract relations and obtain an entity relation graph;

[0031] The entity relationship diagram is used as the mineralization entity relationship structure of bauxite.

[0032] In this embodiment, the ore body grade of the target bauxite is predicted according to the three-dimensional distribution model of the ore body and the mineralization knowledge map of the bauxite, and the ore body grade prediction information obtained specifically includes:

[0033] Aligning the three-dimensional distribution model of the ore body with the metallogenic knowledge graph of the bauxite, and then extracting the spatial characteristics of the ore body in the three-dimensional distribution model of the ore body, and extracting the metallogenic characteristics in the metallogenic knowledge graph of the bauxite;

[0034] Performing feature fusion on the ore body spatial feature and the mineralization feature to obtain an ore body fusion feature vector of the target bauxite;

[0035] The ore body fusion feature vector is input into a prediction model based on deep learning to perform ore body grade prediction to obtain ore body grade prediction information.

[0036] In a second aspect, the present application provides a bauxite resource prediction and analysis system for executing a bauxite resource prediction and analysis method, the prediction and analysis system comprising:

[0037] Image acquisition module, used to collect surface remote sensing images of target bauxite mines;

[0038] A distribution feature extraction module is used to perform distortion correction on the surface remote sensing image according to the digital elevation data of the target bauxite mine to obtain a surface remote sensing corrected image of the target bauxite mine, perform image masking on the surface remote sensing corrected image to obtain a surface remote sensing masked image of the target bauxite mine, and extract the surface resource distribution characteristics of the target bauxite mine from the surface remote sensing masked image;

[0039] A model building module is used to obtain geological exploration data of a target bauxite mine, generate a three-dimensional distribution model of the ore body of the target bauxite mine through the surface resource distribution characteristics and the geological exploration data, obtain a bauxite database, and build a mineralization knowledge map of bauxite based on the bauxite database;

[0040] The ore body information prediction module is used to predict the ore body grade of the target bauxite according to the ore body three-dimensional distribution model and the bauxite mineralization knowledge map to obtain ore body grade prediction information.

[0041] The technical solution provided by the embodiments disclosed in this application has the following beneficial effects:

[0042] By collecting surface remote sensing images of the target bauxite; performing distortion correction on the surface remote sensing images according to the digital elevation data of the target bauxite to obtain the surface remote sensing corrected images of the target bauxite; performing image masking on the surface remote sensing corrected images to obtain the surface remote sensing mask images of the target bauxite; extracting the surface resource distribution characteristics of the target bauxite from the surface remote sensing mask images; acquiring geological exploration data of the target bauxite, generating a three-dimensional distribution model of the ore body of the target bauxite through the surface resource distribution characteristics and the geological exploration data, acquiring a bauxite database, and constructing a mineralization knowledge map of bauxite according to the bauxite database; predicting the ore body grade of the target bauxite according to the three-dimensional distribution model of the ore body and the mineralization knowledge map of the bauxite to obtain ore body grade prediction information.

[0043] It can be seen that in this application, firstly, the use of digital elevation data for distortion correction can make the image closer to the actual surface situation and improve the geometric accuracy of the image. The image masking process can remove the ground object information irrelevant to the target bauxite, which is helpful to more accurately extract the spatial distribution of the ore body. Through the surface resource distribution characteristics, the location and scale of bauxite resources can be more accurately identified; then, by combining geological exploration data with surface resource distribution characteristics, the complementarity of the data is enhanced, the accuracy and reliability of three-dimensional modeling of the ore body can be improved, and the comprehensive understanding of mineral resources can be improved by constructing a mineralization knowledge map; finally, through the effective combination of the three-dimensional distribution model of the ore body and the mineralization knowledge map, the ore grade of bauxite can be more accurately predicted.

[0044] In summary, the technical solution adopted in this application can improve the quality of remote sensing images and enhance the complementarity of data to improve the reliability of bauxite resource prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0046] Figure 1 is a flow chart of the bauxite resource prediction and analysis method provided in this application;

[0047] Figure 2 is an exemplary flow chart for determining the surface remote sensing correction image of the target bauxite mine provided by the present application;

[0048] Figure 3 It is an exemplary flow chart for constructing a mineralization knowledge map of bauxite according to the present application;

[0049] Figure 4 It is a module structure diagram of the bauxite resource prediction and analysis system provided in the present application, wherein the prediction and analysis system includes: an image acquisition module 100, a distribution feature extraction module 200, a model construction module 300 and an ore body information prediction module 400. DETAILED DESCRIPTION

[0050] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0051] The embodiment of the present application provides a bauxite resource prediction and analysis method and system, the core of which is to collect the surface remote sensing image of the target bauxite; perform distortion correction on the surface remote sensing image according to the digital elevation data of the target bauxite to obtain the surface remote sensing correction image of the target bauxite, perform image masking on the surface remote sensing correction image to obtain the surface remote sensing mask image of the target bauxite, and extract the surface resource distribution characteristics of the target bauxite from the surface remote sensing mask image; obtain the geological exploration data of the target bauxite, generate the ore body three-dimensional distribution model of the target bauxite through the surface resource distribution characteristics and the geological exploration data, obtain the bauxite database, and construct the mineralization knowledge map of bauxite according to the bauxite database; predict the ore body grade of the target bauxite according to the ore body three-dimensional distribution model and the mineralization knowledge map of the bauxite, and obtain the ore body grade prediction information. The above scheme can improve the quality of remote sensing images and enhance the complementarity of data to improve the reliability of bauxite resource prediction.

[0052] Embodiment 1

[0053] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods. Figure 1 As shown in the figure, this figure is an exemplary flow chart of a bauxite resource prediction and analysis method according to this embodiment of the present application, and the prediction and analysis method includes the following steps:

[0054] In step S1, a surface remote sensing image of a target bauxite mine is collected.

[0055] In the specific implementation, the surface remote sensing image of the target bauxite is collected by the sensor carried by the drone; the sensor used in this application is a multispectral sensor, which can capture visible light and near-infrared band information for analyzing the mineral reflection characteristics of the surface.

[0056] In step S2, the surface remote sensing image is subjected to distortion correction according to the digital elevation data of the target bauxite mine to obtain a surface remote sensing corrected image of the target bauxite mine, the surface remote sensing corrected image is subjected to image masking to obtain a surface remote sensing mask image of the target bauxite mine, and the surface resource distribution characteristics of the target bauxite mine are extracted from the surface remote sensing mask image.

[0057] Preferably, in this embodiment, reference Figure 2 As shown in FIG. 1 , this figure is an exemplary flow chart of determining the surface remote sensing corrected image of the target bauxite in an embodiment of the present application. In this embodiment, the surface remote sensing image is subjected to distortion correction based on the digital elevation data of the target bauxite to obtain the surface remote sensing corrected image of the target bauxite, which can be specifically implemented by the following steps:

[0058] First, in step S21, a pre-built image correction model is obtained;

[0059] Then, in step S22, a calibration base point is selected according to the image calibration model and the digital elevation data of the target bauxite;

[0060] Finally, in step S23, the surface remote sensing image is subjected to distortion correction through the correction base points to obtain a surface remote sensing corrected image of the target bauxite mine.

[0061] In specific implementation, first, a pre-built image correction model can be obtained. The image correction model is established through historical data, remote sensing images and known geographic reference points, and is used to correct image distortion. The image correction model in this application is a geometric correction model, which mainly considers the influence of factors such as terrain undulation, sensor viewing angle, and flight path on the image position; then, the digital elevation data of the target bauxite mine can be obtained, so as to select the correction base point based on the image correction model and the digital elevation data. The correction base point refers to a reference point with a known position in the image and the ground real coordinate system. Multiple correction base points that can cover the entire target bauxite mining area are selected to ensure that image deformation in different areas can be corrected; Finally, the ground coordinates of the selected correction base point are matched with the corresponding pixel coordinates of the surface remote sensing image. The image registration method can be used to utilize the relationship between the spatial coordinates of the correction base point and the pixel position of the surface remote sensing image to solve the conversion relationship between the surface remote sensing image and the ground true coordinates. Among them, the matching method adopted in this application is the nearest neighbor registration method. According to the image correction model established in the early stage and the position of the correction base point, the surface remote sensing image is geometrically corrected through mathematical transformation (such as affine transformation, perspective transformation, etc.), so as to obtain the surface remote sensing corrected image of the target bauxite. The image geometric distortion caused by external errors can be eliminated through correction, thereby eliminating the influence of distortion on information extraction.

[0062] In this embodiment, performing image masking on the surface remote sensing corrected image includes performing vegetation masking, water body masking and town masking on the surface remote sensing corrected image respectively.

[0063] It should be noted that the information in the surface remote sensing corrected images is often weak information, which is easily affected by factors such as vegetation and water bodies, resulting in the information not being obvious or being masked in the surface remote sensing corrected images. Therefore, before extracting the information, this application adopts an image masking method to mark the interference factors and other parts of the image as 0 and 1, respectively, and make masks for different interference factors to achieve the purpose of removing interference and highlighting information.

[0064] In specific implementation, the surface remote sensing corrected image can be subjected to vegetation masking, that is, the vegetation and non-vegetation areas in the surface remote sensing corrected image can be distinguished by the vegetation index, so that the pixels in the vegetation area in the surface remote sensing corrected image are marked as 0, wherein the vegetation index is an indicator used to measure the degree of vegetation growth in the corresponding area. In actual implementation, the vegetation index can be determined in the following way: vegetation index = (near infrared band - red band) / (near infrared band + red band). When the vegetation index is greater than the set threshold, the area is determined to be a vegetation area. In addition, the method of water body masking and town masking for the surface remote sensing corrected image is the same as the vegetation masking method, the difference is that the water body The indicators used for the mask and town mask are the water body index and the town index respectively; among them, the water body index is an indicator used to measure the water body coverage of the corresponding area. In actual implementation, the water body index can be determined in the following way: water body index = (green band-near infrared band) / (green band + near infrared band); the town index is an indicator used to measure the town coverage of the corresponding area. In actual implementation, the town index can be determined in the following way: town index = (short-wave infrared band-near infrared band) / (short-wave infrared band + near infrared band); the image mask of the surface remote sensing correction image can be completed in the above way, thereby obtaining the surface remote sensing mask image of the target bauxite.

[0065] In this embodiment, the surface resource distribution characteristics of the target bauxite can be extracted from the surface remote sensing mask image in the following manner, namely:

[0066] Obtaining the spectral characteristics of bauxite;

[0067] Extracting the sensitive band of the target bauxite in the surface remote sensing mask image based on the spectral characteristics of the bauxite;

[0068] The sensitive bands are used to construct the surface resource distribution characteristics of the target bauxite.

[0069] In the specific implementation, first, the spectral characteristics of bauxite can be obtained. The spectral characteristics of bauxite refer to the reflectivity or radiation characteristics of bauxite in different bands. The spectral characteristics can be identified in remote sensing images. The spectral characteristics of bauxite are different in different remote sensing bands. For example, bauxite may have a higher reflectivity in certain specific bands, such as the shortwave infrared band (SWIR), and may show a lower reflectivity in the near infrared band (NIR); then, based on the spectral characteristics of bauxite, the sensitive bands of the target bauxite can be extracted from the surface remote sensing mask image. Sensitive bands refer to bands with obvious changes in the spectral characteristics of bauxite and are easy to distinguish between bauxite and other landforms. Generally speaking, sensitive bands are most sensitive to changes in the spectral reflectivity of bauxite. It can effectively distinguish bauxite from other minerals or landforms (such as vegetation, water bodies, buildings, etc.). According to the spectral characteristics of bauxite, the band with the largest contrast or the strongest response can be selected as the sensitive band in the surface remote sensing mask image; finally, the sensitive band can be used to construct the surface resource distribution characteristics of the target bauxite, that is, the bauxite area in the surface remote sensing mask image can be extracted based on the sensitive band, and then the possible bauxite area can be marked through spatial analysis methods, and the marking results can be corrected through morphological operations (such as expansion, corrosion, etc.) to make the bauxite area more continuous and accurate. The final marking result can be used as the surface resource distribution feature of the target bauxite, that is, the surface resource distribution feature represents the distribution shape characteristics of the surface area of ​​the target bauxite.

[0070] It should be noted that the quality of remote sensing images can be improved through distortion correction. Using digital elevation data for distortion correction can make the images closer to the actual surface conditions, improve the geometric accuracy of the images, and ensure that the data extracted from the images are correct. The image masking process (including vegetation, water bodies, urban areas, etc.) can remove ground information that is not related to the target bauxite. The masked images only retain the surface features related to bauxite, which helps to more accurately extract the spatial distribution of the ore body. Through the surface resource distribution characteristics, the location and scale of bauxite resources can be more accurately identified, providing accurate input data for subsequent ore body predictions.

[0071] In step S3, the geological exploration data of the target bauxite is obtained, a three-dimensional distribution model of the ore body of the target bauxite is generated through the surface resource distribution characteristics and the geological exploration data, a bauxite database is obtained, and a mineralization knowledge map of bauxite is constructed based on the bauxite database.

[0072] In specific implementation, geological exploration data of the target bauxite can be queried and obtained through the Geological Survey Bureau and the Mineral Resources Management Department. The geological exploration data includes the drilling data, geophysical data and geochemical data of the target bauxite. Among them, the drilling data includes the drilling location (latitude and longitude or projection coordinates), the depth and thickness of the ore layer and the ore grade; the geophysical data includes gravity, magnetic or resistivity profiles to obtain the density, conductivity or magnetic characteristics of the ore layer; the geochemical data includes the vertical distribution of element content (such as , content).

[0073] In this embodiment, the three-dimensional distribution model of the target bauxite ore body is generated by using the surface resource distribution characteristics and the geological exploration data in the following manner, namely:

[0074] Generating a surface resource distribution map of the target bauxite based on the surface resource distribution characteristics and the digital elevation data of the target bauxite;

[0075] The surface resource distribution map and the geological exploration data are input into a three-dimensional modeling tool to generate a three-dimensional distribution model of the ore body of the target bauxite mine.

[0076] In specific implementation, first, a surface resource distribution map can be generated based on the surface resource distribution characteristics and digital elevation data, that is, by combining the surface resource distribution characteristics with the digital elevation data, the resource distribution information of the ore body can be accurately mapped into three-dimensional space, and a surface resource distribution map of the target bauxite mine can be generated. This distribution map shows the resource concentration in different locations of the mining area, and can reflect the topographic undulations of the mining area and the spatial changes of resources; then, the surface resource distribution map and geological exploration data can be input into a three-dimensional modeling tool, and professional three-dimensional modeling software (such as AutoCAD, ArcGIS, Leapfrog, Geosoft, etc.) can be used to import the surface resource distribution map and geological exploration data. By combining the surface resource distribution map with the exploration data, the three-dimensional modeling tool can simulate the distribution of the ore body in space and generate a three-dimensional distribution model of the ore body. The three-dimensional distribution model of the ore body can clearly show the shape, volume and spatial position of the ore body.

[0077] In this embodiment, the bauxite database can be obtained from a public mineral resource information database provided by a geological exploration agency; the bauxite database contains various basic data about bauxite, such as geological characteristics, ore deposit information, exploration data, ore grade, mining history, etc.

[0078] Preferably, in this embodiment, reference Figure 3 As shown in the figure, this figure is an exemplary flow chart of constructing a metallogenic knowledge graph of bauxite in an embodiment of the present application. In this embodiment, the metallogenic knowledge graph of bauxite is constructed based on the bauxite database, which can be implemented by the following steps:

[0079] First, in step S31, ore body information is extracted from the bauxite database to obtain mineralization information of the bauxite;

[0080] Then, in step S32, semantic feature extraction is performed on the mineralization information to obtain a plurality of semantic feature vectors;

[0081] Secondly, in step S33, all semantic feature vectors are subjected to relationship extraction to obtain the mineralization entity relationship structure of bauxite;

[0082] Finally, in step S34, the mineralization entity relationship structure is used to construct a mineralization knowledge graph of bauxite.

[0083] In the specific implementation, first, ore body information can be extracted from the bauxite database, that is, various information related to the ore body can be collected from the bauxite database, that is, the mineralization information of bauxite, including the spatial position of the ore body, ore grade, geometric shape of the ore body, mineralization process and other basic mineralization data; then, semantic feature extraction can be performed on the mineralization information, that is, in the extracted mineralization information, key semantic features such as the mineral composition of the ore, the structural characteristics of the ore deposit, the mineralization conditions, magmatic activity, etc. are identified, and the semantic features are converted into vectors, thereby obtaining multiple semantic feature vectors. Usually, word embedding (Word2Vec, BERT and other models) is used to convert the semantic information in text data into numerical feature vectors that can be understood by computers; secondly, relationship extraction can be performed on all semantic feature vectors to obtain the mineralization entity relationship structure of bauxite; finally, the mineralization entity relationship structure can be used to construct the mineralization knowledge graph of bauxite, that is, the mineralization entity relationship structure is converted into the mineralization knowledge graph of bauxite using visualization tools (such as Neo4jBloom).

[0084] In this embodiment, the following method is used to extract relations from all semantic feature vectors to obtain the mineralization entity relation structure of bauxite, namely:

[0085] Get all entities in mineralization information;

[0086] Input each entity and all semantic feature vectors into the relation extraction model to extract relations and obtain an entity relation graph;

[0087] The entity relationship diagram is used as the mineralization entity relationship structure of bauxite.

[0088] In the specific implementation, first, the various entity types involved in the mineralization information are clarified, including ore bodies, rock formations, minerals, ore deposit types, mineralization, mineralization periods, mineralization conditions, etc., and the extracted entities are standardized to ensure their consistency, such as unifying mineral names, rock formation types, etc., to facilitate subsequent semantic relationship extraction and graph construction; then, each entity and all semantic feature vectors can be input into the relationship extraction model for relationship extraction, that is, all entities and all semantic feature vectors are organized into a format suitable for model input, and a suitable deep learning-based relationship extraction model is selected for relationship extraction. In this application, the deep learning-based relationship extraction model is selected as the BiLSTM model, so that based on the output of the relationship extraction model, an entity relationship graph is generated, and the nodes in the entity relationship graph represent entities, and the edges represent the relationships between entities; finally, the entity relationship graph can be used as the mineralization entity relationship structure of bauxite.

[0089] It should be noted that by combining geological exploration data with the distribution characteristics of surface resources, the complementarity of data is enhanced, the accuracy and reliability of ore body three-dimensional modeling can be improved, and the comprehensive understanding of mineral resources can be improved by constructing a mineralization knowledge map. The mineralization knowledge map can display the geology, ore bodies, minerals and other related information of the mining area in a structured manner to build a comprehensive ore body knowledge system. Through the construction of the map, the complex relationship between different geological elements can be revealed, such as the correlation between ore bodies and ore deposit types, ore bodies and mineralization periods, etc. In this way, it can not only help geologists better understand the mineralization laws of bauxite, but also provide more scientific data support to guide mining and resource management in mining areas.

[0090] In step S4, the ore body grade of the target bauxite is predicted based on the ore body three-dimensional distribution model and the bauxite mineralization knowledge map to obtain ore body grade prediction information.

[0091] In this embodiment, the ore body grade of the target bauxite is predicted according to the three-dimensional distribution model of the ore body and the mineralization knowledge map of the bauxite, and the ore body grade prediction information can be obtained in the following manner, namely:

[0092] Aligning the three-dimensional distribution model of the ore body with the metallogenic knowledge graph of the bauxite, and then extracting the spatial characteristics of the ore body in the three-dimensional distribution model of the ore body, and extracting the metallogenic characteristics in the metallogenic knowledge graph of the bauxite;

[0093] Performing feature fusion on the ore body spatial feature and the mineralization feature to obtain an ore body fusion feature vector of the target bauxite;

[0094] The ore body fusion feature vector is input into a prediction model based on deep learning to perform ore body grade prediction to obtain ore body grade prediction information.

[0095] In the specific implementation, first of all, it is necessary to spatially and semantically align the ore body three-dimensional distribution model with the metallogenic knowledge graph of bauxite, that is, it is necessary to ensure that the spatial coordinates in the ore body three-dimensional distribution model (such as the spatial position, depth, morphology, etc. of the ore body) correspond to the ore body characteristics defined in the metallogenic knowledge graph (such as ore deposit type, ore body formation conditions, ore grade, etc.), so that the ore body spatial characteristics can be extracted in the ore body three-dimensional distribution model, and the metallogenic characteristics can be extracted in the metallogenic knowledge graph of bauxite; then, the ore body spatial characteristics and the metallogenic characteristics are feature fused to obtain a unified ore body fusion feature vector, and a deep neural network (DNN) can be used to learn these The joint representation of features captures the potential relationship by mapping the feature space to obtain the fused feature vector of the ore body; finally, the fused feature vector of the ore body can be input into the prediction model based on deep learning to predict the ore body grade. In this application, a deep neural network (DNN) is used to predict the ore body grade. The obtained fused feature vector of the ore body is input into the deep neural network. The deep neural network performs nonlinear mapping on these features through a multi-layer network structure to generate the final ore body grade prediction result, that is, the ore body grade prediction information. The ore body grade prediction information includes the average grade of the ore body, the distribution characteristics of the ore body, the grade differences in different regions, etc.

[0096] It should be noted that through the effective combination of the three-dimensional distribution model of the ore body and the mineralization knowledge map, the ore grade of bauxite can be predicted more accurately. By extracting and integrating the spatial characteristics of the ore body with the mineralization characteristics and using the deep learning model for prediction, the grade distribution of the ore body can be evaluated more comprehensively and accurately. This not only improves the accuracy of ore grade prediction, but also effectively utilizes the complementarity of different data sources to provide support for the efficient development of bauxite resources.

[0097] It can be seen that in this application, firstly, the use of digital elevation data for distortion correction can make the image closer to the actual surface situation and improve the geometric accuracy of the image. The image masking process can remove the ground object information irrelevant to the target bauxite, which is helpful to more accurately extract the spatial distribution of the ore body. Through the surface resource distribution characteristics, the location and scale of bauxite resources can be more accurately identified; then, by combining geological exploration data with surface resource distribution characteristics, the complementarity of the data is enhanced, the accuracy and reliability of three-dimensional modeling of the ore body can be improved, and the comprehensive understanding of mineral resources can be improved by constructing a mineralization knowledge map; finally, through the effective combination of the three-dimensional distribution model of the ore body and the mineralization knowledge map, the ore grade of bauxite can be more accurately predicted.

[0098] In summary, the technical solution adopted in this application can improve the quality of remote sensing images and enhance the complementarity of data to improve the reliability of bauxite resource prediction.

[0099] Embodiment 2

[0100] This application provides a bauxite resource prediction and analysis system, reference Figure 4 As shown, this figure is a schematic diagram of the prediction analysis system shown in this embodiment of the present application, and the prediction analysis system includes:

[0101] An image acquisition module 100 is used to acquire a surface remote sensing image of a target bauxite mine;

[0102] The distribution feature extraction module 200 is used to perform distortion correction on the surface remote sensing image according to the digital elevation data of the target bauxite mine to obtain a surface remote sensing corrected image of the target bauxite mine, perform image masking on the surface remote sensing corrected image to obtain a surface remote sensing masked image of the target bauxite mine, and extract the surface resource distribution characteristics of the target bauxite mine from the surface remote sensing masked image;

[0103] The model building module 300 is used to obtain geological exploration data of the target bauxite, generate a three-dimensional distribution model of the ore body of the target bauxite through the surface resource distribution characteristics and the geological exploration data, obtain a bauxite database, and build a mineralization knowledge map of the bauxite based on the bauxite database;

[0104] The ore body information prediction module 400 is used to predict the ore body grade of the target bauxite according to the ore body three-dimensional distribution model and the bauxite mineralization knowledge map to obtain ore body grade prediction information.

[0105] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0106] A person skilled in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, the storage medium including a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically-erasable programmable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0107] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

Claims

1. A bauxite resource prediction and analysis method, characterized in that: The prediction analysis method comprises the following steps: Collect surface remote sensing images of the target bauxite mine; According to the digital elevation data of the target bauxite mine, the surface remote sensing image is subjected to distortion correction to obtain a surface remote sensing corrected image of the target bauxite mine, the surface remote sensing corrected image is subjected to image masking to obtain a surface remote sensing masked image of the target bauxite mine, and the surface resource distribution characteristics of the target bauxite mine are extracted from the surface remote sensing masked image; Acquire geological exploration data of the target bauxite mine, generate a three-dimensional distribution model of the ore body of the target bauxite mine through the surface resource distribution characteristics and the geological exploration data, acquire a bauxite database, and construct a mineralization knowledge map of bauxite based on the bauxite database; The ore body grade of the target bauxite is predicted according to the three-dimensional distribution model of the ore body and the metallogenic knowledge map of the bauxite to obtain ore body grade prediction information, wherein the ore body grade of the target bauxite is predicted according to the three-dimensional distribution model of the ore body and the metallogenic knowledge map of the bauxite to obtain the ore body grade prediction information specifically includes: Aligning the three-dimensional distribution model of the ore body with the metallogenic knowledge graph of the bauxite, and then extracting the spatial characteristics of the ore body in the three-dimensional distribution model of the ore body, and extracting the metallogenic characteristics in the metallogenic knowledge graph of the bauxite; Performing feature fusion on the ore body spatial feature and the mineralization feature to obtain an ore body fusion feature vector of the target bauxite; The ore body fusion feature vector is input into a prediction model based on deep learning to perform ore body grade prediction to obtain ore body grade prediction information.

2. A bauxite resource prediction and analysis method according to claim 1, characterized in that: The surface remote sensing images of the target bauxite mine are collected by sensors carried by drones.

3. A bauxite resource prediction and analysis method according to claim 1, characterized in that: The surface remote sensing image is subjected to distortion correction according to the digital elevation data of the target bauxite mine to obtain the surface remote sensing corrected image of the target bauxite mine, specifically comprising: Get pre-built image correction models; Selecting a calibration base point based on the image calibration model and digital elevation data of the target bauxite mine; The surface remote sensing image is subjected to distortion correction through the correction base points to obtain a surface remote sensing corrected image of the target bauxite mine.

4. A bauxite resource prediction and analysis method according to claim 1, characterized in that: Performing image masking on the surface remote sensing corrected image includes respectively performing vegetation masking, water body masking and town masking on the surface remote sensing corrected image.

5. The bauxite resource prediction and analysis method according to claim 1, characterized in that: The surface resource distribution characteristics of the target bauxite extracted from the surface remote sensing mask image specifically include: Obtaining the spectral characteristics of bauxite; Extracting the sensitive band of the target bauxite in the surface remote sensing mask image based on the spectral characteristics of the bauxite; The sensitive bands are used to construct the surface resource distribution characteristics of the target bauxite.

6. A bauxite resource prediction and analysis method according to claim 1, characterized in that: Generating a three-dimensional distribution model of a target bauxite ore body by using the surface resource distribution characteristics and the geological exploration data specifically includes: Generating a surface resource distribution map of the target bauxite based on the surface resource distribution characteristics and the digital elevation data of the target bauxite; The surface resource distribution map and the geological exploration data are input into a three-dimensional modeling tool to generate a three-dimensional distribution model of the ore body of the target bauxite mine.

7. A bauxite resource prediction and analysis method according to claim 1, characterized in that: The construction of the mineralization knowledge map of bauxite based on the bauxite database specifically includes: Extracting ore body information from the bauxite database to obtain mineralization information of the bauxite; Extracting semantic features from the mineralization information to obtain multiple semantic feature vectors; Relationship extraction is performed on all semantic feature vectors to obtain the mineralization entity relationship structure of bauxite; The mineralization entity relationship structure is used to construct a mineralization knowledge graph of bauxite.

8. A bauxite resource prediction and analysis method as claimed in claim 7, characterized in that: Relationship extraction is performed on all semantic feature vectors to obtain the mineralization entity relationship structure of bauxite, which includes: Get all entities in mineralization information; Input each entity and all semantic feature vectors into the relation extraction model to extract relations and obtain an entity relation graph; The entity relationship diagram is used as the mineralization entity relationship structure of bauxite.

9. A bauxite resource prediction and analysis system, used to execute a bauxite resource prediction and analysis method according to any one of claims 1 to 8, characterized in that: The predictive analysis system comprises: Image acquisition module, used to collect surface remote sensing images of target bauxite mines; A distribution feature extraction module is used to perform distortion correction on the surface remote sensing image according to the digital elevation data of the target bauxite mine to obtain a surface remote sensing corrected image of the target bauxite mine, perform image masking on the surface remote sensing corrected image to obtain a surface remote sensing masked image of the target bauxite mine, and extract the surface resource distribution characteristics of the target bauxite mine from the surface remote sensing masked image; A model building module is used to obtain geological exploration data of a target bauxite mine, generate a three-dimensional distribution model of the ore body of the target bauxite mine through the surface resource distribution characteristics and the geological exploration data, obtain a bauxite database, and build a mineralization knowledge map of bauxite based on the bauxite database; The ore body information prediction module is used to predict the ore body grade of the target bauxite according to the ore body three-dimensional distribution model and the bauxite mineralization knowledge map to obtain ore body grade prediction information.

Images