Mineral detection method, system and equipment based on artificial intelligence and medium

Abstract

The invention relates to a mineral detection method, system and device based on artificial intelligence and a medium, and the method comprises the steps: constructing a three-dimensional ore body model through fusing geological exploration data, real-time market price and mineral separation technical parameters; based on a dynamic optimization algorithm and price fluctuation simulation, iteratively solving a grade-cost optimal solution, and generating recoverable boundary data under different price conditions; utilizing a machine learning model to carry out availability classification on the ore body blocks, and generating a spatial distribution data set containing recoverable resource and associated resource labels; and in combination with real-time market parameters and label data, the emergency resource potential is rapidly evaluated, and a mineable area map and a productivity planning scheme are output. According to the method, the limitation of traditional static threshold evaluation is broken through, dynamic optimization of resource availability and intelligent improvement of emergency response capability are realized, the utilization rate of low-grade ore and associated resources is remarkably improved, the resource guarantee decision-making period is shortened, and a scientific basis is provided for mineral development in a complex market environment.

Description

Technical Field

[0001] The present invention relates to the technical field of mineral resource detection, and in particular to a mineral detection method, system, equipment and medium based on artificial intelligence. Background Art

[0002] Efficient detection and accurate assessment of mineral resources have long been a core concern in the mining industry. With growing global demand for minerals and increasing market volatility, scientifically predicting resource availability within complex geological conditions and a dynamic economic environment has become a key challenge in improving mining efficiency and ensuring resource security. Traditional mineral exploration methods are often based on static geological models and empirical threshold settings, delineating mineable areas through statistical analysis of exploration data. While these methods have some applicability in stable market environments, they struggle to adapt to the real-time impacts of price fluctuations, cost changes, and unexpected events on resource assessment.

[0003] Existing technologies typically rely on fixed grade thresholds for mineral mineability assessments, failing to incorporate dynamic parameters such as market prices and breakthroughs in mineral processing technology. This results in a significant amount of idle low-grade ores and associated resources. Furthermore, traditional methods often require manual data recollection and model adjustments to address sudden resource supply disruptions. This results in long response times and insufficient decision-making basis, making it difficult to rapidly formulate emergency mining plans. Furthermore, the disconnected analysis of geological data and economic parameters further limits the adaptability of resource assessment results to market fluctuations. Summary of the Invention

[0004] Based on this, the purpose of the present invention is to provide an artificial intelligence-based mineral exploration method, system, equipment and medium that can dynamically integrate multi-source data, optimize mineability standards in real time and quickly generate emergency response plans.

[0005] The purpose of the present invention is achieved by the following scheme:

[0006] In a first aspect, the present invention provides a mineral exploration method based on artificial intelligence, comprising the following steps:

[0007] S1: Obtain geological exploration data of the target mining area, economic parameters related to the economic efficiency of the mineral resources, and mineral processing technical parameters related to the mineral processing process, and perform data fusion processing on the geological exploration data, economic parameters, and mineral processing technical parameters to construct a three-dimensional ore body model containing the ore body block data set and economic mineral processing constraint information;

[0008] S2: Monitor price fluctuations and mining costs in the mineral market in real time, generate real-time market resource data, input the market resource data into the three-dimensional ore body model for processing, calculate the dynamic mineability of the ore body resources, and generate a mineable boundary data set under different price conditions;

[0009] S3: Perform resource classification processing on the mineable boundary dataset, predict the availability of ore body blocks and assign labels using a machine learning model, and generate an ore body block dataset containing availability labels. The ore body block dataset is used to indicate the spatial distribution of mineable resources, associated resources, and temporarily unmineable resources.

[0010] S4: Conduct emergency resource assessment on the ore body block data set, screen mineable blocks based on availability tags, calculate emergency resource quantities based on real-time market resource data, and generate a mineable area coordinate map and production capacity planning plan.

[0011] In one embodiment, S1 of a mineral exploration method based on artificial intelligence provided by the present invention specifically includes the following steps:

[0012] S11: Based on the sensor network and market monitoring system, the exploration data, market prices and mineral processing process parameters of the target mining area are collected in real time to generate geological exploration data of the target mining area, economic parameters related to the economic efficiency of the mineral resources and mineral processing technical parameters related to the mineral processing process;

[0013] S12: Performing three-dimensional spatial reconstruction processing on the drill hole location data, ore composition data, and ore layer thickness data in the geological exploration data based on a spatial interpolation algorithm, calculating the ore composition distribution in the unsampled area using a Kriging interpolation algorithm, and generating three-dimensional ore body morphology data;

[0014] S13: performing regular grid division processing on the three-dimensional morphological data of the ore body, dividing the ore body into cubic units of equal volume according to a preset size, and generating an ore body block data set containing position coordinates, ore grade, and reserve information;

[0015] S14: Based on the historical market price data in the economic parameters and the recovery efficiency data in the mineral processing technical parameters, the ore body block data set is subjected to multi-dimensional data fusion processing. The economic constraints and geological parameters are integrated through the matrix weighting algorithm to construct a three-dimensional ore body model containing the ore body block data set and economic mineral processing constraint information.

[0016] In one embodiment, S2 of the artificial intelligence-based mineral exploration method provided by the present invention specifically includes the following steps:

[0017] S21: Based on the market data collection system, the price fluctuation data and mining costs of the mineral market are monitored in real time. The price trend characteristics and cost fluctuation parameters are extracted through the time series analysis algorithm to generate real-time market resource data.

[0018] S22: Inputting real-time market resource data into the three-dimensional ore body model to simulate price fluctuations, combining the cost threshold and ore recovery rate parameters of the economic ore processing constraint information in the three-dimensional ore body model to generate a simulation data set containing multiple sets of dynamic price change scenarios;

[0019] S23: Based on the dynamic optimization algorithm, the mineability calculation process is performed on the simulation data set and the ore body block data set in the three-dimensional ore body model, and the grade-cost optimal solution of the ore body block is solved through the objective function iteration to generate a dynamic mineable grade threshold;

[0020] S24: Fit the spatial distribution data of the ore body based on the dynamic mineable grade threshold, use the Kriging spatial interpolation algorithm to generate the boundaries of the continuous mineable area, and output the mineable boundary data set under different price conditions.

[0021] In one embodiment, S23 of the artificial intelligence-based mineral exploration method provided by the present invention specifically includes the following steps:

[0022] S231: performing objective function decomposition processing on the simulation data set and the ore body block data set in the three-dimensional ore body model based on a dynamic optimization algorithm, and splitting the objective function of the mineability calculation into an ore grade constraint term and an economic cost constraint term;

[0023] S232: Iteratively optimize the ore grade constraint and the economic cost constraint, and solve the grade-cost joint optimal solution using a gradient descent algorithm until the change in the mineable grade threshold satisfies a preset convergence condition, thereby generating a converged grade threshold;

[0024] S233: Perform economic feasibility verification on the convergent grade threshold, screen the threshold results that meet the mining profit constraints, and generate a dynamic mineable grade threshold.

[0025] In one embodiment, S3 of the artificial intelligence-based mineral exploration method provided by the present invention specifically includes the following steps:

[0026] S31: Perform feature extraction on the ore body blocks in the mineable boundary dataset, calculate the grade gradient of each block using the sliding window method, generate the grade difference value of the adjacent area using the neighborhood analysis algorithm, and combine it with the geological structure density statistical method to generate a multidimensional feature dataset containing grade gradient, neighborhood grade difference and geological structure density;

[0027] S32: Processing the multidimensional feature data set based on the pre-trained machine learning model, generating the probability value of the availability of each ore body block through feature standardization and model inference calculation;

[0028] S33: Perform dynamic threshold determination processing on the availability probability value to generate an ore body block data set containing availability labels. The availability labels include classification identifiers of recoverable resources, associated resources, and temporarily unrecoverable resources.

[0029] In one embodiment, S32 of the artificial intelligence-based mineral exploration method provided by the present invention specifically includes the following steps:

[0030] S321: Processing the multidimensional feature dataset based on the feature standardization method, eliminating feature dimension differences through the Z-score standardization algorithm, and generating a standard feature dataset;

[0031] S322: Perform inference calculations on the standard feature data set through a pre-trained machine learning model, calculate the feature weights and probability distribution of each block layer by layer based on the forward propagation algorithm, and generate the probability value of the availability of each ore body block.

[0032] In one embodiment, S4 of the artificial intelligence-based mineral exploration method provided by the present invention specifically includes the following steps:

[0033] S41: Perform emergency resource assessment on the ore body block data set, filter the data set based on its availability tags, extract minable blocks and associated resource blocks, and generate an emergency resource candidate set;

[0034] S42: Based on the price and cost parameters in the real-time market resource data, an economic evaluation is performed on the candidate set of emergency resources, and the recoverable resource volume and potential resource volume are calculated through a cost-benefit analysis model to generate an emergency resource volume including economic evaluation indicators;

[0035] S43: Performing geospatial mapping of the emergency resource volume, associating the resource volume data with the ore body coordinates through GIS spatial analysis technology, and generating a mineable area coordinate map with the thermal value of the mineable area marked;

[0036] S44: Based on the multi-objective optimization algorithm, the coordinate map of the mineable area is optimized for production capacity, balancing mining costs, market demand and resource sustainability parameters, and generating a production capacity planning scheme with priority sorting.

[0037] In a second aspect, the present invention provides an artificial intelligence-based mineral exploration system, which is configured with the following modules:

[0038] Data acquisition module, used to obtain geological exploration data of the target mining area, economic parameters related to the economics of the mineral products, and mineral processing technical parameters related to the mineral processing technology;

[0039] The model building module is used to fuse geological exploration data, economic parameters and mineral processing technical parameters to construct a three-dimensional ore body model containing ore body block data sets and economic mineral processing constraint information;

[0040] The mineability calculation module is used to monitor price fluctuations and mining costs in the mineral market in real time, generate real-time market resource data, input the market resource data into the three-dimensional ore body model for processing, calculate the dynamic mineability of the ore body resources, and generate a mineable boundary data set under different price conditions;

[0041] The label assignment module is used to classify the resources of the mineable boundary dataset, predict the availability of ore blocks and assign labels through machine learning models, and generate ore block datasets containing availability labels. The ore block datasets are used to indicate the spatial distribution of mineable resources, associated resources, and temporarily unmineable resources.

[0042] The resource assessment and planning module is used to perform emergency resource assessment on the ore body block data set, screen the mineable blocks according to the availability tags, calculate the emergency resource volume in combination with real-time market resource data, and generate the mineable area coordinate map and production capacity planning scheme.

[0043] In a third aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, it implements any of the above-mentioned artificial intelligence-based mineral exploration methods.

[0044] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-mentioned artificial intelligence-based mineral exploration methods.

[0045] In summary, the artificial intelligence-based mineral exploration method provided by the present invention can achieve efficient integration of geological exploration data, economic parameters and mineral processing process parameters and accurate construction of three-dimensional ore body models through the synergistic effect of multi-source data fusion and dynamic optimization algorithm, so as to break through the response hysteresis of traditional static evaluation models to market price fluctuations and changes in mining costs; the dynamic mineability calculation mechanism based on real-time market resource data can achieve real-time optimization and multi-scenario adaptation of the mineable boundary of ore body resources, and effectively solve the problem of idle low-grade mines and associated resources caused by fixed grade thresholds; through the intelligent prediction and label classification of the availability of ore body blocks by machine learning models, it can achieve refined analysis of resource spatial distribution characteristics, and provide multi-dimensional data support for emergency resource assessment; the emergency resource potential assessment system combined with real-time market parameters and label data can achieve dynamic matching of mining plans and market fluctuations, so as to achieve rapid response and scientific decision-making in the scenario of sudden resource supply interruption.

[0046] For better understanding and implementation, the present invention is described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 A schematic diagram of a flow chart of a mineral exploration method based on artificial intelligence provided in an embodiment of the present application;

[0048] Figure 2 A schematic diagram of a process for generating a mineable boundary dataset under different price conditions provided in an embodiment of the present application;

[0049] Figure 3 A schematic diagram of a process for generating an ore body block dataset containing availability tags provided in an embodiment of the present application;

[0050] Figure 4 A schematic structural diagram of an artificial intelligence-based mineral exploration system provided in another embodiment of the present application. DETAILED DESCRIPTION

[0051] To facilitate understanding of the present invention, the present invention will be described more fully below with reference to the accompanying drawings. The drawings illustrate preferred embodiments of the present invention. However, the present invention may be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and comprehensive understanding of the disclosure.

[0052] Unless otherwise defined, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification are for the purpose of describing specific embodiments only and are not intended to limit the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0053] In one embodiment, Figure 1 As shown, a mineral exploration method based on artificial intelligence is provided. This embodiment uses the method applied to a terminal as an example for illustration. It is understandable that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0054] S1: Obtain geological exploration data of the target mining area, economic parameters related to the economic efficiency of the mineral resources, and mineral processing technical parameters related to the mineral processing process, and fuse the geological exploration data, economic parameters, and mineral processing technical parameters to construct a three-dimensional ore body model containing the ore body block data set and economic mineral processing constraint information.

[0055] Specifically, the system collects comprehensive geological exploration data from the target mining area, including but not limited to drill hole data, geological profiles, geophysical exploration data, and geochemical exploration data. Geophysical exploration data includes gravity, magnetic, and electromagnetic measurement results, while geochemical exploration data includes soil sampling and rock analysis results. This data can provide key information such as the geological structure, ore body distribution, and rock type of the mining area. At the same time, it collects economic parameters related to the economic viability of mineral products, such as historical data on market price fluctuations of mineral products, transportation costs, labor costs, and equipment depreciation expenses. In addition, the system also obtains mineral processing technical parameters related to the mineral processing process, such as mineral processing recovery rate, concentrate grade, and mineral processing reagent costs.

[0056] The system integrates the collected geological exploration data, economic parameters, and mineral processing parameters through data fusion. For geological exploration data, three-dimensional interpolation of drill hole data is performed to generate a three-dimensional framework of the mining area's geological body. Geophysical and geochemical data are used to assign physical properties to the geological body, further enhancing the authenticity of the geological model. For economic parameters and mineral processing technical parameters, a parameter database associated with the geological model is established. Through spatial analysis and data association algorithms, economic and mineral processing information is integrated into the geological model. Ultimately, a three-dimensional ore body model is constructed, containing a collection of ore body block data and economic mineral processing constraint information. This model intuitively displays the ore body's geological characteristics and the spatial distribution of its economic and mineral processing attributes.

[0057] S2: Monitor price fluctuations and mining costs in the mineral market in real time, generate real-time market resource data, input the market resource data into the three-dimensional ore body model for processing, calculate the dynamic mineability of the ore body resources, and generate a mineable boundary data set under different price conditions.

[0058] Specifically, the system connects to the mineral market information platform and enterprise cost management system via a real-time data interface, continuously acquiring data on mineral market price fluctuations and changes in mining costs. Market price data includes spot prices from major global metal exchanges, futures price indices, and price differences between different grades of mineral products. Mining cost data covers energy price fluctuations, equipment maintenance cost adjustments, and mining operating costs. The system cleans and standardizes this real-time data and integrates it into a real-time market resource dataset, including time series curves of price changes and trend analysis charts of cost changes. This data is then fed into a pre-built three-dimensional ore body model, and the system calculates the dynamic recoverability of the ore resources based on an economic calculation model.

[0059] The economic calculation model comprehensively considers the geological characteristics, mining costs, beneficiation costs, and market prices of each ore body segment, using economic indicators such as net present value (NPV), internal rate of return (IRR), and break-even grade for evaluation. Geological characteristics include grade, thickness, and ore type; mining costs include blasting, excavation, and transportation costs; and beneficiation costs are calculated based on beneficiation processes and reagent consumption. For example, the system calculates the NPV value of each ore body segment under different price conditions. When the market price is above a certain threshold, the segment is considered economically mineable; when the price is below the threshold, the segment is deemed unmineable. Based on these calculation results, the system dynamically generates a dataset of mineable boundaries under different price conditions. This dataset digitally records the boundary coordinates of mineable and unmineable ore bodies, the corresponding price conditions, and economic benefit indicators. As market prices and costs fluctuate in real time, the system automatically updates the mineable boundaries, ensuring the timeliness and accuracy of mineral resource assessments and providing real-time support for mining companies' mining decisions.

[0060] S3: Perform resource classification processing on the mineable boundary dataset, predict the availability of ore body blocks and assign labels through machine learning models, and generate ore body block datasets containing availability labels. The ore body block datasets are used to indicate the spatial distribution of mineable resources, associated resources, and temporarily unmineable resources.

[0061] Specifically, the system divides mineral resources according to preset classification rules and algorithms. The classification process is based on the analysis of the geological characteristics of the ore body blocks, including grade, thickness, and spatial connectivity, and is combined with their location in the mineable boundary dataset and the corresponding price conditions. Based on this, the system divides mineral resources into three categories: mineable resources, which refer to ore body blocks with clear mining value under current economic conditions. Their grade and thickness meet the requirements of mineral processing technology and have significant economic benefits; associated resources, which refer to ore body blocks that coexist with mineable resources but are uneconomical to mine alone due to low grade or mineral processing technology limitations, but can be mined together with the main mineral mining process and have potential utilization value; temporarily unmineable resources, which refer to ore body blocks that have no mining value under current technical and economic conditions due to too low grade, too thin thickness, or too high mining costs.

[0062] To further enhance the accuracy of resource assessment, the system utilizes a machine learning model to predict and label the exploitability of ore body segments. This model is trained on a vast amount of historical mineral data. This data covers geological characteristics of the ore body segments, such as grade distribution curves and ore body morphology; economic parameters, such as price and cost data from different historical periods; mineral processing technical parameters, such as the history of process improvements and corresponding recovery rate changes; and actual mining and utilization data, such as whether the ore body segments are mined and their economic benefits after mining. During application, the system inputs the model's current geological, economic, and technical parameters for the ore body segments. The model, through complex algorithms, outputs a probability of exploitability for each segment. The system then assigns an exploitability label to each ore body segment based on a predefined probability threshold, such as labeling a probability greater than 80% as high, 50% to 80% as medium, and below 50% as low.

[0063] Ultimately, the system generates a dataset containing detailed exploitability labels for each ore block. This dataset not only includes the spatial coordinates and geological attributes of the ore blocks, but also records their exploitability labels and corresponding prediction confidence levels. This dataset comprehensively indicates the distribution of recoverable resources, associated resources, and currently unrecoverable resources in three dimensions, providing strong data support for subsequent resource development planning and technological improvements, helping mining companies to more rationally prioritize and prioritize resource development efforts.

[0064] S4: Conduct emergency resource assessment on the ore body block data set, screen mineable blocks based on availability tags, calculate emergency resource quantities based on real-time market resource data, and generate a mineable area coordinate map and production capacity planning plan.

[0065] Specifically, the system selects mineable blocks based on the availability tags in the ore body block data set. The screening process follows preset logical rules, and the system compares the availability tags of each ore body block with the mineability standards one by one. For example, blocks labeled as high or medium availability and meeting specific economic and mineral processing conditions will be selected. The system extracts these qualified blocks and records key information such as their spatial location, ore reserves, and quality parameters. At the same time, the system reads and integrates the latest market resource data in real time, including current mineral product market prices, price fluctuation trends, market supply and demand, and changes in mining costs.

[0066] The system integrates the parameters of the selected mineable blocks and real-time market data into an emergency resource calculation model. This model considers several key factors: ore reserves calculated based on the volume of the ore body and ore density; ore grade, which affects the total recoverable amount; mining recovery rates determined by mining processes and equipment performance; beneficiation recovery rates calculated based on beneficiation technical parameters; and market price factors, which determine the economic value of the resource. Through a series of formulas and algorithms, the model calculates the emergency resource volume that these mineable blocks can quickly bring to market in the event of an unexpected resource supply disruption. The results include detailed emergency production estimates for different metals or minerals, their corresponding economic value, and a cost-benefit analysis.

[0067] After calculating the emergency resource volume, the system further generates a coordinate map of the mineable area and a production capacity planning plan. To create the coordinate map, the system converts the spatial coordinate data of the selected mineable blocks into geographic coordinate format and incorporates spatial information such as the mining area's topography, transportation routes, and mining equipment layout. Using geographic information system technology, the system creates a distribution map that accurately displays the location, extent, and interrelationships of the mineable area, providing a visual basis for equipment positioning and operation path planning. The production capacity planning plan is based on the emergency resource volume of the mineable blocks, an optimization algorithm for the mining sequence that considers ore grade gradients and mining difficulty, and the company's production capacity constraints, such as equipment processing capacity, staffing, and capital investment. By simulating different mining scenarios and time schedules, the system generates a detailed plan that includes mining phase divisions, production targets for each phase, equipment configuration plans, and personnel scheduling. This ensures rapid, efficient, and economical mining of mineral resources in emergency situations, safeguarding the economic benefits of mining companies and market supply stability.

[0068] In summary, the artificial intelligence-based mineral exploration method provided by the present invention can achieve efficient integration of geological exploration data, economic parameters and mineral processing process parameters and accurate construction of three-dimensional ore body models through the synergistic effect of multi-source data fusion and dynamic optimization algorithm, so as to break through the response hysteresis of traditional static evaluation models to market price fluctuations and changes in mining costs; the dynamic mineability calculation mechanism based on real-time market resource data can achieve real-time optimization and multi-scenario adaptation of the mineable boundary of ore body resources, and effectively solve the problem of idle low-grade mines and associated resources caused by fixed grade thresholds; through the intelligent prediction and label classification of the availability of ore body blocks by machine learning models, it can achieve refined analysis of resource spatial distribution characteristics, and provide multi-dimensional data support for emergency resource assessment; the emergency resource potential assessment system combined with real-time market parameters and label data can achieve dynamic matching of mining plans and market fluctuations, so as to achieve rapid response and scientific decision-making in the scenario of sudden resource supply interruption.

[0069] The mineral exploration provided by the present invention significantly improves the adaptability of resource assessment models to complex economic environments and geological conditions through the technical path of dynamic data drive and intelligent algorithm iteration. It can break through the technical bottlenecks of traditional methods in data segmentation analysis, threshold setting and emergency response lag, effectively shorten the resource security decision-making cycle, and provide intelligent technical support for the sustainable development of mineral resources and strategic reserve management.

[0070] In one embodiment, S1 of a mineral exploration method based on artificial intelligence provided by the present invention specifically includes the following steps:

[0071] S11: Based on the sensor network and market monitoring system, the exploration data, market prices and mineral processing process parameters of the target mining area are collected in real time to generate the geological exploration data of the target mining area, economic parameters related to the economic efficiency of the mineral and mineral processing technical parameters related to the mineral processing process.

[0072] Specifically, the system uses a sensor network and market monitoring system to collect real-time exploration data, market prices, and mineral processing parameters for target mining areas. The sensor network, widely distributed across the mining area, accurately captures geological information, including key data such as drill hole location, ore composition, and ore layer thickness. These sensors utilize advanced measurement technologies, such as laser ranging and spectral analysis, to ensure accurate and real-time data. Simultaneously, the market monitoring system closely monitors dynamic changes in the mineral market, collecting real-time data on market price fluctuations, recovery efficiency data, and other economically relevant information within mineral processing parameters. This data is systematically integrated to generate geological exploration data for the target mining area, economic parameters related to the economic viability of the mineral, and mineral processing parameters related to the mineral processing process, providing a comprehensive data foundation for subsequent mineral exploration. In this process, the system fully utilizes its efficient data collection and integration capabilities to ensure timely and accurate recording of various data types, providing a rich and reliable data source for subsequent processing steps.

[0073] S12: Based on the spatial interpolation algorithm, the drill hole location data, ore composition data and ore layer thickness data in the geological exploration data are reconstructed in three dimensions. The ore composition distribution in the unsampled area is calculated by the Kriging interpolation algorithm to generate three-dimensional morphological data of the ore body.

[0074] Specifically, after acquiring geological exploration data, the system performs three-dimensional spatial reconstruction on the drill hole location data, ore composition data, and ore layer thickness data. Preferably, the system uses a spatial interpolation algorithm to locate the drill hole location data in three-dimensional space, constructing an initial framework for the mining area. Then, based on the distribution of ore composition data at each drill hole location, the Kriging interpolation algorithm is applied to calculate the ore composition distribution in unsampled areas. The Kriging interpolation algorithm analyzes the correlation between the ore composition and spatial location of known sampling points to establish a variogram model and predict the ore composition of unknown areas. The system then performs a three-dimensional fitting based on the ore layer thickness data and, combined with the interpolation results of the ore composition, generates three-dimensional morphological data for the ore body.

[0075] For example, when processing a metal deposit, the system uses drill hole location data to determine the approximate extent of the ore body. Using the metal content information from the ore composition data, the system uses a kriging interpolation algorithm to calculate the metal content distribution within unsampled areas within the ore body. Simultaneously, the undulating morphology of the ore body is constructed based on the ore layer thickness data. The resulting three-dimensional ore body morphology data details the ore body's boundaries, internal structure, and spatial variations in ore composition. This provides a precise geometric model for subsequent ore body segmentation and resource assessment, ensuring the accuracy and reliability of mineral resource assessments.

[0076] S13: performing regular grid division processing on the three-dimensional morphological data of the ore body, dividing the ore body into cubic units of equal volume according to a preset size, and generating an ore body block data set containing position coordinates, ore grade, and reserve information.

[0077] Specifically, the system divides the ore body into cubic cells of equal volume based on preset size parameters. During the division process, the system first determines the minimum circumscribed cuboid of the ore body and then performs grid division according to the preset cubic cell side lengths. For cubic cells located near the ore body boundary, the system determines their spatial relationship with the ore body's three-dimensional morphological data and uses truncation or merging to ensure that all cubic cells are completely contained within the ore body. The position coordinates of each cubic cell are calculated using its row and column numbers in a regular grid, and the system records the three-dimensional coordinates of each cell's center point. Ore grade is calculated by calculating the spatial average of the ore composition within the cell, while reserve information is calculated based on the volume and ore density of the cubic cells obtained from the geological database based on the ore type. The system generates a data set for the ore body's segmented data, containing position coordinates, ore grade, and reserve information. Each data item corresponds to a cubic cell, providing a structured data foundation for subsequent integration of economic beneficiation constraint information and resource assessment, ensuring the refinement and accuracy of resource assessment.

[0078] S14: Based on the historical market price data in the economic parameters and the recovery efficiency data in the mineral processing technical parameters, the ore body block data set is subjected to multi-dimensional data fusion processing. The economic constraints and geological parameters are integrated through the matrix weighting algorithm to construct a three-dimensional ore body model containing the ore body block data set and economic mineral processing constraint information.

[0079] Specifically, the system establishes a data fusion model, correlating historical market price data with the ore grade of each ore body segment, and calculating the potential sales revenue of each segment under different price conditions. Simultaneously, it combines mineral processing recovery efficiency data and ore composition to calculate the recoverable amount of metal or minerals during the beneficiation process. A matrix weighting algorithm integrates economic constraints such as mining and beneficiation costs with geological parameters such as ore grade and reserves. The matrix weighting algorithm constructs a weight matrix based on pre-set weight coefficients, quantizes the data for each factor, and then performs matrix operations to obtain the integrated economic beneficiation constraint information for each ore body segment.

[0080] The system integrates the fused economic beneficiation constraint information for each ore body block with the ore body block data set to construct a three-dimensional ore body model that incorporates both the ore body block data set and the economic beneficiation constraint information. This model is stored digitally in the system database and uses a visual interface to display the ore body's geological characteristics and the spatial distribution of its economic beneficiation attributes. This provides comprehensive and accurate data support for mining companies' resource assessment and mining decisions, ensuring the economic feasibility and efficiency optimization of mineral resource development.

[0081] In one embodiment, Figure 2 As shown, S2 of the mineral exploration method based on artificial intelligence provided by the present invention specifically includes the following steps:

[0082] S21: Based on the market data collection system, the price fluctuation data and mining costs of the mineral market are monitored in real time. The price trend characteristics and cost fluctuation parameters are extracted through the time series analysis algorithm to generate real-time market resource data.

[0083] Specifically, the system collects real-time data from global mineral market trading platforms, covering spot and futures price trends for major metals and minerals, as well as economic parameters closely related to the economic viability of mineral production, such as transportation and labor costs. The system also monitors real-time changes in mining costs, including fluctuations in energy prices and equipment maintenance expenses.

[0084] To accurately extract price trend characteristics and cost fluctuation parameters, the system utilizes a time series analysis algorithm. This algorithm decomposes collected market price data, extracting features such as short-term fluctuation amplitude, medium-term trend direction, and long-term cyclical changes. For cost fluctuations, the algorithm analyzes the fluctuation range and frequency of various costs and identifies key parameters of cost fluctuations. The system integrates these features and parameters to generate real-time market resource data, including time series curves of price changes and statistical analysis results of cost fluctuations. This data not only reflects current market dynamics but also provides accurate input for subsequent price fluctuation simulations and feasibility calculations, ensuring timely and accurate decision-making.

[0085] S22: Inputting real-time market resource data into the three-dimensional ore body model to simulate price fluctuations, and combining the cost threshold and ore processing recovery rate parameters of the economic ore processing constraint information in the three-dimensional ore body model to generate a simulation data set containing multiple sets of dynamic price change scenarios.

[0086] Specifically, the system utilizes the economic mineral processing constraint information stored in the three-dimensional ore body model, including cost thresholds and mineral processing recovery rate parameters, to generate multiple simulated data sets for dynamic price change scenarios. During the simulation process, the system first adjusts the potential sales revenue of the ore body segments based on real-time market price data. Then, combined with the cost threshold, it determines which ore body segments are economically viable under different price conditions. The mineral processing recovery rate parameter is used to calculate the actual recoverable resource volume of the ore body segments under different price conditions. By simulating different price change scenarios, the system generates simulated data sets containing the economic value of the ore body segments under each scenario. These data sets not only cover the mineability of the ore body segments under different price conditions but also provide detailed economic benefit analysis, providing comprehensive data support for subsequent mineability calculations and ensuring scientific and comprehensive decision-making.

[0087] S23: Based on the dynamic optimization algorithm, the mineability calculation and processing of the simulation data set and the ore body block data set in the three-dimensional ore body model are performed, and the grade-cost optimal solution of the ore body block is solved through the objective function iteration to generate a dynamic mineable grade threshold.

[0088] Specifically, the system uses the grade and cost of the ore body blocks as core variables to construct an objective function, aiming to maximize the economic benefits of mining. During the calculation process, the system comprehensively considers the ore grade, mining cost, mineral processing recovery cost and the impact of market price fluctuations of the ore body blocks. Through the iterative solution of the objective function, the system gradually approaches the grade-cost optimal solution of the ore body blocks. This process involves a detailed economic evaluation of each ore body block to determine the optimal mining strategy for each block under different price conditions. Finally, the system generates dynamic mineable grade thresholds, which reflect the economic mineability boundaries of the ore body blocks under current market conditions. The generation of dynamic mineable grade thresholds provides key quantitative indicators for the subsequent definition of mineable areas, helping mining companies to make accurate mining decisions and ensure the economic feasibility and efficiency optimization of resource development. Preferably, the dynamic mineable grade thresholds are generated through the following steps:

[0089] S231: Based on the dynamic optimization algorithm, the objective function decomposition processing is performed on the simulation data set and the ore body block data set in the three-dimensional ore body model, and the objective function of the mineability calculation is split into the ore grade constraint item and the economic cost constraint item.

[0090] Specifically, the system applies a dynamic optimization algorithm to decompose the objective function of the simulation dataset and the ore body block data set within the 3D orebody model. The system first considers the overall objective function and splits it into ore grade constraints and economic cost constraints. The ore grade constraint is set based on the ore grade data within the ore body block data set, reflecting the impact of ore quality on mineability and ensuring that the incoming ore grade meets industrial requirements. The economic cost constraint is constructed by combining economic parameters such as mining cost and beneficiation cost data to reflect the economic constraints of the mining and beneficiation processes and ensure the economic benefits of mining activities.

[0091] Through mathematical transformation, the system decomposes the objective function into two key parts: ore grade constraint and economic cost constraint, preparing for subsequent iterative optimization processing and ensuring that each constraint is fully considered in the optimization process, thereby accurately calculating the mineability of the ore body blocks and improving the accuracy and reliability of resource assessment.

[0092] S232: Iteratively optimize the ore grade constraint and the economic cost constraint, and solve the grade-cost joint optimal solution through the gradient descent algorithm until the change in the mineable grade threshold meets the preset convergence condition, thereby generating a converged grade threshold.

[0093] Specifically, the system first initializes the mineable grade threshold, and then during the iterative process, calculates the objective function value under the current threshold to evaluate the degree of match between ore grade and economic cost. According to the gradient descent algorithm, the system automatically adjusts the threshold and updates the iteration along the direction where the objective function decreases fastest. After each iteration, the system accurately calculates the change in the mineable grade threshold and compares it with the preset convergence condition. When the change is less than the preset convergence threshold, the system determines that the iterative process has converged and generates the final convergence grade threshold. This threshold represents the grade boundary at which the ore body blocks achieve optimal economic benefits under current economic and geological conditions, providing an accurate basis for subsequent economic feasibility verification, and helping mining companies achieve refined management and efficient utilization of resource development.

[0094] S233: Perform economic feasibility verification on the convergent grade threshold, screen the threshold results that meet the mining profit constraints, and generate a dynamic mineable grade threshold.

[0095] Specifically, the system constructs a mining profit model based on economic parameters such as market price, mining cost, and beneficiation cost, and calculates the expected mining profit for each ore body segment at a convergence grade threshold. The system sets mining profit constraints, such as minimum profit expectations, and compares the calculated expected profit against these constraints one by one. The system selects convergence grade thresholds that meet the mining profit constraints and integrates them to generate dynamic mineable grade thresholds. These thresholds not only reflect the geological characteristics of the ore body but also incorporate real-time economic factors to ensure their economic feasibility and effectiveness in actual mining. The resulting dynamic mineable grade thresholds will guide mining companies in rationally planning mining areas, optimizing resource allocation, improving the economic benefits of resource development, and achieving sustainable development in a complex and volatile market environment.

[0096] S24: Fit the spatial distribution data of the ore body based on the dynamic mineable grade threshold, use the Kriging spatial interpolation algorithm to generate the boundaries of the continuous mineable area, and output the mineable boundary data set under different price conditions.

[0097] Specifically, the system uses the Kriging spatial interpolation algorithm to fully leverage the spatial correlation and structural information of ore body segments to accurately estimate the mineability of ore in unsampled areas. During the fitting process, the system first analyzes the spatial distribution characteristics of the ore body segments to determine the continuity and changing trends of the ore bodies. Then, the Kriging interpolation algorithm is used to interpolate the unsampled areas to generate smooth boundaries of the mineable areas. These boundaries not only take into account the geological characteristics of the ore bodies but also incorporate the influence of economic parameters to ensure that the definition of the mineable areas is scientific and economically reasonable. Finally, the system outputs a dataset of mineable boundaries under different price conditions, which details the spatial extent and boundary morphology of the mineable areas. These datasets provide intuitive and detailed spatial data support for mining planning, helping mining companies to rationally plan mining areas and achieve the scientific development and efficient utilization of mineral resources.

[0098] The AI-based mineral exploration method provided above, through the synergy of real-time market data collection and dynamic optimization algorithms, can accurately extract and simulate price fluctuation trends and mining cost parameters, thereby constructing a multi-dimensional dynamic price change scenario dataset. A grade-cost joint optimization mechanism based on objective function iteration can achieve dynamic adjustment of the mineability threshold of ore body blocks, breaking through the limitations of traditional fixed grade thresholds in responding to market price fluctuations and mineral processing technology parameters. The Kriging spatial interpolation algorithm is used to continuously fit the optimization results to the boundaries, enabling the adaptive generation of mineable area boundaries under different economic scenarios, effectively addressing the problem of insufficient analysis of the spatial distribution of low-grade ores and associated resources by static models. This technical approach significantly improves the real-time and adaptability of resource mineability assessment by dynamically embedding market prices with coupled calculations of geological and economic parameters. It can accurately quantify resource development potential in complex market environments, provide dynamic data support for emergency resource scheduling and capacity planning, and overcome the technical bottlenecks of traditional mineral exploration methods in terms of threshold solidification, data fragmentation, and scenario adaptability.

[0099] In one embodiment, S3 of the artificial intelligence-based mineral exploration method provided by the present invention specifically includes the following steps:

[0100] S31: Feature extraction is performed on the ore body blocks in the mineable boundary data set. The grade gradient of each block is calculated using the sliding window method, and the grade difference value of the adjacent area is generated using the neighborhood analysis algorithm. Combined with the geological structure density statistical method, a multidimensional feature data set containing grade gradient, neighborhood grade difference and geological structure density is generated.

[0101] Specifically, the system uses a sliding window method to calculate the grade gradient of each block. By moving a fixed-size window in three-dimensional space, the grade data of the ore body blocks within the window are differentially calculated. The system automatically adjusts the window size according to the geological complexity and data density of the ore body to ensure the accuracy of the calculation results. Next, the system calls the neighborhood analysis algorithm, takes each ore body block as the center, analyzes the surrounding adjacent areas, calculates the mean and standard deviation of the grade difference between the adjacent areas and the central block, and generates a neighborhood grade difference value. The neighborhood analysis algorithm automatically identifies and excludes outliers to reduce the impact of data noise. At the same time, the system combines the geological structure density statistical method to analyze the geological structure characteristics of the area where the ore body blocks are located, and statistically calculates the density information of structures such as faults and folds. The geological structure density statistical method comprehensively considers the type, scale and distribution density of the structure to fully reflect the impact of the geological structure on the ore body. Ultimately, the system integrates grade gradients, neighborhood grade differences, and geological structure density data to generate a multidimensional feature data set, providing comprehensive feature input for subsequent machine learning model processing, ensuring that the feature data can accurately reflect the geological and economic properties of the ore body blocks, and improving the accuracy and reliability of resource assessment.

[0102] S32: Process the multidimensional feature data set based on the pre-trained machine learning model, and generate the probability value of the availability of each ore body block through feature standardization and model inference calculation.

[0103] Specifically, the system standardizes the input multi-dimensional feature data, scales the feature data such as grade gradient, neighborhood grade difference and geological structure density to a unified numerical range, eliminates the influence of the dimension and numerical range differences between different features, and ensures that the model can accurately assign weights to the features. At the same time, the system inputs the standardized feature data into the machine learning model, and the model performs complex nonlinear mapping calculations through the internal neural network structure. Based on the mapping relationship between the features and availability learned during the pre-training process, the model infers the availability of each ore body block and outputs an availability probability value. This probability value reflects the possibility of the ore body block being effectively utilized under the current economic and technical conditions, provides a quantitative basis for the subsequent allocation of availability labels, and helps mining companies accurately identify mineral resources with potential economic value and optimize resource development strategies. Preferably, the availability probability value of each ore body block is generated through the following steps:

[0104] S321: Process the multidimensional feature dataset based on the feature standardization method, eliminate feature dimension differences through the Z-score standardization algorithm, and generate a standard feature dataset.

[0105] Specifically, when processing feature data, the system uses the Z-Score normalization algorithm on the multidimensional feature data set. This process first involves the precise calculation of the mean and standard deviation of each feature dimension. The system traverses each eigenvalue in the multidimensional feature data set, accumulates and sums them, and calculates the mean μ. Then, it traverses the data set again and calculates the sum of the squares of the differences between each eigenvalue and the mean μ to obtain the standard deviation σ. Specifically, the calculation formulas for the mean μ and standard deviation σ are:

[0106]

[0107]

[0108] Where N is the number of samples, x i is the eigenvalue of the i-th sample; after obtaining the mean and standard deviation, the system applies a normalization formula to each eigenvalue, eliminating dimensional differences between different features. The resulting standardized feature dataset not only preserves the distribution characteristics of the original data but also makes different features comparable, providing high-quality data input for subsequent machine learning inference calculations. When performing these calculations, the system strictly adheres to mathematical definitions and algorithmic steps to ensure that each data point is accurately converted, thereby maintaining data integrity and accuracy.

[0109] S322: Perform inference calculations on the standard feature data set through a pre-trained machine learning model, calculate the feature weights and probability distribution of each block layer by layer based on the forward propagation algorithm, and generate the probability value of the availability of each ore body block.

[0110] Specifically, the pre-trained model is based on a deep neural network architecture and consists of multiple hidden layers, each composed of a large number of neurons. During inference, the system takes a standardized feature dataset as input and uses a forward propagation algorithm to calculate the feature weights and probability distributions for each ore block layer by layer. The input layer receives the normalized feature vector and passes it to the first hidden layer. In each hidden layer, the system performs a linear transformation on the input data—the product of the weight matrix and the input vector plus a bias term. It then performs a nonlinear transformation using nonlinear activation functions such as ReLU or Sigmoid to enhance the model's expressive power. This process is repeated sequentially through each hidden layer, with the output of each layer serving as the input for the next layer, gradually extracting high-level features of the data. Finally, in the output layer, the system calculates the exploitability probability of each ore block based on the output of the last layer. A softmax function is typically used to convert the output into a probability distribution. The generated probability reflects the exploitability of each ore block under current economic and technical conditions, providing a quantitative basis for subsequent resource classification. Throughout this process, the system strictly adheres to the pre-set model architecture and parameter settings, ensuring the accuracy and reliability of the inference results, thereby supporting effective mineral resource management decisions.

[0111] S33: Perform dynamic threshold determination processing on the availability probability value to generate an ore body block data set containing availability labels. The availability labels include classification identifiers of recoverable resources, associated resources, and temporarily unrecoverable resources.

[0112] Specifically, the system determines the classification thresholds for the probability of exploitation based on a preset dynamic threshold strategy, taking into account current market conditions and the company's economic goals. This dynamic threshold strategy automatically adjusts the thresholds based on factors such as market price fluctuations and changes in mining costs to ensure the economic rationality and timeliness of the classification results. For example, when market prices rise, the system appropriately lowers the threshold for recoverable resources, allowing more ore blocks to be classified as recoverable. Conversely, when market prices fall, the system raises the threshold, reducing the number of recoverable resources. Based on these thresholds, the system classifies the probability of exploitation for each ore block, generating a dataset of ore blocks containing availability labels.

[0113] The availability label clearly identifies whether an ore body block is a mineable resource, an associated resource, or a temporarily unmineable resource. The mineable resource label indicates that the ore body block has high economic value and mineability under current conditions; the associated resource label indicates that, while the ore body block cannot be mined alone, it has potential for comprehensive utilization during the mining of the main mineral species; and the temporarily unmineable resource label means that the ore body block is not mineable under current technical and economic conditions. The resulting dataset with availability labels provides intuitive and scientific guidance for resource planning and mining decisions for mining companies, helping them achieve rational resource allocation and efficient utilization, thereby improving their economic benefits and market competitiveness.

[0114] The present application provides an artificial intelligence-based mineral exploration method that can achieve refined classification of resource availability of ore blocks through the technical path of multi-dimensional feature extraction and dynamic threshold determination. The grade gradient and difference value calculation mechanism based on the sliding window method and neighborhood analysis can achieve quantitative characterization of the spatial distribution characteristics of the ore body, which is used to break through the problem of insufficient analysis of geological structure influencing factors by traditional methods; through the standardized processing and probabilistic inference of multi-dimensional feature data by pre-trained machine learning models, it can achieve dynamic prediction of the potential for utilization of ore blocks, and significantly improve the identification accuracy of low-grade ores and associated resources; combined with the dynamic threshold determination strategy of real-time economic parameter adjustment, it can achieve dynamic adaptation of resource classification identification and market environment, and effectively solve the hysteresis of fixed classification standards to external variables such as price fluctuations and mineral processing technology progress.

[0115] The mineral exploration method provided in this application significantly enhances the spatial resolution and economic adaptability of resource classification through the deep coupling of geological characteristics and machine learning models. It can achieve accurate portraits of the resource endowments of complex mining areas, provide a scientific basis for emergency resource development priority determination and sustainable mining planning, and break through the technical bottlenecks of traditional mineral exploration methods in terms of single feature dimensions, rigid classification standards and lack of dynamic adaptability.

[0116] In one embodiment, S4 of the artificial intelligence-based mineral exploration method provided by the present invention specifically includes the following steps:

[0117] S41: Perform emergency resource assessment on the ore body block data set, perform screening based on the availability labels of the ore body block data set, extract minable blocks and associated resource blocks, and generate an emergency resource candidate set.

[0118] Specifically, when the system performs emergency resource assessment on the ore body block data set, it performs screening operations based on the availability tags in the data set. The availability tags include classification identifiers for recoverable resources, associated resources, and temporarily unrecoverable resources. By setting label screening conditions, the system automatically identifies and extracts blocks identified as recoverable resources and associated resources to construct an emergency resource candidate set. This process involves efficient query and screening of large-scale data sets. The system uses optimized database query algorithms to ensure the rapid execution of screening operations. During the screening process, the system preliminarily sorts the extracted blocks according to preset priority rules so that subsequent economic evaluations can focus on the most promising resource blocks, thereby improving overall processing efficiency.

[0119] S42: Based on the price and cost parameters in the real-time market resource data, the economic evaluation of the emergency resource candidate set is carried out, and the recoverable resource volume and potential resource volume are calculated through the cost-benefit analysis model to generate the emergency resource volume including the economic evaluation indicators.

[0120] Specifically, real-time market resource data includes key economic indicators such as current market prices, mining costs, and transportation costs. The system inputs these parameters into the cost-benefit analysis model, which calculates the recoverable resources and potential resources of each resource block through mathematical formulas. Specifically, the system determines the sales revenue based on the ore grade and real-time market price, calculates the total cost based on the mining cost and beneficiation cost, and then derives economic evaluation indicators such as the net present value (NPV) and internal rate of return (IRR) for each block. These indicators intuitively reflect the economic feasibility and potential value of emergency resources. The system integrates the calculation results into an emergency resource data set containing economic evaluation indicators, providing a reliable basis for subsequent production capacity planning, helping enterprises make scientific and reasonable decisions in a complex and changing market environment, and ensuring economic benefits and resource utilization efficiency.

[0121] S43: Perform geospatial mapping of the emergency resource volume, associate the resource volume data with the ore body coordinates through GIS spatial analysis technology, and generate a mineable area coordinate map with the thermal value of the mineable area marked.

[0122] Specifically, the system precisely associates emergency resource data with ore body coordinates and, through spatial data processing algorithms, intuitively displays the distribution of emergency resources on a geographic coordinate map of the mining area. During this process, the system first spatially processes the emergency resource data, converting it into a spatial data format that matches the geographic coordinates. Then, using the visualization capabilities of GIS technology, it generates a thermal map based on the resource size and economic evaluation indicators. The color depth intuitively reflects the resource richness and economic benefits. The system ensures the accuracy and visualization of geospatial mapping. The generated coordinate map of the mineable area not only clearly marks the location and scope of the mineable area, but also visually presents the economic value distribution of the resources through thermal values. This provides intuitive and scientific spatial data support for subsequent capacity planning, helping enterprise managers and decision makers to clearly understand the geographic distribution and economic potential of emergency resources, thereby more efficiently formulating mining strategies and resource development plans.

[0123] S44: Based on the multi-objective optimization algorithm, the coordinate map of the mineable area is optimized for production capacity, balancing mining costs, market demand and resource sustainability parameters, and generating a production capacity planning scheme with priority sorting.

[0124] Specifically, the system comprehensively considers multiple key parameters, including mining costs, market demand, and resource sustainability. Using a multi-objective optimization algorithm, it establishes a mathematical model that transforms these parameters into quantifiable optimization objectives and constraints. During the optimization process, the system first analyzes market demand data to determine resource demand under different time periods and market scenarios. Then, combining mining cost data with resource sustainability parameters, it employs an iterative calculation method to identify a production capacity planning solution that minimizes mining costs and maximizes resource sustainability while still meeting market demand.

[0125] During the calculation process, the multi-objective optimization algorithm continuously adjusts and balances the weights between various objectives, ensuring that the resulting capacity planning solution is both economically feasible and meets market supply and demand, while also taking into account the long-term sustainable use of resources. Ultimately, the system generates a prioritized capacity planning solution that clearly specifies the mining sequence, capacity allocation, and mining schedule for each exploitable area. This capacity planning solution not only provides detailed guidance for a company's daily production operations but also helps it quickly adjust its capacity layout in emergency situations, achieving an optimal balance between economic benefits, market demand, and resource sustainability, thereby enhancing the company's overall competitiveness and sustainable development capabilities.

[0126] Preferably, if Figure 4 As shown, the present invention provides an artificial intelligence-based mineral exploration system 500, which is configured with the following modules:

[0127] Data acquisition module 510, used to obtain geological exploration data of the target mining area, economic parameters related to the economics of the mineral products, and mineral processing technical parameters related to the mineral processing technology;

[0128] The model building module 520 is used to perform data fusion processing on geological exploration data, economic parameters and mineral processing technical parameters to construct a three-dimensional ore body model including ore body block data set and economic mineral processing constraint information;

[0129] The mineability calculation module 530 is used to monitor the price fluctuation data and mining costs of the mineral market in real time, generate real-time market resource data, input the market resource data into the three-dimensional ore body model for processing, calculate the dynamic mineability of the ore body resources, and generate a mineable boundary data set under different price conditions;

[0130] The label assignment module 540 is used to perform resource classification processing on the mineable boundary dataset, predict the availability of ore body blocks and assign labels using a machine learning model, and generate an ore body block dataset containing availability labels. The ore body block dataset is used to indicate the spatial distribution of mineable resources, associated resources, and temporarily unmineable resources.

[0131] The resource assessment and planning module 550 is used to perform emergency resource assessment processing on the ore body block data set, screen the mineable blocks according to the availability tags, calculate the emergency resource volume in combination with real-time market resource data, and generate a mineable area coordinate map and production capacity planning plan.

[0132] In summary, the artificial intelligence-based mineral exploration system provided by the present invention can realize the efficient integration of geological exploration data, economic parameters and mineral processing process parameters and the accurate construction of three-dimensional ore body models through the synergistic effect of multi-source data fusion and dynamic optimization algorithm, so as to break through the response hysteresis of traditional static evaluation models to market price fluctuations and mining cost changes; the dynamic mineability calculation mechanism based on real-time market resource data can achieve real-time optimization and multi-scenario adaptation of the mineable boundary of ore body resources, and effectively solve the problem of idle low-grade mines and associated resources caused by fixed grade thresholds; through the intelligent prediction and label classification of the availability of ore body blocks by machine learning models, it can realize the refined analysis of resource spatial distribution characteristics, and provide multi-dimensional data support for emergency resource assessment; the emergency resource potential assessment system combined with real-time market parameters and label data can achieve dynamic matching of mining plans and market fluctuations, so as to realize rapid response and scientific decision-making in the scenario of sudden resource supply interruption.

[0133] Preferably, the model building module 520 provided in this application is configured with units:

[0134] The ore body morphology data generation unit is used to perform three-dimensional spatial reconstruction processing on the drill hole location data, ore composition data and ore layer thickness data in the geological exploration data based on the spatial interpolation algorithm, calculate the ore composition distribution of the unsampled area through the Kriging interpolation algorithm, and generate the three-dimensional ore body morphology data;

[0135] The ore body block data integration unit is used to perform regular grid division processing on the ore body's three-dimensional morphological data, divide the ore body into cubic units of equal volume according to preset sizes, and generate an ore body block data set containing location coordinates, ore grade, and reserve information;

[0136] The ore body data fusion unit is used to perform multi-dimensional data fusion processing on the ore body block data set based on the market price history data in the economic parameters and the recovery efficiency data in the mineral processing technical parameters. It integrates the economic constraints and geological parameters through the matrix weighted algorithm to construct a three-dimensional ore body model that includes the ore body block data set and economic mineral processing constraint information.

[0137] Preferably, the admissibility calculation module 530 provided in this application is configured with units:

[0138] The market monitoring unit is used to monitor the price fluctuation data and mining costs of the mineral market in real time based on the market data collection system. It extracts price trend characteristics and cost fluctuation parameters through time series analysis algorithms to generate real-time market resource data.

[0139] A simulation data generation unit is used to input real-time market resource data into the three-dimensional ore body model to simulate price fluctuations, and combine the cost threshold and ore recovery rate parameters of the economic ore processing constraint information in the three-dimensional ore body model to generate a simulation data set containing multiple sets of dynamic price change scenarios;

[0140] A dynamic threshold generation unit is used to perform mineability calculations on the simulated data set and the ore body block data set in the three-dimensional ore body model based on a dynamic optimization algorithm, and to iteratively solve the grade-cost optimal solution of the ore body block through the objective function to generate a dynamic mineable grade threshold;

[0141] The mining boundary integration unit is used to fit the spatial distribution data of the ore body based on the dynamic mining grade threshold, generate the continuous mining area boundary using the Kriging spatial interpolation algorithm, and output the mining boundary data set under different price conditions.

[0142] Preferably, the dynamic threshold generation unit includes an objective function decomposition subunit, an iterative optimization solution subunit, and an economic feasibility verification subunit. The objective function decomposition subunit is used to perform objective function decomposition processing on the simulation data set and the ore body block data set based on the dynamic optimization algorithm, splitting the objective function of the mineability calculation into ore grade constraint terms and economic cost constraint terms; the iterative optimization solution subunit is used to iteratively optimize the ore grade constraint terms and the economic cost constraint terms, solving the grade-cost joint optimal solution through the gradient descent algorithm until the change in the mineable grade threshold meets the preset convergence condition, thereby generating a converged grade threshold; the economic feasibility verification subunit is used to perform economic feasibility verification processing on the converged grade threshold, screening the threshold results that meet the mining profit constraint, and generating a dynamic mineable grade threshold.

[0143] Preferably, the label allocation module 540 provided in this application is configured with units:

[0144] A multidimensional feature data set generation unit is used to perform feature extraction processing on the ore body blocks in the mineable boundary data set, calculate the grade gradient of each block through the sliding window method, generate the grade difference value of the adjacent area through the neighborhood analysis algorithm, and combine the geological structure density statistical method to generate a multidimensional feature data set containing grade gradient, neighborhood grade difference and geological structure density;

[0145] An availability probability value generation unit is used to process the multidimensional feature data set based on a pre-trained machine learning model, and generate the availability probability value of each ore body block through feature standardization and model inference calculation;

[0146] The ore body block data set generation unit is used to perform dynamic threshold judgment processing on the availability probability value and generate an ore body block data set containing availability tags. The availability tags include classification identifiers of recoverable resources, associated resources and temporarily unrecoverable resources.

[0147] Preferably, the availability probability value generation unit includes a feature standardization subunit and a probability value inference subunit. The feature standardization subunit is used to process the multidimensional feature data set based on the feature standardization method, eliminate feature dimension differences through the Z-score standardization algorithm, and generate a standard feature data set; the probability value inference subunit is used to infer and calculate the standard feature data set through a pre-trained machine learning model, calculate the feature weight and probability distribution of each block layer by layer based on the forward propagation algorithm, and generate the availability probability value of each ore body block.

[0148] Preferably, the resource assessment and planning module 550 provided in this application is configured with units:

[0149] The emergency resource candidate integration unit is used to perform emergency resource assessment on the ore body block data set, screen the ore body block data set based on its availability tags, extract the mineable blocks and associated resource blocks, and generate an emergency resource candidate set;

[0150] The emergency resource calculation unit is used to perform economic evaluation on the emergency resource candidate set based on the price and cost parameters in the real-time market resource data, calculate the recoverable resource volume and potential resource volume through the cost-benefit analysis model, and generate the emergency resource volume including economic evaluation indicators;

[0151] The mineable area generation unit is used to perform geospatial mapping of emergency resource quantities, associate resource quantity data with ore body coordinates through GIS spatial analysis technology, and generate a mineable area coordinate map marked with the thermal value of the mineable area;

[0152] The capacity planning scheme generation unit is used to optimize the capacity of the mineable area coordinate map based on a multi-objective optimization algorithm, balance mining costs, market demand and resource sustainability parameters, and generate a capacity planning scheme including priority sorting.

[0153] In one embodiment, the present application also provides a computer device including a memory and a processor, wherein the memory stores a computer program, and the processor implements the above-mentioned artificial intelligence-based mineral exploration method when executing the computer program.

[0154] In one embodiment, the present application also provides a computer-readable storage medium having a computer program stored thereon, which implements the above-mentioned artificial intelligence-based mineral exploration method when executed by a processor.

[0155] In the description of this specification, the reference terms "one embodiment," "some embodiments," "example," "specific example," or "some examples" mean that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. Moreover, the specific features, structures, materials, or characteristics described may be combined in any appropriate manner in any one or more embodiments or examples. In addition, those skilled in the art may combine and integrate different embodiments or examples described in this specification, as well as features of different embodiments or examples, unless they are mutually inconsistent.

[0156] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to the partial description of the method embodiments. The device embodiments described above are merely illustrative, wherein the components described as separate parts may or may not be physically separated, and the parts displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the disclosed solution. A person of ordinary skill in the art can understand and implement it without expending creative work.

[0157] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any person skilled in the art can easily conceive of various modifications or substitutions within the technical scope disclosed in this application, and such modifications or substitutions should be included within the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A mineral exploration method based on artificial intelligence, characterized in that: The following steps are involved: S1: obtaining geological exploration data of the target mining area, economic parameters related to the economic efficiency of the mineral products, and mineral processing technical parameters related to the mineral processing process, and fusing the geological exploration data, the economic parameters, and the mineral processing technical parameters to construct a three-dimensional ore body model including an ore body block data set and economic mineral processing constraint information; S2: Real-time monitoring of price fluctuation data and mining costs in the mineral market to generate real-time market resource data, inputting the market resource data into the three-dimensional ore body model for processing, calculating the dynamic mineability of the ore body resources, and generating a mineable boundary data set under different price conditions; S3: performing resource classification processing on the mineable boundary dataset, predicting the availability of ore body blocks and assigning labels using a machine learning model, and generating an ore body block dataset containing availability labels, wherein the ore body block dataset is used to indicate the spatial distribution of mineable resources, associated resources, and temporarily unmineable resources; S4: Perform emergency resource assessment on the ore body block data set, screen the mineable blocks according to the availability tags, calculate the emergency resource volume in combination with the real-time market resource data, and generate a mineable area coordinate map and production capacity planning scheme.

2. The method according to claim 1, characterized in that Said S1 comprises: S11: Based on the sensor network and market monitoring system, the exploration data, market prices and mineral processing process parameters of the target mining area are collected in real time to generate geological exploration data of the target mining area, economic parameters related to the economic efficiency of the mineral resources and mineral processing technical parameters related to the mineral processing process; S12: performing three-dimensional spatial reconstruction processing on the drill hole location data, ore composition data, and ore layer thickness data in the geological exploration data based on a spatial interpolation algorithm, calculating the ore composition distribution of the unsampled area using a Kriging interpolation algorithm, and generating three-dimensional ore body morphology data; S13: performing regular grid division processing on the three-dimensional morphological data of the ore body, dividing the ore body into cubic units of equal volume according to a preset size, and generating an ore body block data set containing position coordinates, ore grade, and reserve information; S14: Based on the market price history data in the economic parameters and the recovery efficiency data in the mineral processing technical parameters, multi-dimensional data fusion processing is performed on the ore body block data set, and economic constraints and geological parameters are integrated through a matrix weighted algorithm to construct the three-dimensional ore body model containing the ore body block data set and economic mineral processing constraint information.

3. The method according to claim 1, characterized in that The S2 includes: S21: Based on the market data collection system, the price fluctuation data and mining costs of the mineral market are monitored in real time. The price trend characteristics and cost fluctuation parameters are extracted through the time series analysis algorithm to generate real-time market resource data. S22: Inputting the real-time market resource data into the three-dimensional ore body model to perform price fluctuation simulation processing, and combining the cost threshold and ore dressing recovery rate parameters of the economic ore dressing constraint information in the three-dimensional ore body model to generate a simulation data set containing multiple sets of dynamic price change scenarios; S23: performing mineability calculation processing on the simulation data set and the ore body block data set in the three-dimensional ore body model based on a dynamic optimization algorithm, solving the grade-cost optimal solution of the ore body block through objective function iteration, and generating a dynamic mineable grade threshold; S24: fitting the spatial distribution data of the ore body based on the dynamic mineable grade threshold, generating continuous mineable area boundaries using the Kriging spatial interpolation algorithm, and outputting a mineable boundary data set under different price conditions.

4. The method according to claim 3, characterized in that The S23 includes: S231: performing objective function decomposition processing on the simulation data set and the ore body block data set in the three-dimensional ore body model based on a dynamic optimization algorithm, and splitting the objective function of mineability calculation into an ore grade constraint term and an economic cost constraint term; S232: performing iterative optimization processing on the ore grade constraint and the economic cost constraint, solving the grade-cost joint optimal solution by a gradient descent algorithm, until the change in the mineable grade threshold satisfies a preset convergence condition, thereby generating a converged grade threshold; S233: Performing economic feasibility verification processing on the convergent grade threshold, screening threshold results that meet the mining profit constraint, and generating a dynamic mineable grade threshold.

5. The method according to claim 1, wherein The S3 includes: S31: performing feature extraction processing on the ore body blocks in the mineable boundary data set, calculating the grade gradient of each block by a sliding window method, generating the grade difference value of the adjacent area by a neighborhood analysis algorithm, and combining the geological structure density statistical method to generate a multidimensional feature data set including the grade gradient, neighborhood grade difference and geological structure density; S32: Processing the multidimensional feature data set based on a pre-trained machine learning model, and generating an availability probability value for each ore body block through feature standardization and model inference calculation; S33: Performing dynamic threshold determination processing on the availability probability value to generate an ore body block data set containing availability tags, wherein the availability tags include classification identifiers of recoverable resources, associated resources, and temporarily unrecoverable resources.

6. The method according to claim 5, characterized in that The S32 includes: S321: Processing the multidimensional feature dataset based on a feature standardization method, eliminating feature dimension differences through a Z-score standardization algorithm, and generating a standard feature dataset; S322: Inferring and calculating the standard feature data set through a pre-trained machine learning model, calculating the feature weight and probability distribution of each block layer by layer based on the forward propagation algorithm, and generating the probability value of the availability of each ore body block.

7. The method according to any one of claims 1 to 6, characterized in that The S4 includes: S41: performing emergency resource assessment processing on the ore body block data set, screening the ore body block data set based on its availability tags, extracting mineable blocks and associated resource blocks, and generating an emergency resource candidate set; S42: Based on the price and cost parameters in the real-time market resource data, an economic evaluation is performed on the candidate set of emergency resources, and recoverable resources and potential resources are calculated using a cost-benefit analysis model to generate an emergency resource volume including economic evaluation indicators; S43: Performing geospatial mapping on the emergency resource volume, associating the resource volume data with the ore body coordinates using GIS spatial analysis technology, and generating a mineable area coordinate map with the thermal value of the mineable area marked; S44: Based on a multi-objective optimization algorithm, the coordinate map of the mineable area is subjected to capacity optimization processing, mining costs, market demand and resource sustainability parameters are balanced, and a capacity planning scheme including priority sorting is generated.

8. A mineral exploration system based on artificial intelligence, characterized in that: The system comprises: Data acquisition module, used to obtain geological exploration data of the target mining area, economic parameters related to the economics of the mineral products, and mineral processing technical parameters related to the mineral processing technology; A model building module is used to perform data fusion processing on the geological exploration data, the economic parameters and the mineral processing technical parameters to build a three-dimensional ore body model including an ore body block data set and economic mineral processing constraint information; a mineability calculation module for monitoring price fluctuations and mining costs in the mineral market in real time, generating real-time market resource data, inputting the market resource data into the three-dimensional ore body model for processing, calculating the dynamic mineability of the ore body resources, and generating a mineable boundary data set under different price conditions; a label assignment module for performing resource classification processing on the mineable boundary dataset, predicting the availability of ore body blocks and assigning labels using a machine learning model, and generating an ore body block dataset containing availability labels, wherein the ore body block dataset is used to indicate the spatial distribution of mineable resources, associated resources, and temporarily unmineable resources; The resource assessment and planning module is used to perform emergency resource assessment processing on the ore body block data set, screen the mineable blocks according to the availability tags, calculate the emergency resource volume in combination with the real-time market resource data, and generate a mineable area coordinate map and production capacity planning scheme.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images