Intelligent construction system for mineral deposit model spatial data sample library

Abstract

The invention provides an intelligent construction system for a mineral deposit model spatial data sample library, and the system comprises a mineral deposit model sample library management module which is used for constructing a source spatial database, issuing a task to a sample library construction workbench module, monitoring the issued task, and carrying out the reprocessing of a delivery task of the sample library construction workbench module, constructing an ore deposit model sample library; the sample library construction workbench module is used for processing the published task and delivering the completed task to the ore deposit model sample library management module; and the prospecting forceful area delineation module is used for obtaining and processing data in the ore deposit model sample library, delineating a prospecting favorable area and performing task guidance on the sample library construction workbench module. By performing standardized processing, structured storage and intelligent management on multi-dimensional geoscience data of a typical ore deposit, a standardized sample set which can be used for machine learning, deep learning or expert system training is formed, and data support is provided for delineating a favorable prospecting area of an unknown area.

Description

Technical Field

[0001] The present invention relates to the technical field of spatial data sample library construction, and in particular to an intelligent construction system for a mineral deposit model spatial data sample library. Background Art

[0002] Against the backdrop of growing global resource demand, the exploration and development of mineral resources is crucial for ensuring national economic security and sustainable social development. Prospecting prediction, a crucial preliminary step in mineral exploration, aims to identify favorable areas with potential mineral exploration value through analysis of multi-source data, including geological, geophysical, and geochemical data. However, traditional prospecting prediction methods suffer from fragmented data, disorganized management, unscientific task allocation, and inaccurate prospecting target positioning, severely hindering their accuracy and efficiency. Summary of the Invention

[0003] The purpose of the present invention is to provide an intelligent construction system for a spatial data sample library of a mineral deposit model. By standardizing, structuring, and intelligently managing the multi-dimensional geological data of typical mineral deposits, a standardized sample set that can be used for machine learning, deep learning, or expert system training is formed, providing data support for the delineation of favorable prospecting areas in unknown areas.

[0004] To achieve the above objectives, the present invention provides a system for intelligently constructing a spatial data sample library of a mineral deposit model, the system comprising: The ore deposit model sample library management module is used to build the source space database, publish tasks to the sample library construction workbench module, monitor the published tasks, and reprocess the tasks delivered by the sample library construction workbench module to build the ore deposit model sample library; The sample library construction workbench module is used to process the release tasks and deliver the completed tasks to the deposit model sample library management module; The module for delineating favorable areas for mineral exploration is used to obtain data from the ore deposit model sample library for processing, delineate favorable areas for mineral exploration, and provide task guidance for the sample library construction workbench module.

[0005] Furthermore, the ore deposit model sample library management module includes: The data acquisition unit is used to collect spatial data of typical mineral deposits. It uses a multimodal data adaptive acquisition engine to parse and identify heterogeneous data in the spatial data of typical mineral deposits and align the spatial data of typical mineral deposits to a unified geographic coordinate system. The source space database unit is used to classify and manage typical mineral deposit spatial data based on mineral deposit genesis type, including the establishment of a classification system and data maintenance and update; The task publishing unit is used to publish the spatial data sample production task to the sample library construction workbench module for the typical mineral deposit spatial data in the source spatial database unit; Task monitoring unit, used to monitor the progress and workload of all sample preparation tasks in the sample library construction workbench module; The delivery task detection unit is used to receive tasks delivered by the sample library construction workbench module and perform quality detection on the delivered tasks; The ore deposit model sample warehouse unit is used to reprocess the typical ore deposit spatial data after sample production, build an ore deposit model sample library, and perform visual display.

[0006] Furthermore, the source space database unit includes: The hierarchical classification management subunit is used to establish a classification system based on the genetic type of the ore deposit to classify the spatial data of typical ore deposits; The spatial data lake-warehouse integrated management subunit is used to establish a unified data storage format and standard, and to store the data in the hierarchical and classified management units in multiple nodes through distributed storage technology; The data dynamic real-time update and maintenance sub-unit is used to record the update history of typical mineral deposit spatial data by establishing a data version control mechanism, and use data detection algorithms to repair and update errors and anomalies in the typical mineral deposit spatial data in the node.

[0007] Furthermore, the task issuing unit is specifically used to perform the following operations: Analyze the data in the source spatial database and extract the key features of the data; Process key features through machine learning algorithms to obtain the processing task requirements for each data; The processing task requirements are allocated to the processing units corresponding to the sample library construction workbench module through the matching algorithm.

[0008] Furthermore, the deposit model sample warehouse unit includes: The sample temporary management subunit is used to classify and manage the spatial data of typical mineral deposits after sample preparation; The sample data processing subunit is used to convert the spatial data of all typical mineral deposits in the sample temporary management subunit into a format using a vector-to-raster algorithm, and then cut and segment them into blocks according to uniform specifications to form single-channel sample sets and multi-channel sample sets; The sample data enhancement subunit is used to calculate the covariance matrix of the data in each single-channel sample set and multi-channel sample set, apply noise distribution perturbation to each data, and perform whitening / de-whitening processing through the covariance matrix, and finally perform geometric transformation synchronously; The ore deposit model sample management subunit is used to build an ore deposit model sample library, establish data description and metadata information for each single-channel data set, and use data fusion technology to integrate and associate multi-channel data sets to form a multi-channel block data set; The ore deposit model sample visualization subunit is used to visualize the ore deposit model sample library through cloud computing and big data processing technology, combined with GIS and Web visualization engine.

[0009] Furthermore, the sample library construction workbench module includes: Data visualization unit, used to visualize the spatial data of the deposit model in the release task, as well as the data in the positive sample creation, negative sample creation and annotation process; Positive sample production unit, used to produce positive samples based on typical mineral deposit spatial data in the release task; Negative sample production unit, used to produce negative samples based on typical mineral deposit spatial data in the release task; The quality inspection unit is used to perform quality inspection on the typical mineral deposit spatial data that has been marked. If it fails, positive samples or negative samples will be re-produced. If it passes, it will be delivered to the mineral deposit model sample library management module.

[0010] Furthermore, the positive sample production unit includes: The manual annotation subunit is used to manually annotate the spatial data of typical mineral deposits in the release task; The automatic annotation subunit is used to automatically annotate the spatial data of typical mineral deposits in the published tasks; The intelligent labeling subunit is used to intelligently label the spatial data of typical mineral deposits in the release task.

[0011] Furthermore, the positive sample production unit is specifically used to perform the following operations: Assign typical mineral deposit spatial data in the release task to different annotation units according to preset rules; Assign typical mineral deposit spatial data with clear structure and clear rules to automatic annotation subunits, and annotate the assigned data using an algorithm based on preset rules; Assign the remaining typical mineral deposit spatial data and the automatically labeled confidence level below the preset threshold to the intelligent labeling subunit, and label the assigned data using the deep learning model; Assign the typical mineral deposit spatial data with annotation confidence lower than the preset threshold in the intelligent annotation subunit to the manual annotation subunit, and have experts annotate the assigned data; The results of manual labeling, automatic labeling, and intelligent labeling are integrated, the area of ​​each positive sample region is calculated, and the summaries are performed to obtain the total area of ​​the positive samples.

[0012] Furthermore, the negative sample generation unit is specifically used to perform the following operations: Determine the distribution range of the abnormal area of ​​each layer of the typical mineral deposit spatial data in the release task, and integrate the distribution of abnormal areas of each layer to construct a multi-dimensional geological feature space; For each layer, the preset anomaly recognition algorithm is used to identify the abnormal areas in each layer, and the abnormal areas of all layers are spatially superimposed and merged to obtain the comprehensive coverage area of ​​all abnormal areas; Perform spatial difference calculation on the overall range of the study area and the comprehensive coverage area to obtain the blank area not covered by the abnormality. The blank area is the potential spatial distribution range of the negative sample. According to the total area of ​​positive samples, the target value of the total area of ​​negative samples is determined, the range of negative sample data is set in the blank area, and the spatial distribution of negative samples is automatically calculated in the blank area using a random function.

[0013] Furthermore, the module for identifying the promising areas for mineral exploration includes: The intelligent algorithm library unit is used to train the data in the ore deposit model sample library through machine learning algorithms and deep learning algorithms to obtain multiple AI prospecting favorable area delineation models. The AI ​​prospecting favorable area delineation models are used to provide task guidance for the sample library construction workbench module; The ore deposit model sample processing unit is used to analyze and process the data in the ore deposit model sample library to obtain a prediction data set; The favorable prospecting area delineation unit is used to input the predicted data set into the AI ​​favorable prospecting area delineation model for inference and delineate favorable prospecting areas.

[0014] Compared with the prior art, the present invention has the following beneficial effects: The present invention provides an intelligent construction system for a mineral deposit model spatial data sample library. The mineral deposit model sample library management module focuses on the construction of the source space database, task release and monitoring, and reprocessing after task delivery, and performs macro-control of the entire sample library construction process through centralized management. The sample library construction workbench module focuses on the specific processing of tasks. By operating various complex data processing tasks and accurately delivering the results, it improves the efficiency of sample library construction and provides a high-quality data foundation for subsequent mineral deposit model construction. The module for delineating promising areas for mineral exploration utilizes the high-quality data processed in the mineral deposit model sample library management module, combined with analysis algorithms, to quickly delineate areas with potential mineral resources, shorten the prediction cycle of mineral exploration targets, and improve the scientificity and reliability of prediction results; at the same time, its task guidance function can feed back the data requirements discovered during the delineation process to the sample library construction workbench module, providing direction for the subsequent optimization of the sample library. The present invention solves the problems of inefficient, scattered, and disconnected construction of the mineral deposit model knowledge base, establishes a knowledge closed-loop feedback mechanism, significantly improves the efficiency and accuracy of intelligent mineral exploration prediction, and promotes the intelligent and knowledge-based upgrade of mineral exploration. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without any creative work. Figure 1 A schematic diagram of the structure of an intelligent construction system for a mineral deposit model spatial data sample library provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0016] The present invention will be further described in detail below with reference to the accompanying drawings and examples. It will be understood that the specific embodiments described herein are intended only to illustrate the present invention and are not intended to limit the present invention. It should also be noted that, for ease of description, the accompanying drawings only illustrate portions relevant to the present invention, not all structures.

[0017] Reference Figure 1 This embodiment provides a system for intelligently constructing a mineral deposit model spatial data sample library, the system comprising: The ore deposit model sample library management module is used to build the source space database, publish tasks to the sample library construction workbench module, monitor the published tasks, and reprocess the tasks delivered by the sample library construction workbench module to build the ore deposit model sample library; The sample library construction workbench module is used to process the release tasks and deliver the completed tasks to the deposit model sample library management module; The module for delineating favorable areas for mineral exploration is used to obtain data from the ore deposit model sample library for processing, delineate favorable areas for mineral exploration, and provide task guidance for the sample library construction workbench module.

[0018] In this embodiment, the ore deposit model sample library management module manages the source space database and the ore deposit model sample library, including the construction of the source space database, task publishing and monitoring, and task delivery and reprocessing, thereby building the ore deposit model sample library. First, typical ore deposit spatial data is collected and standardized to build the source space database. Then, tasks are planned and published to the sample library construction workbench module, and the progress and status of task execution are monitored in real time. Finally, the delivered task results are received, data quality is checked and reprocessed, and the ore deposit model sample library is built to provide accurate data support for subsequent mineral exploration and prediction.

[0019] The Sample Library Construction Workbench module handles task reception and preparation, data processing and analysis, and task completion and delivery. It first receives tasks issued by the Mineral Deposit Model Sample Library Management module, extracts and prepares the required data from the source spatial database, then processes and analyzes the data according to the task requirements, applying appropriate algorithms. Finally, the processed and analyzed results are delivered to the Mineral Deposit Model Sample Library Management module, facilitating the construction and improvement of the Mineral Deposit Model Sample Library.

[0020] The Delineation of Promising Areas module implements data acquisition and preprocessing, delineation of favorable prospecting areas, and task guidance and feedback. The module first acquires and preprocesses multi-source data from a deposit model sample library. It then uses algorithms such as data mining to analyze the data and delineate favorable prospecting areas. Finally, based on the delineation results, it provides task guidance to the sample library construction workbench module and collects feedback to optimize the delineation method, thereby improving the accuracy and efficiency of prospecting.

[0021] As a preferred embodiment, the ore deposit model sample library management module includes: The data acquisition unit is used to collect typical mineral deposit spatial data. It uses the multimodal data adaptive acquisition engine to parse and identify the heterogeneous data in the typical mineral deposit spatial data, and align the typical mineral deposit spatial data to a unified geographic coordinate system.

[0022] The source space database unit is used to carry out hierarchical and classified management of typical mineral deposit spatial data based on mineral deposit genesis type, including the establishment of a classification system and data maintenance and update.

[0023] The task publishing unit is used to publish the spatial data sample production task to the sample library construction workbench module for the typical mineral deposit spatial data in the source spatial database unit.

[0024] The task monitoring unit is used to monitor the progress and workload of all sample production tasks in the sample library construction workbench module.

[0025] The delivery task detection unit is used to receive the tasks delivered by the sample library construction workbench module and perform quality detection on the delivered tasks.

[0026] The ore deposit model sample warehouse unit is used to reprocess the typical ore deposit spatial data after sample production, build an ore deposit model sample library, and perform visual display.

[0027] In this embodiment, the data acquisition unit collects and stores spatial data on typical mineral deposits, including structural, stratigraphic, rock mass, geochemical, geophysical, remote sensing, heavy sand, ore bodies, mineralization, alteration, and lithology data. A multimodal data adaptive acquisition engine is used to analyze heterogeneous data, automatically identifying different structures such as points, lines, bins, and raster data. This allows for intelligent spatial registration of coordinates, automatically matching them to a unified map projection coordinate system, ensuring the accuracy and integrity of the collected data and improving data quality.

[0028] The source space database unit realizes the management of typical mineral deposit spatial data. The typical mineral deposit model data is hierarchically and classified based on the mineral deposit genesis type, including the establishment of classification, data maintenance and update, and the managed data types include vector data and raster data.

[0029] The task publishing unit first obtains the spatial data of typical mineral deposits from the source spatial database unit, plans and defines specific spatial data sample production tasks according to the requirements of the mineral deposit model sample library construction, clarifies the goals and requirements of the tasks, and then publishes these tasks to the sample library construction workbench module through the system interface or interface.

[0030] The task monitoring unit includes progress monitoring and workload monitoring. Progress monitoring tracks the completion status of all sample production tasks in real time, counts the number of completed sample production tasks, and calculates the completion rate. Statistics on the number of completed sample production tasks and the completion rate for different mineral types are collected separately. This is used to analyze the progress of sample production for different mineral types, identify potential progress discrepancies, and adjust resource allocation in a timely manner.

[0031] Workload monitoring counts the number of sample preparation tasks for each mineral type and calculates its proportion of the total workload, providing a clear understanding of the distribution of sample preparation workload for different mineral types. Similarly, by combining mineral types and genetic types, we analyze the distribution of sample preparation workload for different types of deposits to rationally allocate resources and adjust task plans.

[0032] The delivery task detection unit ensures high-quality delivery of sample production and improves data reliability, including: Intelligent acceptance begins with an automated preliminary check of the sample production spatial data using machine learning algorithms. This data is compared with the original data in the source spatial database to test its completeness, accuracy, and consistency. For example, it checks whether the annotations of spatial data such as structures and strata are consistent with the actual geological characteristics and whether the spatial distribution of the data is reasonable.

[0033] Expert-assisted review: Based on the automatic preliminary inspection, an expert-assisted review mechanism is introduced. Geological experts conduct sampling review of the sample production spatial data, focusing on the sample production of complex geological features and key data, and provide review opinions.

[0034] Feedback and correction: For problems found during the acceptance process, a detailed feedback report is automatically generated and sent to the sample maker. The annotator corrects the sample production space annotation data based on the feedback report and submits it for acceptance again until it passes the acceptance.

[0035] Data submission: If the sample production space data passes the acceptance, the system will automatically submit the completed sample data package to the "Sample Temporary Management Subunit" using data privacy computing encryption technology to ensure the security and non-tamperability of the data, and record the submission time and related information.

[0036] The ore deposit model sample warehouse unit classifies and manages the data of completed sample production, and organizes and manages the data by ore type and ore deposit genesis type, including sample rasterization standard cutting, sample data enhancement, ore deposit model sample library management, and visualization processing.

[0037] As a preferred embodiment, the source space database unit includes: The hierarchical classification management subunit is used to establish a classification system according to the genesis type of the ore deposit and classify the spatial data of typical ore deposits.

[0038] The spatial data lake-warehouse integrated management sub-unit is used to establish a unified data storage format and standard, and to store the data in the hierarchical and classified management units in multiple nodes through distributed storage technology.

[0039] The data dynamic real-time update and maintenance sub-unit is used to record the update history of typical mineral deposit spatial data by establishing a data version control mechanism, and use data detection algorithms to repair and update errors and anomalies in the typical mineral deposit spatial data in the node.

[0040] In this embodiment, the hierarchical classification management subunit establishes a classification system according to different mineral deposit genesis types, such as endogenous mineral deposits, exogenous mineral deposits, metamorphic mineral deposits, etc., to accurately classify various types of spatial data.

[0041] The spatial data lake-warehouse integrated management subunit establishes unified data storage formats and standards to achieve efficient storage and management of different types of data. Furthermore, it uses distributed storage technology to disperse data across multiple nodes, increasing storage capacity and access speed, while enhancing data security and reliability.

[0042] The dynamic, real-time data update and maintenance subunit comprehensively manages the data structure and content of spatial data layers, such as structures and stratigraphic layers, for typical mineral deposit models. This includes update and maintenance tasks such as data addition, modification, and deletion. A data version control mechanism records data update history and ensures data traceability. Furthermore, automated data detection algorithms promptly identify errors and anomalies in the data, automatically repairing and updating them, improving the efficiency and accuracy of data updates and maintenance.

[0043] As a preferred embodiment, the task issuing unit is specifically configured to perform the following operations: Analyze the data in the source spatial database and extract the key features of the data.

[0044] The key features are processed through machine learning algorithms to obtain the processing task requirements for each data.

[0045] The processing task requirements are allocated to the processing units corresponding to the sample library construction workbench module through the matching algorithm.

[0046] In this embodiment, the task issuing unit first extracts key features of a typical spatial dataset of mineral deposit models from the source spatial data warehouse, such as data type, data volume, complexity, and data relevance. These features are processed using a machine learning algorithm to understand the unique requirements of each dataset. Based on the data feature analysis results, personalized sample preparation tasks are then automatically generated. An optimization algorithm is used to appropriately assign tasks to appropriate annotators, taking into account their skill level, experience, and workload. Sample preparation personnel also obtain tasks through encrypted authentication, ensuring data security and fair task allocation.

[0047] As a preferred embodiment, the ore deposit model sample warehouse unit includes: The sample temporary management subunit is used to classify and manage the spatial data of typical mineral deposits after sample preparation.

[0048] The sample data processing subunit is used to convert the format of all typical mineral deposit spatial data in the sample temporary management subunit through the vector-to-raster algorithm, and then cut and segment them according to uniform specifications to form single-channel sample sets and multi-channel sample sets.

[0049] The sample data enhancement subunit is used to calculate the covariance matrix of the data in each single-channel sample set and multi-channel sample set respectively, apply noise distribution perturbation to each data, and perform whitening / de-whitening processing through the covariance matrix, and finally perform geometric transformation synchronously.

[0050] The ore deposit model sample management subunit is used to build an ore deposit model sample library, establish data description and metadata information for each single-channel data set, and use data fusion technology to integrate and associate multi-channel data sets to form a multi-channel block data set.

[0051] The ore deposit model sample visualization subunit is used to visualize the ore deposit model sample library through cloud computing and big data processing technology, combined with GIS and Web visualization engine.

[0052] In this embodiment, the sample temporary management subunit implements classified management of typical ore deposit spatial data after sample preparation, and adopts classified organization management based on ore type and ore deposit genesis type.

[0053] The sample data processing subunit performs unified data analysis and processing on all typical mineral deposit spatial data of the sample temporary management subunit. It implements a dynamic programming adaptive resolution vector-to-raster algorithm through MPI scheduling of CPU and GPU collaborative computing to achieve format conversion. It then performs block cutting of uniform specifications (e.g., 28*28 pixels) to form single-channel sample sets and multi-channel sample sets. A single-channel sample set is a block data set of a single raster element of a typical mineral deposit spatial data, such as a single-channel data set of structure and a single-channel data set of Au element anomaly. A multi-channel sample set is a block data set of n elements of a typical mineral deposit spatial data, such as a four-channel block data set of structure, rock mass, Au element anomaly, and magnetic anomaly.

[0054] The sample data enhancement subunit generates high-quality enhanced samples for input data with multiple dimensions or modalities by preserving inter-channel correlations, mining cross-channel features, or introducing new transformation strategies to improve the model's generalization capabilities. Using multi-factor sample data enhancement technology, simulated noise is added to the metallogenic elements of the deposit model based on the spatial data of the deposit model, and the correlation between elements is maintained through single-factor covariance matrix constraints. First, the covariance matrix of each metallogenic element data is calculated to capture the correlation between elements. Second, a perturbation that conforms to the noise distribution is applied to each element, and covariance matrix whitening / de-whitening is performed to ensure that the perturbed data still conforms to the true data distribution. Then, geometric transformations are performed simultaneously to maintain spatial consistency. Finally, through the cross-scale CycleGAN sample data enhancement technology, a dual generator network is constructed to convert a small-scale metallogenic element into large-scale data to make up for the lack of samples, and the cycle consistency loss is used to ensure the reversibility of the conversion.

[0055] The ore deposit model sample management subunit implements a classification management system and data management technology, enabling efficient management of multiple complex data types and different channel data sets, improving the efficiency of data retrieval, call and analysis, while the data authority management mechanism ensures data security and privacy. Specifically, it includes: Classification management system, marking result classification management system, classification management according to mineral types and genetic types, and establishment of independent storage directories and indexes for sample data of different mineral types and genetic types.

[0056] Single-channel data sample set management: Create detailed data descriptions and metadata for each single-channel data set, including data source, acquisition time, data accuracy, etc. Quickly locate and obtain the required single-channel data set through keyword search and conditional screening.

[0057] Multi-channel data sample set management uses data fusion technology to integrate and associate data sets from different channels to form multi-channel segmented data sets. Users can use visualization tools to intuitively view the relationship between multi-channel data and conduct comprehensive analysis.

[0058] Data permission management,In order to ensure the security and privacy of data, a data permission management mechanism is introduced to,implement hierarchical authorization access to the annotation result data based on,the user's identity and permissions.

[0059] The mineral deposit model sample visualization subunit is based on cloud computing and big data processing technologies, combined with GIS and Web visualization engines, to achieve data query and visualization, including innovative technologies for global visualization, innovative technologies for typical mineral deposit model source space data visualization, and innovative technologies for sample library data visualization.

[0060] As a preferred embodiment, the sample library construction workbench module includes: The data visualization unit is used to visualize the spatial data of the deposit model in the release task, as well as the data in the positive sample preparation, negative sample preparation and labeling process.

[0061] The positive sample production unit is used to produce positive samples based on typical mineral deposit spatial data in the release task.

[0062] The negative sample production unit is used to produce negative samples based on typical mineral deposit spatial data in the release task.

[0063] The quality inspection unit is used to perform quality inspection on the typical mineral deposit spatial data that has been marked. If it fails, positive samples or negative samples will be re-produced. If it passes, it will be delivered to the mineral deposit model sample library management module.

[0064] In this embodiment, the general tools used in the sample library construction workbench to process the release task include: structural data processing tools, structural data processing tools, stratum processing tools, mineralization information extraction tools, point transition gridding tools, editing processing tools and spatial analysis tools. Positive sample production and annotation tools include: manual positive sample production and annotation tools, automated positive sample production and annotation tools and intelligent positive sample production and annotation tools. After the labeling is completed, the label management realizes the rapid query, viewing, modification and deletion of the input label information through a scientific classification management system and efficient operations, thereby improving the efficiency of data management, and providing users with flexible expansion capabilities to meet the personalized needs of different users. Providing optional recommendations when entering labels reduces the time and error rate of manual label entry, and improves the accuracy and efficiency of labeling, including classification management, operation functions, custom extensions and intelligent recommendations.

[0065] The data visualization unit uses professional visualization tools and technologies, such as geographic information system software and 3D visualization engines, to present the spatial data of the ore deposit model in the release task, as well as the data from the positive sample creation, negative sample creation, and annotation processes, in the form of intuitive graphics, charts, and maps. By setting different colors, symbols, and layers, it clearly presents the spatial distribution of the ore deposit, geological structural characteristics, distribution of sample data, and annotation information, enabling users to quickly understand and analyze the data, providing intuitive data support and decision-making basis for the construction of the ore deposit model sample library and subsequent prospecting and prediction work.

[0066] The positive sample production unit realizes data annotation through automatic annotation, intelligent annotation and manual annotation, including structural data space annotation, stratigraphic data space annotation, rock mass data space annotation, geochemical data space annotation, geophysical data space annotation, remote sensing data space annotation, heavy sand data space annotation, alteration data space annotation and lithology data space annotation.

[0067] The number of negative samples produced by the negative sample production unit is generally required to be equal to the number of positive samples. By producing negative samples based on spatial data, the strategy adopted is to make the total area of ​​negative samples close to the total area of ​​positive samples.

[0068] The quality inspection unit conducts quality inspection on the marked deposit models. Unqualified ones will be reworked and qualified ones will be delivered, including: Multi-source data fusion analysis organically combines data from different sources, utilizes the complementarity of multi-source data, identifies the characteristics and properties of mineral deposits, and improves the quality and reliability of data.

[0069] Deep learning model construction, based on deep learning algorithms, builds a neural network model for mineral deposit model data quality inspection. By learning and training a large amount of labeled mineral deposit model spatial data, it automatically extracts features and patterns in the data to achieve quality assessment of the mineral deposit model.

[0070] The automated quality inspection process combines multi-source data fusion analysis and deep learning models to realize automatic quality inspection of labeled mineral deposit models, including feature extraction, model prediction, result evaluation and other links. It can quickly and accurately detect errors and anomalies in mineral deposit models, and improve the efficiency and automation of quality inspection.

[0071] Visual display and interaction present the quality inspection results in an intuitive graphical manner. The platform can be used to view the quality assessment results of the deposit model, including indicators such as the model's accuracy, completeness, and consistency, while also providing interactive functions.

[0072] As a preferred embodiment, the positive sample preparation unit includes: The manual annotation subunit is used to manually annotate the spatial data of typical mineral deposits in the release task.

[0073] The automatic labeling subunit is used to automatically label the spatial data of typical mineral deposits in the release task.

[0074] The intelligent labeling subunit is used to intelligently label the spatial data of typical mineral deposits in the release task.

[0075] In this embodiment, positive sample production refers to the sample annotation of typical mineral deposit spatial data. The overall idea is to integrate the three methods of manual positive sample spatial annotation, automated positive sample spatial annotation, and intelligent positive sample spatial annotation, and combine label management and annotation tools to achieve comprehensive and efficient positive sample spatial annotation of the mineral deposit model and typical mineral deposit spatial data of typical mineral deposits. In the positive sample spatial annotation process, the experience and knowledge of geological experts and the advantages of artificial intelligence algorithms are fully considered, and the accuracy and efficiency of annotation are improved through human-computer interaction. At the same time, big data processing technology and cloud computing platforms are used to achieve rapid processing and storage of large-scale spatial data.

[0076] As a preferred embodiment, the positive sample preparation unit is specifically configured to perform the following operations: The typical mineral deposit spatial data in the release task are assigned to different annotation units according to preset rules.

[0077] The spatial data of typical mineral deposits with clear structure and clear rules are assigned to the automatic labeling subunit, and the assigned data are labeled using an algorithm based on preset rules.

[0078] The remaining and automatically labeled typical mineral deposit spatial data with confidence levels lower than a preset threshold are assigned to the intelligent labeling subunit, and the assigned data are labeled using a deep learning model.

[0079] The typical mineral deposit spatial data with annotation confidence lower than the preset threshold in the intelligent annotation subunit are assigned to the manual annotation subunit, and the assigned data are annotated by experts.

[0080] The results of manual labeling, automatic labeling, and intelligent labeling are integrated, the area of ​​each positive sample region is calculated, and the summaries are performed to obtain the total area of ​​the positive samples.

[0081] In this embodiment, a hierarchical labeling process of human-machine collaboration is used to achieve full coverage and accurate labeling of data from simple to complex. The controllability and traceability of data quality are ensured through the confidence threshold mechanism, providing high-confidence positive sample support for subsequent intelligent mineral prospecting models.

[0082] As a preferred embodiment, the negative sample generating unit is specifically configured to perform the following operations: Determine the distribution range of the abnormal area of ​​each layer of the typical mineral deposit spatial data in the release task, and integrate the distribution of the abnormal areas of each layer to construct a multi-dimensional geological feature space.

[0083] For each layer, the preset anomaly recognition algorithm is used to identify the abnormal areas in each layer, and the abnormal areas of all layers are spatially superimposed and merged to obtain the comprehensive coverage area of ​​all abnormal areas.

[0084] A spatial difference operation is performed between the overall range of the study area and the comprehensive coverage area to obtain the blank area that is not covered by the abnormality. The blank area is the potential spatial distribution range of the negative samples.

[0085] According to the total area of ​​positive samples, the target value of the total area of ​​negative samples is determined, the range of negative sample data is set in the blank area, and the spatial distribution of negative samples is automatically calculated in the blank area using a random function.

[0086] In this embodiment, the spatial distribution range of negative samples is set within the blank area not covered by the anomaly after the spatial superposition and merging of the anomaly areas in the spatial data of the typical mineral deposit. Within this blank area, two random parameters, the number and area of ​​negative samples, are set. A random function is used to automatically calculate the distribution of negative samples, with the requirement that the total area covered by the final calculated negative samples approximates the total area of ​​the positive samples.

[0087] 1) Select typical mineralization element data of mineral deposits Select the spatial data of typical mineral deposits, including structure, stratigraphy, lithology, remote sensing, alteration and other mineralization element layer data, and determine the distribution range of abnormal areas in each layer.

[0088] 2) Constructing a multidimensional geological feature space By integrating the data of each metallogenic element layer, a multidimensional geological feature space including structural features, stratigraphic features, lithologic features, remote sensing features, and alteration features is constructed. Assume that there are n metallogenic element layers, and the feature vector of the i-th layer is , where m is the feature dimension, the fused multidimensional geological feature space can be expressed as , providing rich geological information for subsequent identification of abnormal areas.

[0089] 3) Abnormal area identification and spatial superposition and merging For each metallogenic element layer, the preset anomaly recognition algorithm is used to identify the abnormal areas in each layer. Taking the threshold method based on statistical analysis as an example, let the characteristic value of the i-th layer be , whose mean is , the standard deviation is , setting the threshold coefficient to k, the judgment formula for the abnormal area is:

[0090] If the above conditions are met, the area is an abnormal area. When using a classification algorithm based on machine learning, the classification model is ,when , the area is determined to be an abnormal area.

[0091] Then, the abnormal areas of all mineralization element layers are spatially superimposed and merged. Represents the set of abnormal areas of the i-th layer, through The comprehensive coverage area of ​​all abnormal areas in the study area is obtained.

[0092] 4) Determine the spatial distribution range of negative samples Perform spatial difference calculation on the overall range R of the study area and the above-mentioned comprehensive anomaly coverage area A, namely , we get the blank area N not covered by the anomaly, which is the potential space distribution range of the negative sample.

[0093] 5) Set the number of negative samples and area random parameters According to the total area of ​​the positive samples , determine the target value of the total area of ​​negative samples ,satisfy ,in is the area fluctuation coefficient and can be adjusted according to actual conditions.

[0094] In the blank area N, the range of the number of negative samples is set to , the random parameter range of a single negative sample area is Assume that the blank area is ,in Indicates rounding up. Indicates rounding down.

[0095] 6) Automatic distribution calculation of negative samples The random function is used to automatically calculate the distribution of negative samples in the blank area.

[0096] Using the stratified random sampling method, the blank area N is divided into s sub-areas , let the area of ​​the jth sub-region be , the number of negative samples generated in the jth sub-region is ,satisfy: and in is the area of ​​a single negative sample generated in the jth sub-region, and Specifically, according to the set number of negative samples and the single area random parameter, the position and range of negative samples are randomly generated in the blank area, and the total area of ​​the generated negative samples is calculated in real time during the generation process. ,when Stop generating when t is the number of negative samples generated.

[0097] 7) Negative sample verification and adjustment Verify the generated negative samples to see if their spatial distribution is within the blank area N and if their total area is close to that of the positive samples. If there are any discrepancies, adjust the random parameters and recalculate the distribution of negative samples until the requirements are met.

[0098] As a preferred embodiment, the module for delineating the potential prospecting area includes: The intelligent algorithm library unit is used to train the data in the ore deposit model sample library through machine learning algorithms and deep learning algorithms to obtain multiple AI prospecting favorable area delineation models, and use the AI ​​prospecting favorable area delineation models to provide task guidance for the sample library construction workbench module.

[0099] The ore deposit model sample processing unit is used to analyze and process the data in the ore deposit model sample library to obtain a prediction data set; The favorable prospecting area delineation unit is used to input the predicted data set into the AI ​​favorable prospecting area delineation model for inference and delineate favorable prospecting areas.

[0100] In this embodiment, the intelligent algorithm library unit is used to perform a collection of machine learning and deep learning algorithms for prospecting and prediction, including random forests, support vector machines (SVMs), convolutional neural networks (CNNs), unified neural networks (UNets), and AlexNets. Training is performed using all algorithms in the "intelligent algorithm library" and the prospecting and prediction annotated dataset in the "ore deposit model sample library." This training generates multiple "AI model files" for use in the next step of inference. The AI ​​prospecting favorable area delineation model guides the task of constructing a workbench for the sample library, achieving an intelligent closed-loop "prediction-feedback-optimization." It uses model prediction results to reversely identify data defects (such as annotation blind spots and rule failures) and automatically generates supplementation, correction, and optimization tasks, upgrading the sample library from static storage to a continuously iterative "living knowledge engine." Based on confidence diagnosis based on model feedback, annotation resources are allocated in a targeted manner. Data with clear rules is automatically processed by the algorithm, and difficult data is prioritized for manual annotation by experts. New mineralization patterns mined by the model are converted into annotation rules in real time and injected into the workbench.

[0101] The ore deposit model sample processing unit uses fast spatial indexing and graph-cutting analysis techniques to extract corresponding spatial data from the ore deposit model sample library based on the user-defined prediction area spatial coordinate range, mineral type, genetic type, and prospecting prediction element layer data. This includes multi-scale, multi-element geophysical, chemical, remote sensing, and mineral-related vector or raster spatial data. The extracted data forms a "prediction area spatial layer dataset," which includes structural spatial data, stratigraphic spatial data, rock mass spatial data, geochemical spatial data, geophysical spatial data, remote sensing spatial data, and heavy sand spatial data.

[0102] The unit for delineating favorable areas for mineral exploration uses the algorithm model in the intelligent algorithm library and the corresponding "AI model file" to perform intelligent reasoning on the prediction data set formed after "rasterization and segmentation of spatial data". After the intelligent reasoning is completed, a favorable area for mineral exploration prediction is generated.

[0103] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An intelligent construction system for a mineral deposit model spatial data sample library, characterized in that: The system comprises: The ore deposit model sample library management module is used to build the source space database, publish tasks to the sample library construction workbench module, monitor the published tasks, and reprocess the tasks delivered by the sample library construction workbench module to build the ore deposit model sample library; The sample library construction workbench module is used to process the release tasks and deliver the completed tasks to the deposit model sample library management module; The module for delineating favorable areas for mineral exploration is used to obtain data from the ore deposit model sample library for processing, delineate favorable areas for mineral exploration, and provide task guidance for the sample library construction workbench module.

2. The intelligent construction system for the mineral deposit model spatial data sample library according to claim 1 is characterized in that: The ore deposit model sample library management module includes: The data acquisition unit is used to collect spatial data of typical mineral deposits. It uses a multimodal data adaptive acquisition engine to parse and identify heterogeneous data in the spatial data of typical mineral deposits and align the spatial data of typical mineral deposits to a unified geographic coordinate system. The source space database unit is used to classify and manage typical mineral deposit spatial data based on mineral deposit genesis type, including the establishment of a classification system and data maintenance and update; The task publishing unit is used to publish the spatial data sample production task to the sample library construction workbench module for the typical mineral deposit spatial data in the source spatial database unit; Task monitoring unit, used to monitor the progress and workload of all sample preparation tasks in the sample library construction workbench module; The delivery task detection unit is used to receive tasks delivered by the sample library construction workbench module and perform quality detection on the delivered tasks; The ore deposit model sample warehouse unit is used to reprocess the typical ore deposit spatial data after sample production, build an ore deposit model sample library, and perform visual display.

3. The intelligent construction system for the mineral deposit model spatial data sample library according to claim 2 is characterized in that: The source space database unit includes: The hierarchical classification management subunit is used to establish a classification system based on the genetic type of the ore deposit to classify the spatial data of typical ore deposits; The spatial data lake-warehouse integrated management subunit is used to establish a unified data storage format and standard, and to store the data in the hierarchical and classified management units in multiple nodes through distributed storage technology; The data dynamic real-time update and maintenance sub-unit is used to record the update history of typical mineral deposit spatial data by establishing a data version control mechanism, and use data detection algorithms to repair and update errors and anomalies in the typical mineral deposit spatial data in the node.

4. The intelligent construction system for the mineral deposit model spatial data sample library according to claim 2 is characterized in that: The task publishing unit is used to perform the following operations: Analyze the data in the source spatial database and extract the key features of the data; Process key features through machine learning algorithms to obtain the processing task requirements for each data; The processing task requirements are allocated to the processing units corresponding to the sample library construction workbench module through the matching algorithm.

5. The intelligent construction system for the mineral deposit model spatial data sample library according to claim 2 is characterized in that: The deposit model sample warehouse unit includes: The sample temporary management subunit is used to classify and manage the spatial data of typical mineral deposits after sample preparation; The sample data processing subunit is used to convert the spatial data of all typical mineral deposits in the sample temporary management subunit into a format using a vector-to-raster algorithm, and then cut and segment them into blocks according to uniform specifications to form single-channel sample sets and multi-channel sample sets; The sample data enhancement subunit is used to calculate the covariance matrix of the data in each single-channel sample set and multi-channel sample set, apply noise distribution perturbation to each data, and perform whitening / de-whitening processing through the covariance matrix, and finally perform geometric transformation synchronously; The ore deposit model sample management subunit is used to build an ore deposit model sample library, establish data description and metadata information for each single-channel data set, and use data fusion technology to integrate and associate multi-channel data sets to form a multi-channel block data set; The ore deposit model sample visualization subunit is used to visualize the ore deposit model sample library through cloud computing and big data processing technology, combined with GIS and Web visualization engine.

6. The intelligent construction system for the mineral deposit model spatial data sample library according to claim 1 is characterized in that: The sample library construction workbench modules include: Data visualization unit, used to visualize the spatial data of the deposit model in the release task, as well as the data in the positive sample creation, negative sample creation and annotation process; Positive sample production unit, used to produce positive samples based on typical mineral deposit spatial data in the release task; Negative sample production unit, used to produce negative samples based on typical mineral deposit spatial data in the release task; The quality inspection unit is used to perform quality inspection on the typical mineral deposit spatial data that has been marked. If it fails, positive samples or negative samples will be re-produced. If it passes, it will be delivered to the mineral deposit model sample library management module.

7. The intelligent construction system for the mineral deposit model spatial data sample library according to claim 6 is characterized in that: The positive sample production unit includes: The manual annotation subunit is used to manually annotate the spatial data of typical mineral deposits in the release task; The automatic annotation subunit is used to automatically annotate the spatial data of typical mineral deposits in the published tasks; The intelligent labeling subunit is used to intelligently label the spatial data of typical mineral deposits in the release task.

8. The intelligent construction system for the mineral deposit model spatial data sample library according to claim 7 is characterized in that: The positive sample production unit is specifically used to perform the following operations: Assign typical mineral deposit spatial data in the release task to different annotation units according to preset rules; Assign typical mineral deposit spatial data with clear structure and clear rules to automatic annotation subunits, and annotate the assigned data using an algorithm based on preset rules; Assign the remaining typical mineral deposit spatial data and the automatically labeled confidence level below the preset threshold to the intelligent labeling subunit, and label the assigned data using the deep learning model; Assign the typical mineral deposit spatial data with annotation confidence lower than the preset threshold in the intelligent annotation subunit to the manual annotation subunit, and have experts annotate the assigned data; The results of manual labeling, automatic labeling, and intelligent labeling are integrated, the area of ​​each positive sample region is calculated, and the summaries are performed to obtain the total area of ​​the positive samples.

9. The intelligent construction system for the mineral deposit model spatial data sample library according to claim 8, characterized in that: The negative sample generation unit is used to perform the following operations: Determine the distribution range of the abnormal area of ​​each layer of the typical mineral deposit spatial data in the release task, and integrate the distribution of abnormal areas of each layer to construct a multi-dimensional geological feature space; For each layer, the preset anomaly recognition algorithm is used to identify the abnormal areas in each layer, and the abnormal areas of all layers are spatially superimposed and merged to obtain the comprehensive coverage area of ​​all abnormal areas; Perform spatial difference calculation on the overall range of the study area and the comprehensive coverage area to obtain the blank area not covered by the abnormality. The blank area is the potential spatial distribution range of the negative sample. According to the total area of ​​positive samples, the target value of the total area of ​​negative samples is determined, the range of negative sample data is set in the blank area, and the spatial distribution of negative samples is automatically calculated in the blank area using a random function.

10. The intelligent construction system for the mineral deposit model spatial data sample library according to claim 1, characterized in that: The module for identifying potential prospecting areas includes: The intelligent algorithm library unit is used to train the data in the ore deposit model sample library through machine learning algorithms and deep learning algorithms to obtain multiple AI prospecting favorable area delineation models. The AI ​​prospecting favorable area delineation models are used to provide task guidance for the sample library construction workbench module; The ore deposit model sample processing unit is used to analyze and process the data in the ore deposit model sample library to obtain a prediction data set; The favorable prospecting area delineation unit is used to input the predicted data set into the AI ​​favorable prospecting area delineation model for inference and delineate favorable prospecting areas.

Images