A three-dimensional quantitative prediction method and system for concealed mineral deposits

Abstract

This invention provides a three-dimensional quantitative prediction method and system for concealed mineral deposits, relating to the field of geological exploration technology. It involves collecting an equal number of mineralized and non-mineralized model samples and dividing them into three-dimensional grid cells. Feature values ​​such as stratigraphic systems, rock types, and structural characteristics, along with the presence or absence of mineral deposits, are obtained as target variables. Stratified sampling is used to divide the dataset, construct a random forest model, and set and optimize parameters. By querying geological logs, geological feature data such as stratigraphic systems, rock types, and structural characteristics of the area to be detected are extracted, compiled into a geological feature data table, and imported into professional geological modeling software to construct an accurate three-dimensional geological model. This provides a reliable geological basis for concealed mineral deposit prediction. The model is then further divided into grids, and the required feature data is input to predict concealed mineral deposits, achieving precise three-dimensional spatial positioning and prediction of concealed mineral deposits. This provides a scientific basis for mineral exploration and development and improves the accuracy of resource assessment.

Description

Technical Field

[0001] This invention relates to the field of geological exploration technology, specifically to a three-dimensional quantitative prediction method and system for concealed mineral deposits. Background Technology

[0002] With the ever-increasing demand for mineral resources, accurate detection and quantitative prediction of concealed ore bodies have become key tasks in the field of geological exploration. Traditional geological exploration methods often have limitations when faced with complex geological conditions and concealed ore bodies.

[0003] In existing technologies, CN118070971A discloses a three-dimensional quantitative prediction method, computer equipment, medium, and product for concealed mineralization. This method primarily acquires geological data of the target mining area through geological maps, geophysical data, and geochemical data. Based on the theory of metallogenic systems and the three-dimensional exploration area prospecting prediction theory, it analyzes the metallogenic process to determine ore-controlling elements and uses machine learning algorithms to determine the probability of ore bodies existing at each spatial location within the target mining area. However, its data types are relatively traditional and singular. In complex concealed environments, relying solely on this data is insufficient to accurately capture the unique ore-controlling elements of a specific mining area. For example, key information such as special geological structures, changes in rock microstructure, and mineralization alteration characteristics may not be fully reflected, thus affecting the accurate analysis of the metallogenic process and the prediction of ore body location.

[0004] Meanwhile, existing prediction methods lack visualization capabilities, failing to intuitively display the three-dimensional structure of concealed ore within geological bodies. This makes it difficult for geologists to understand and analyze prediction results, hindering in-depth exploration of the relationship between geological bodies and concealed ore from a three-dimensional spatial perspective, and impeding the development of scientifically sound exploration and mining plans.

[0005] Furthermore, with the continuous increase in the amount of geological data, traditional data processing and analysis methods are facing challenges in terms of efficiency and accuracy. In the era of big data, how to more efficiently integrate and analyze massive amounts of geological data and uncover the hidden mineralization information within them has become an urgent problem to be solved.

[0006] Therefore, the present invention aims to provide a more advanced and comprehensive three-dimensional quantitative prediction solution for concealed mineral deposits.

[0007] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0008] The purpose of this invention is to provide a three-dimensional quantitative prediction method and system for concealed mineral deposits, so as to solve the problems mentioned in the background art.

[0009] To achieve the above objectives, the present invention provides the following technical solution:

[0010] A three-dimensional quantitative prediction method for concealed mineral deposits, comprising the following steps:

[0011] Step 1: Collect an equal number of mineralized and non-mineralized model samples, divide all model samples into grid cells, and use the stratigraphic thickness, strike, dip, and dip angle of the stratigraphic system of each grid cell as stratigraphic characteristic values, classify and encode the rock type as rock characteristic values, and use the fault strike, dip, dip angle, and fault displacement of the structural features as structural characteristic values.

[0012] Step 2: Take the presence or absence of minerals as the target variable, associate the target variable with its corresponding stratigraphic, rock, and structural feature values ​​to form a set of geological feature data, label the set of all geological feature data as model data, use the model data as the training set, construct a random forest model, set the maximum depth, the minimum number of samples required for node splitting, and the minimum number of samples for leaf nodes, and then use the accuracy to evaluate and optimize the random forest model.

[0013] Step 3: Extract geological feature data of the area to be tested by querying geological logs, save it as a geological feature data table, and import it into geological modeling software to build a three-dimensional geological model. The geological feature data includes stratigraphic system, rock type, and structural feature data.

[0014] Step 4: Divide the constructed 3D geological model into cubic meshes. Based on the geological feature data table, input the geological feature data corresponding to each mesh unit into the optimized random forest model. The optimized random forest model will then predict the mesh units that may contain concealed minerals, thereby realizing the location prediction of concealed minerals in 3D space.

[0015] Furthermore, the collected mineralized and non-mineralized model samples are first divided into grids, and then the formation thickness at the j-th measurement point within the i-th grid cell is defined as h. ij The total number of measurement points in the grid cell is n, i is used to distinguish different grid cells, each grid cell has its own independent set of measurement point data and corresponding formation thickness characteristic value calculation process, and j is a variable for numbering each measurement point in the i-th grid cell;

[0016] The formula for calculating the characteristic value of the formation thickness is:

[0017]

[0018] Let the strike angle of the stratigraphy at the j-th measurement point within the i-th grid cell be denoted as α. ij First, set the angle value α ijConverted to radians for calculation, the converted radian value is:

[0019]

[0020] In the formula, θ ij Let x be the strike angle in radians of the formation at the j-th measurement point within the i-th grid cell, and then calculate x. ij =cosθ ij and y ij =sinθ ij ,

[0021] The formula for calculating the characteristic value of the stratigraphic strike is:

[0022]

[0023] In the formula, θ i The eigenvalue representing the stratigraphic strike of the i-th grid cell in radians.

[0024] Then take the radian value θ i Converted into angular values ​​as characteristic values ​​of stratigraphic strike:

[0025]

[0026] In the formula, α i The characteristic value representing the stratigraphic strike of the i-th grid cell is denoted as .

[0027] Let the dip angle of the stratigraphy at the j-th measurement point within the i-th grid cell be denoted as β. ij First, convert it to radians:

[0028]

[0029] In the formula, Represented as the stratigraphic dip angle in radians at the j-th measurement point within the i-th grid cell, then calculated... and

[0030] The formula for calculating the characteristic value of the stratigraphic dip is:

[0031]

[0032] In the formula, It is represented as the characteristic value of the stratigraphic dip of the i-th grid cell under the radian value;

[0033] Then convert it into an angle value as a characteristic value of the stratigraphic dip:

[0034]

[0035] Let the dip angle of the formation at the j-th measurement point within the i-th grid cell be denoted as γ. ij The formula for calculating the characteristic value of the dip angle of the strata is:

[0036]

[0037] In the formula, γ i The eigenvalue representing the dip angle of the formation in the i-th grid cell is denoted as .

[0038] Identify all possible rock types and assign a unique code C to each. k Where k = 1, 2, 3, ..., f, f represents the total number of all possible different rock types within the i-th grid cell, forming a rock type coding list. The distribution area of ​​the k-th rock type within the i-th grid cell is set to A. ik The total area is S i The formula for the area proportion of the k-th rock type within this grid cell is expressed as:

[0039]

[0040] In the formula, p ik Let A represent the area percentage of the k-th rock type within the i-th grid cell. ik S represents the distribution area of ​​the k-th rock type within the i-th grid cell. i Represented as the total area of ​​the i-th grid cell,

[0041] Within each grid cell, the area percentage p of different rock types is compared. ik Extract the rock type with the highest proportion and use the code of that rock type as the rock type code feature value of this grid cell.

[0042] Furthermore, using the presence or absence of mineralization as a stratification variable, with a total number of mineralized and non-mineralized samples of e, a random number generator is used to generate random numbers within the range of 1 to e for all mineralized samples. Use a unique random number generator to add the corresponding numbered mineralized samples to the training set. For non-mineralized samples, use a random number generator to generate a random number within the range of 1 to e. Use a unique random number to add all the corresponding unmineralized samples to the training set. When the model accurately predicts the proportion of mineralized and unmineralized samples in the training set, i.e., the accuracy rate, is higher than 90%, training stops.

[0043] Furthermore, the maximum depth is initially set to 8, and a random forest model is constructed using the pre-divided training set. The minimum number of samples required for node splitting is set to 8, and the minimum number of samples for leaf nodes is set to 5.

[0044] Furthermore, in each iteration of the random forest model training, the sample data from the training set is input into the current random forest model for prediction. Let TP be the number of correctly predicted mineralized samples and TN be the number of correctly predicted non-mineralized samples in the training set, and let e be the total number of mineralized and non-mineralized samples. After each training iteration, the accuracy is calculated according to the formula:

[0045]

[0046] In the formula, Accuracy represents the accuracy rate, e represents the total number of mineralized and non-mineralized samples, TP represents the number of correct predictions for mineralized samples, and TN represents the number of correct predictions for non-mineralized samples.

[0047] Training should stop when the accuracy rate is above 90%.

[0048] Furthermore, the names, thicknesses, strikes, dips, and dip angles of each stratum in the stratigraphic system data are saved as a geological feature data table in a unified text format. This data is then imported into geological modeling software, where different rock types are distinguished by different colors or textures. This allows the model to visually reflect the distribution characteristics of rock types. Structural feature data is integrated into the model to show the location, shape, and intersection relationship of faults with strata in three-dimensional space, thus constructing a complete three-dimensional geological model that includes stratigraphic structure, rock types, and structural features.

[0049] Furthermore, feature data is extracted from the constructed 3D model of the area to be detected in the same way as when training the model. This includes stratum thickness, strike, dip, and dip angle; rock type encoding; and fault strike, dip, dip angle, and displacement. The 3D space of the area to be detected is divided into cubic meshes, with each mesh cell considered as an independent sample point. All sample data are input into the trained random forest model. Based on the input feature data, and utilizing the relationship between geological features and mineralization learned during training, a probability threshold μ is set after the model's prediction results are output. When the probability value of the predicted result of a certain mesh cell corresponding to the presence of concealed minerals is higher than the probability threshold μ, it is determined that there are concealed minerals at the location of this mesh cell; otherwise, if the probability value does not reach the probability threshold μ, it is determined that there are no concealed minerals at that location.

[0050] Additionally, a three-dimensional quantitative prediction system for concealed mineral deposits is provided. This system is used to execute the aforementioned three-dimensional quantitative prediction method for concealed mineral deposits, including:

[0051] The feature extraction module is used to collect an equal number of mineralized and non-mineralized model samples, and divide them into three-dimensional grid cells. The stratigraphic thickness, strike, dip, and dip angle of each grid cell are used as feature values, the rock type is classified and coded as feature values, the fault strike, dip, dip angle, and fault displacement of structural features are used as feature values, and the presence or absence of minerals is used as the target variable.

[0052] The feature extraction module is used to collect an equal number of mineralized and non-mineralized model samples, divide all model samples into grid cells, and use the stratigraphic thickness, strike, dip, and dip angle of the stratigraphic structure of each grid cell as stratigraphic feature values, classify and encode rock types as rock feature values, and use the fault strike, dip, dip angle, and fault displacement of structural features as structural feature values.

[0053] The model building module is used to take the presence or absence of minerals as the target variable, associate the target variable with its corresponding stratigraphic, rock, and structural feature values ​​to form a set of geological feature data, label the set of all geological feature data as model data, use the model data as the training set, build a random forest model, and after setting the maximum depth, the minimum number of samples required for node splitting, and the minimum number of samples for leaf nodes, the accuracy is used to evaluate and optimize the random forest model.

[0054] The 3D modeling module is used to extract geological feature data of the area to be detected by querying geological logs, save it as a geological feature data table, and import it into geological modeling software to build a 3D geological model. The geological feature data includes stratigraphic structure, rock type, and structural feature data.

[0055] The concealed ore location module is used to divide the constructed 3D geological model into cubic meshes. Based on the geological feature data table, the geological feature data corresponding to each mesh unit is input into the optimized random forest model. The optimized random forest model predicts the mesh units that may contain concealed ore, thereby realizing the location prediction of concealed ore in 3D space.

[0056] Compared with the prior art, the beneficial effects of the present invention are:

[0057] By collecting an equal number of mineralized and non-mineralized model samples and dividing them into three-dimensional grid cells, including multiple feature values ​​such as stratigraphic structure, rock type and tectonic features, as well as the target variable of whether or not mineralization exists, the random forest model avoids over-learning of one type of sample and under-learning of another type of sample due to the imbalance of sample quantity, thereby improving the model's ability to identify and generalize different mineralization conditions.

[0058] After dividing the mineralized and non-mineralized model samples into three-dimensional grid units, the variation of the stratigraphic strike in different small regions can be clearly shown. For each grid unit, multiple feature values ​​of stratigraphic structure, rock type and tectonic features can be obtained in detail. The various geological features within each grid unit are interconnected and influence each other, working together to affect the mineralization process, providing rich information for revealing the mineralization mechanism. After multiple experiments and parameter adjustments, the accuracy of the model can reach a high level, effectively improving the accuracy of predicting concealed mineralization. Attached Figure Description

[0059] Figure 1 This is a schematic diagram of the overall method flow of the present invention.

[0060] Figure 2 This is a schematic diagram of the overall system modules of the present invention. Detailed Implementation

[0061] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0062] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0063] Example:

[0064] Please see Figure 1 The present invention provides a technical solution:

[0065] A three-dimensional quantitative prediction method for concealed mineral deposits, comprising the following steps:

[0066] Step 1: Collect an equal number of mineralized and non-mineralized model samples, divide all model samples into grid cells, and use the stratigraphic thickness, strike, dip, and dip angle of the stratigraphic system of each grid cell as stratigraphic characteristic values, classify and encode the rock type as rock characteristic values, and use the fault strike, dip, dip angle, and fault displacement of the structural features as structural characteristic values.

[0067] Based on geological exploration reports from multiple explored mining areas, areas where large quantities of high-grade ore have been mined were identified as mineralized areas, while areas where no economically valuable ore bodies were found after detailed exploration were identified as non-mineralized areas. For each area, geological characteristic data were collected, including the thickness, strike, dip, and dip angle of strata in the stratigraphic system, rock type, and the strike, dip, dip angle, and displacement of faults in the structural features. Ultimately, 500 mineralized and 500 non-mineralized model samples were collected. Each sample contains complete stratigraphic structure, rock type, structural feature data, and a label indicating whether mineralization exists.

[0068] Sample data from mineralized and non-mineralized areas represent different geological outcomes. The combinations of geological features they contain are key to the model's ability to distinguish between mineralized and non-mineralized conditions. During training, the model learns from a large number of samples, enabling it to better identify subtle differences in geological features between mineralized and non-mineralized areas. This allows for more accurate predictions of the presence or absence of concealed mineralization when faced with data from unknown regions. For example, when the model learns that a certain rock type, combined with specific stratum thickness and fault structures, frequently occurs in mineralized areas, it can use these characteristic patterns to predict new areas, reducing the possibility of misjudgments.

[0069] First, the collected mineralized and non-mineralized model samples were divided into grids, with a grid size of 5 meters. This ensured that each grid cell could become a relatively independent and representative geological sample cell, facilitating subsequent data collection and analysis of various geological elements such as strata, structures, and rock types within each small area. This allowed for a more accurate understanding of the relationship between geological features and mineralization.

[0070] Using the volume calculation function of the modeling software Surpac, the volume of the ore portion and the volume of the entire mesh were calculated separately, and then the volume percentage of the ore-bearing portion was calculated:

[0071]

[0072] In the formula, P represents the volume percentage of the mineralized portion, and V... ore This represents the volume of the ore-bearing portion within the grid. During model training and validation using modeling software and related algorithms, using 60% as the decision threshold enables the model to achieve a good balance between accuracy and recall when predicting concealed mineralization. Therefore, when the ore content P in the grid is greater than or equal to 60%, it is determined to be an ore-bearing area; conversely, when the ore content P in the grid is less than 60%, it is determined to be an unore-bearing area.

[0073] Let h be the formation thickness at the j-th measurement point within the i-th grid cell. ijThe total number of measurement points in the grid cell is n, i is used to distinguish different grid cells, each grid cell has its own independent set of measurement point data and corresponding formation thickness characteristic value calculation process, and j is a variable for numbering each measurement point in the i-th grid cell;

[0074] The formula for calculating the characteristic value of the formation thickness is:

[0075]

[0076] In the formula, H i It is a characteristic value of the formation thickness. The formation thickness may not be uniform within a grid cell. Multiple measurement points can more comprehensively reflect its variation and obtain a value that can represent the overall formation thickness level within the grid cell.

[0077] Let the strike angle of the stratigraphy at the j-th measurement point within the i-th grid cell be denoted as α. ij First, set the angle value α ij Converted to radians for calculation, the converted radian value is:

[0078]

[0079] In the formula, θ ij Let x be the strike angle in radians of the formation at the j-th measurement point within the i-th grid cell, and then calculate x. ij =cosθ ij and y ij =sinθ ij The strike of strata refers to the direction of the intersection of the bedding plane and the horizontal plane; it reflects the horizontal extension trend of the strata. First, let's define the angle value α. ij Convert to radian value θ ij This is for subsequent trigonometric function calculations and for calculating the characteristic value θ of the stratigraphic strike using the arctangent function. i It is more convenient and accurate to calculate in radians because the rules for calculating trigonometric functions in radians are simpler and more consistent.

[0080] The formula for calculating the characteristic value of the stratigraphic strike is:

[0081]

[0082] In the formula, θ i It represents the characteristic value of the stratigraphic strike of the i-th grid cell under the radian value.

[0083] Then take the radian value θ i Converted into angular values ​​as characteristic values ​​of stratigraphic strike:

[0084]

[0085] In the formula, α i The characteristic value of the stratigraphic strike of the i-th grid cell is represented as the eigenvalue, and then converted into an angle value α. i To facilitate intuitive understanding.

[0086] Let the dip angle of the stratigraphy at the j-th measurement point within the i-th grid cell be denoted as β. ij First, convert it to radians:

[0087]

[0088] In the formula, Represented as the stratigraphic dip angle in radians at the j-th measurement point within the i-th grid cell, then calculated... and Stratigraphic dip refers to the direction in which strata dip, and it is perpendicular to the stratigraphic strike. Together, they describe the spatial tilt of the strata. The dip angle β is used to define the strata dip angle. ij Convert to radians This is also to facilitate subsequent trigonometric function calculations.

[0089] The formula for calculating the characteristic value of the stratigraphic dip is:

[0090]

[0091] In the formula, This is represented by the characteristic value of the stratigraphic dip of the i-th grid cell under the radian value.

[0092] Then convert it into an angle value as a characteristic value of the stratigraphic dip:

[0093]

[0094] Let the dip angle of the formation at the j-th measurement point within the i-th grid cell be denoted as γ. ij The formula for calculating the characteristic value of the dip angle of the strata is:

[0095]

[0096] In the formula, γ i Let γ be the characteristic value of the dip angle of the strata in the i-th grid cell. The dip angle is the acute angle between the bedding plane of the strata and the horizontal plane, which directly reflects the degree of inclination of the strata in the vertical direction. The dip angle γ of the strata at each measurement point within the grid cell is then used to determine the dip angle. ij The characteristic value γ of the formation dip angle is obtained by arithmetic mean calculation. i This value represents the overall vertical tilt of the strata within the grid cell.

[0097] For constructing feature data, detect whether there are faults within the grid cells. If they exist, calculate the characteristic values ​​of the fault strike, dip, dip angle, and fault displacement.

[0098] Let the fault strike angle at the j-th measurement point within the i-th grid cell be denoted as α. ′ ij First, set the angle value α ij Converted to radians for calculation, the converted radian value is:

[0099]

[0100] In the formula, θ ′ ij Let x be the fault strike angle in radians at the j-th measurement point within the i-th grid cell, then calculate x. ′ ij =cosθ ij and y ′ ij =sinθ ij Fault strike refers to the direction in which a fault extends on the ground or underground horizontal plane; it reflects the fault's tendency to extend horizontally. Similar to the calculation of stratigraphic strike, the angle value α is first... ′ ij Convert to radian value θ ′ ij This is for subsequent trigonometric function calculations and for calculating the characteristic value θ of the fault strike using the arctangent function. ′ i It is more convenient and accurate.

[0101] The formula for calculating the characteristic value of fault strike is:

[0102]

[0103] In the formula, θ ′ i It represents the characteristic value of the fault strike of the i-th grid cell under the radian value.

[0104] Then take the radian value θ ′ i Convert the fault angle value into a characteristic value representing the fault strike:

[0105]

[0106] In the formula, α ′ i The fault strike characteristic value of the i-th grid cell is represented as the characteristic value of the fault strike, and then converted into the angle value α. ′ i This is to facilitate intuitive understanding.

[0107] Let the fault dip angle of the j-th measurement point within the i-th grid cell be denoted as β. ′ ij First, convert it to radians:

[0108]

[0109] In the formula, This is expressed as the fault dip angle in radians at the j-th measurement point within the i-th grid cell, and then calculated. and The fault dip direction refers to the direction in which the fault plane tilts. It is perpendicular to the fault strike and together they describe the tilt state of the fault plane in space. The fault dip angle β... ′ ij Convert to radians This is also to facilitate subsequent trigonometric function calculations.

[0110] The formula for calculating the characteristic value of fault dip is:

[0111]

[0112] In the formula, This is represented by the characteristic value of the fault dip of the i-th grid cell under the radian value.

[0113] Then convert it into an angle value as a characteristic value of the fault dip:

[0114]

[0115] In the formula, β ′ i The eigenvalue representing the fault dip of the i-th grid cell is converted into an angle value β. ′ i This is to facilitate intuitive understanding.

[0116] Let the fault dip angle of the j-th measurement point within the i-th grid cell be denoted as γ. ′ ij The formula for calculating the characteristic value of the fault dip angle is:

[0117]

[0118] In the formula, γ ′ i Let γ represent the characteristic value of the fault dip angle of the i-th grid cell. The fault dip angle is the acute angle between the fault plane and the horizontal plane, which directly reflects the degree of inclination of the fault plane in the vertical direction. The fault dip angle γ is measured at each point within the grid cell. ′ ij The eigenvalue γ of the fault dip angle is obtained by arithmetic mean calculation. ′i This value represents the overall tilt of the fault plane within the grid cell in the vertical direction.

[0119] Let the fault displacement at the j-th measurement point within the i-th grid cell be denoted as η. ij The formula for calculating the characteristic value of fault displacement is:

[0120]

[0121] In the formula, A i The fault displacement is represented by the characteristic value of the fault displacement of the i-th grid cell. Fault displacement is a key parameter describing the characteristics of a fault and can accurately record the distance of relative displacement between the two plates at different locations of the fault.

[0122] Identify all possible rock types and assign a unique code C to each. k Where k = 1, 2, 3, ..., f, f represents the total number of all possible different rock types within the i-th grid cell, forming a rock type coding list. The distribution area of ​​the k-th rock type within the i-th grid cell is set to A. ik The total area is S i The formula for the area proportion of the k-th rock type within this grid cell is expressed as:

[0123]

[0124] In the formula, p ik Let A represent the area percentage of the k-th rock type within the i-th grid cell. ik S represents the distribution area of ​​the k-th rock type within the i-th grid cell. i It is represented as the total area of ​​the i-th grid cell.

[0125] Within each grid cell, the area percentage p of different rock types is compared. ik The rock type with the highest proportion is extracted, and its code is used as the rock type coding feature value for this grid cell. This is because multiple rock types may exist within a single grid cell, but their influence on geological processes and mineralization often varies. By comparing area proportions, the dominant rock type in the grid cell can be identified. This dominant rock type usually plays a key role in the composition and properties of geological bodies.

[0126] Identifying and encoding all possible rock types facilitates differentiation and processing of different rock types in subsequent data analysis and model building. The area percentage (p) of each rock type within the grid cell is calculated. ikThis allows us to understand the distribution of different rock types in the region, i.e., which rock type dominates within the grid cell, which is of great significance for inferring the relationship between mineral formation and rock type.

[0127] The thickness of the strata is one of the important basic data for estimating resource volume;

[0128] The strike of strata describes their orientation on the horizontal plane, helps determine their spatial distribution, and is an important parameter for constructing geological maps and models.

[0129] The dip direction of strata is not only helpful in understanding the mineralization process and inferring the location, range, and shape of concealed ore bodies, but also in reducing the difficulty of mining when formulating mining plans in subsequent underground mining operations.

[0130] The dip angle of strata, that is, the angle between the strata bedding plane and the horizontal plane, has an important influence on the shape and distribution of ore bodies and is helpful in identifying layered ore bodies.

[0131] The distribution range of different rock types within a mining area can help identify potential mineralized zones or deposits by associating rock types with specific types of ore deposits.

[0132] Descriptions of the types and relative contents of minerals in ores help geological prospectors understand the geological characteristics, mineral distribution, and mineralization of the mining area.

[0133] The name of a fault is an identifier for a specific fault, which facilitates rapid location and understanding of the fault.

[0134] The location of the faults was determined in the three-dimensional geological model, which provided key positioning information for constructing accurate three-dimensional geological structures.

[0135] The strike of the fault is described, indicating the direction of the fault plane's extension on the horizontal plane. This is used to analyze the direction of the fault's influence on the continuity of the strata and its control over the migration path of ore-forming fluids on the horizontal plane.

[0136] The dip direction of the fault plane is determined. Faults with different dip directions will cause different displacement modes and results to the ore body, which directly affects the position, shape and continuity of the ore body in three-dimensional space.

[0137] The dip angle of a fault reflects the degree of inclination of the fault plane and has a significant impact on the morphology and distribution of the ore body.

[0138] The fault displacement parameter is the distance of relative movement between the two sides of the fault, including vertical and horizontal displacement. It is of key significance for determining the true distribution range, morphology, and resource quantity assessment of ore bodies.

[0139] Step 2: Take the presence or absence of minerals as the target variable, associate the target variable with its corresponding stratigraphic, rock, and structural feature values ​​to form a set of geological feature data, label the set of all geological feature data as model data, use the model data as the training set, construct a random forest model, set the maximum depth, the minimum number of samples required for node splitting, and the minimum number of samples for leaf nodes, and then use the accuracy to evaluate and optimize the random forest model.

[0140] Data was collected from multiple explored mining areas. After screening and sorting, 1,000 samples were identified, including 500 mineralized samples and 500 non-mineralized samples. Each sample was divided into a three-dimensional grid with a grid size of 5 meters.

[0141] For mineralized samples, a random number generator is used to generate 500 unique random numbers in the range of 1 to 1000. For non-mineralized samples, the same random number generator is used to generate 500 unique random numbers in the range of 1 to 1000.

[0142] A random forest model was constructed. Random forest is an ensemble learning algorithm composed of multiple decision trees, which can integrate the prediction results of multiple decision trees to reduce errors caused by overfitting or data bias in a single decision tree. Initially, the maximum depth was set to 8. Based on preliminary analysis of geological data characteristics and some empirical judgment, it was believed that the relationship between geological factors such as stratigraphic structure, rock type, and tectonic features and mineralization is not extremely simple and direct, requiring a certain depth of decision trees to uncover the complex correlations. However, it could not be too deep to prevent overfitting; 10 was an initial attempt to balance learning ability and generalization ability. The minimum number of samples required for node splitting was set to 8. This was to prevent the decision tree from splitting excessively when data was limited, leading to an overly complex and unstable model. The minimum number of samples for leaf nodes was set to 5 to ensure that leaf nodes had sufficient samples to support their predictions, making the predictions more reliable and representative.

[0143] In each iteration of the random forest model training, sample data from the training set is input into the current random forest model for prediction. Let TP be the number of correctly predicted mineralized samples and TN be the number of correctly predicted non-mineralized samples in the training set, and let e be the total number of mineralized and non-mineralized samples. After each training iteration, the accuracy is calculated according to the formula:

[0144]

[0145] In the formula, Accuracy represents the accuracy rate, e represents the total number of mineralized and non-mineralized samples, TP represents the number of correct predictions for mineralized samples, and TN represents the number of correct predictions for non-mineralized samples. Training stops when the accuracy rate is higher than 90%.

[0146] The model is trained from the start. After each training round, the model is evaluated using the training set, and the overall prediction accuracy of the model on both mineralized and non-mineralized samples in the training set is calculated. For example, after the first round of training, if the model predicts 1000 samples in the training set, correctly predicting mineralization in 400 samples and correctly predicting non-mineralization in 420 samples, then the accuracy is _____. After multiple rounds of training, after the 20th round, the model made predictions on 1000 samples in the training set. Of these, 460 samples correctly predicted mineralization, and 450 samples correctly predicted non-mineralization. Therefore, the accuracy rate was [insert accuracy rate here]. Since the accuracy rate is higher than 90%, training is stopped. At this point, the trained random forest model is obtained and can be used for subsequent hidden mineral prediction.

[0147] Step 3: Extract geological feature data of the area to be tested by querying geological logs, save it as a geological feature data table, and import it into geological modeling software to build a three-dimensional geological model. The geological feature data includes stratigraphic structure, rock type, and structural feature data.

[0148] For the target area to be detected, geological logging is consulted to extract data related to stratigraphic structure, including the name, thickness, strike, dip, and dip angle of each stratum. For rock types, the specific type of each rock is determined according to geological identification standards and descriptions in the logging, and coded according to a pre-set classification and coding system for subsequent data processing and model recognition. For structural features, including the strike, dip, dip angle, and displacement of faults, the extracted stratigraphic structure, rock type, and structural feature data are compiled into a geological feature data table. The names of each stratum generally correspond to specific rock types, fossil assemblages, and geological structures, and some names are globally recognized stratigraphic nomenclatures, such as Cambrian and Ordovician, facilitating rapid location and understanding of stratigraphic conditions and aiding in the inference of mineral formation.

[0149] Then, the prepared stratigraphic data table is imported into the Surpac geological modeling software. Surpac boasts powerful functions and professional algorithms, capable of transforming two-dimensional geological data into three-dimensional visualization models. Based on the imported data, it uses mathematical models and graphics rendering technology to accurately construct the morphology and interrelationships of strata, rocks, and structures in three-dimensional space. In concealed mineralization prediction, three-dimensional geological models help researchers better understand the spatial coupling relationship between geological structures and mineralization processes, thereby more accurately determining the potential location and extent of concealed mineralization.

[0150] In the software environment, a stratigraphic model is constructed sequentially based on the stratigraphic names and thickness data from the stratigraphic data table. For each stratigraphic layer, its position and shape in three-dimensional space are set according to its strike, dip, and dip angle parameters, ensuring the stratigraphic model accurately reflects the actual distribution of underground strata. For rock type data, the software's material assignment function is used to assign corresponding colors or materials to different strata or geological bodies according to a pre-defined correspondence between rock type and material color. This allows for a direct observation of the distribution characteristics of rock types in the three-dimensional model; for example, sandstone strata are set to a yellow material, and granite strata to a gray material. For fault information in the structural feature data, fault objects are created in the software, and the faults are drawn in the three-dimensional model based on their strike, dip, dip angle, and displacement parameters.

[0151] Stratigraphic structure provides the spatial framework for the occurrence of concealed mineralization. Different stratigraphic thicknesses, strikes, dips, and angles influence the migration paths of ore-forming fluids and the locations of mineral deposition. Rock types are closely related to mineralization; certain rock types may be rich in specific ore-forming elements or possess physicochemical properties conducive to their enrichment. Tectonic features, especially faults, often serve as important channels for ore-forming fluids and favorable sites for mineral deposition. Fault activity can alter the permeability and pressure state of strata, promoting the migration and accumulation of ore-forming elements. Integrating these three factors into a three-dimensional geological model can comprehensively reflect the complex relationship between the geological environment and concealed mineralization, providing complete geological data for model-based concealed mineralization prediction.

[0152] Step 4: Divide the constructed 3D geological model into cubic meshes. Based on the geological feature data table, input the geological feature data corresponding to each mesh unit into the optimized random forest model. The optimized random forest model will then predict the mesh units that may contain concealed minerals, thereby realizing the location prediction of concealed minerals in 3D space.

[0153] Activate the mesh generation function in the software. Set the mesh size to 5 meters. The software will divide the entire spatial range of the 3D geological model into several cubic mesh units with a side length of 5 meters according to the set parameters, so that each mesh unit can become a relatively independent and representative geological sample unit.

[0154] From the geological feature data table, obtain the geological feature data corresponding to each grid cell. For stratigraphic structure data, calculate the average stratigraphic thickness at each point within the grid cell as the feature value of the stratigraphic thickness for that cell. Calculate the average stratigraphic dip angle as the feature value of the dip angle. For rock type data, determine the rock type coding feature value based on the distribution area ratio of rock types within the grid cell or the dominant rock type. For structural feature data, detect whether faults exist within the grid cell; if so, calculate the feature values ​​of fault strike, dip, dip angle, and displacement.

[0155] Let h be the formation thickness at the j-th measurement point within the i-th grid cell. ij The total number of measurement points in the grid cell is n, i is used to distinguish different grid cells, each grid cell has its own independent set of measurement point data and corresponding formation thickness characteristic value calculation process, and j is a variable for numbering each measurement point in the i-th grid cell;

[0156] The formula for calculating the characteristic value of the formation thickness is:

[0157]

[0158] Let the strike angle of the stratigraphy at the j-th measurement point within the i-th grid cell be denoted as α. ij First, set the angle value α ij Converted to radians for calculation, the converted radian value is:

[0159]

[0160] In the formula, θ ij Let x be the strike angle in radians of the formation at the j-th measurement point within the i-th grid cell, and then calculate x. ij =cosθ ij and y ij =sinθ ij .

[0161] The formula for calculating the characteristic value of the stratigraphic strike is:

[0162]

[0163] In the formula, θ i It represents the characteristic value of the stratigraphic strike of the i-th grid cell under the radian value.

[0164] Then take the radian value θ i Converted into angular values ​​as characteristic values ​​of stratigraphic strike:

[0165]

[0166] In the formula, α i The characteristic value representing the stratigraphic strike of the i-th grid cell is denoted as .

[0167] Let the dip angle of the stratigraphy at the j-th measurement point within the i-th grid cell be denoted as β. ij First, convert it to radians:

[0168]

[0169] In the formula, Represented as the stratigraphic dip angle in radians at the j-th measurement point within the i-th grid cell, then calculated... and

[0170] The formula for calculating the characteristic value of the stratigraphic dip is:

[0171]

[0172] In the formula, This is represented by the characteristic value of the stratigraphic dip of the i-th grid cell under the radian value.

[0173] Then convert it into an angle value as a characteristic value of the stratigraphic dip:

[0174]

[0175] Let the dip angle of the formation at the j-th measurement point within the i-th grid cell be denoted as γ. oj The formula for calculating the characteristic value of the dip angle of the strata is:

[0176]

[0177] In the formula, γ i The eigenvalue representing the dip angle of the formation in the i-th grid cell is denoted as .

[0178] For constructing feature data, detect whether there are faults within the grid cells. If they exist, calculate the characteristic values ​​of the fault strike, dip, dip angle, and fault displacement.

[0179] Let the fault strike angle at the j-th measurement point within the i-th grid cell be denoted as α. ′ ij First, set the angle value α ij Converted to radians for calculation, the converted radian value is:

[0180]

[0181] In the formula, θ ′ ij Let x be the fault strike angle in radians at the j-th measurement point within the i-th grid cell, then calculate x. ′ ij =cosθ ij and y ′ ij =sinθ ij ,

[0182] The formula for calculating the characteristic value of fault strike is:

[0183]

[0184] In the formula, θ ′ i It represents the characteristic value of the fault strike of the i-th grid cell under the radian value.

[0185] Then take the radian value θ ′ i Convert the fault angle value into a characteristic value representing the fault strike:

[0186]

[0187] In the formula, α ′ i The characteristic value representing the fault strike of the i-th grid cell is denoted as .

[0188] Let the fault dip angle of the j-th measurement point within the i-th grid cell be denoted as β. ′ ij First, convert it to radians:

[0189]

[0190] In the formula, This is expressed as the fault dip angle in radians at the j-th measurement point within the i-th grid cell, and then calculated. and

[0191] The formula for calculating the characteristic value of fault dip is:

[0192]

[0193] In the formula, This is represented by the characteristic value of the fault dip of the i-th grid cell under the radian value.

[0194] Then convert it into an angle value as a characteristic value of the fault dip:

[0195]

[0196] In the formula, β ′ i The eigenvalue representing the fault dip of the i-th grid cell is converted into an angle value β. ′ i This is to facilitate intuitive understanding.

[0197] Let the fault dip angle of the j-th measurement point within the i-th grid cell be denoted as γ. ′ ij The formula for calculating the characteristic value of the fault dip angle is:

[0198]

[0199] In the formula, γ ′ i The eigenvalue representing the fault dip angle of the i-th grid cell is denoted as .

[0200] Let the fault displacement at the j-th measurement point within the i-th grid cell be denoted as η. ij The formula for calculating the characteristic value of fault displacement is:

[0201]

[0202] In the formula, A i The characteristic value of the fault displacement of the i-th grid cell is represented as .

[0203] Identify all possible rock types and assign a unique code C to each. k Where k = 1, 2, 3, ..., f, f represents the total number of all possible different rock types within the i-th grid cell, forming a rock type coding list. The distribution area of ​​the k-th rock type within the i-th grid cell is set to A. ik The total area is S i The formula for the area proportion of the k-th rock type within this grid cell is expressed as:

[0204]

[0205] In the formula, p ik Let A represent the area percentage of the k-th rock type within the i-th grid cell. ik S represents the distribution area of ​​the k-th rock type within the i-th grid cell. i It is represented as the total area of ​​the i-th grid cell.

[0206] Within each grid cell, the area percentage p of different rock types is compared. ik Extract the rock type with the highest proportion and use the code of that rock type as the rock type code feature value of this grid cell.

[0207] After the geological feature data of all grid cells are organized in the prescribed format, they are input into the pre-trained random forest model. The model will predict the probability of hidden mineralization in each grid cell based on the complex relationship between geological features and mineralization learned during the training phase. These feature data undergo a series of complex calculations and decision-making processes within the model, and finally output a probability value representing the probability of hidden mineralization in that grid cell.

[0208] When more than 60% of the decision trees judge a grid cell as having hidden mineral deposits, the grid cell is marked as having hidden mineral deposits. By organizing and analyzing the prediction results of all grid cells, the areas that the model judges to have hidden mineral deposits can be located in three-dimensional space.

[0209] Please see Figure 2 The present invention also provides a three-dimensional quantitative prediction system for concealed minerals, the system being used to execute the above-described three-dimensional quantitative prediction method for concealed minerals, comprising:

[0210] The feature extraction module is used to collect an equal number of mineralized and non-mineralized model samples, divide all model samples into grid cells, and use the stratigraphic thickness, strike, dip, and dip angle of the stratigraphic structure of each grid cell as stratigraphic feature values, classify and encode rock types as rock feature values, and use the fault strike, dip, dip angle, and fault displacement of structural features as structural feature values.

[0211] The model building module is used to take the presence or absence of minerals as the target variable, associate the target variable with its corresponding stratigraphic, rock, and structural feature values ​​to form a set of geological feature data, label the set of all geological feature data as model data, use the model data as the training set, build a random forest model, and after setting the maximum depth, the minimum number of samples required for node splitting, and the minimum number of samples for leaf nodes, the accuracy is used to evaluate and optimize the random forest model.

[0212] The 3D modeling module is used to extract geological feature data of the area to be detected by querying geological logs, save it as a geological feature data table, and import it into geological modeling software to build a 3D geological model. The geological feature data includes stratigraphic structure, rock type, and structural feature data.

[0213] The concealed ore location module is used to divide the constructed 3D geological model into cubic meshes. Based on the geological feature data table, the geological feature data corresponding to each mesh unit is input into the optimized random forest model. The optimized random forest model predicts the mesh units that may contain concealed ore, thereby realizing the location prediction of concealed ore in 3D space.

[0214] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0215] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0216] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple grid units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0217] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A three-dimensional quantitative prediction method for concealed mineral deposits, characterized in that, The specific steps include: Step 1: Collect an equal number of mineralized and non-mineralized model samples, divide all model samples into grid cells, and use the thickness, strike, dip, and dip angle of the stratigraphic system of each grid cell as stratigraphic feature values, classify and encode the rock type as rock feature values, and use the fault strike, dip, dip angle, and fault displacement of the structural features as structural feature values. Step 2: Take the presence or absence of minerals as the target variable, associate the target variable with its corresponding stratigraphic, rock, and structural feature values ​​to form a set of geological feature data, label the set of all geological feature data as model data, use the model data as the training set, construct a random forest model, set the maximum depth, the minimum number of samples required for node splitting, and the minimum number of samples for leaf nodes, and then use the accuracy to evaluate and optimize the random forest model. Step 3: Extract geological feature data of the area to be tested by querying geological logs, save it as a geological feature data table, and import it into geological modeling software to build a three-dimensional geological model. The geological feature data includes stratigraphic system, rock type, and structural feature data. Step 4: Divide the constructed 3D geological model into cubic meshes. Based on the geological feature data table, input the geological feature data corresponding to each mesh unit into the optimized random forest model. The optimized random forest model will then predict the mesh units that may contain concealed minerals, thereby realizing the location prediction of concealed minerals in 3D space.

2. The method for three-dimensional quantitative prediction of concealed mineral deposits according to claim 1, characterized in that: In step 1, the thickness, strike, dip, and dip angle of the stratigraphic system of each grid cell are used as stratigraphic characteristic values; rock types are classified and coded as rock characteristic values; and the strike, dip, dip angle, and displacement of faults are used as structural characteristic values. First, the collected mineralized and non-mineralized model samples are divided into grids. Then, the formation thickness at the j-th measurement point within the i-th grid cell is denoted as h. ij The total number of measurement points in the grid cell is n, i is used to distinguish different grid cells, each grid cell has its own independent set of measurement point data and corresponding formation thickness characteristic value calculation process, and j is a variable for numbering each measurement point in the i-th grid cell; The formula for calculating the characteristic value of the formation thickness is: Let the strike angle of the stratigraphy at the j-th measurement point within the i-th grid cell be denoted as α. ij First, set the angle value α ij Converted to radians for calculation, the converted radian value is: In the formula, θ ij Let x be the strike angle in radians of the formation at the j-th measurement point within the i-th grid cell, and then calculate x. ij =cosθ ij and y ij =sinθ ij , The formula for calculating the characteristic value of the stratigraphic strike is: In the formula, θ i The eigenvalue representing the stratigraphic strike of the i-th grid cell in radians. Then take the radian value θ i Converted into angular values ​​as characteristic values ​​of stratigraphic strike: In the formula, α i The characteristic value representing the stratigraphic strike of the i-th grid cell is denoted as . Let the dip angle of the stratigraphy at the j-th measurement point within the i-th grid cell be denoted as β. ij First, convert it to radians: In the formula, Represented as the stratigraphic dip angle in radians at the j-th measurement point within the i-th grid cell, then calculated... and The formula for calculating the characteristic value of the stratigraphic dip is: In the formula, It is represented as the characteristic value of the stratigraphic dip of the i-th grid cell under the radian value; Then convert it into an angle value as a characteristic value of the stratigraphic dip: Let the dip angle of the formation at the j-th measurement point within the i-th grid cell be denoted as γ. ij The formula for calculating the characteristic value of the dip angle of the strata is: In the formula, γ i The eigenvalue representing the dip angle of the formation in the i-th grid cell is denoted as . Identify all possible rock types and assign a unique code C to each. k Where k = 1, 2, 3, ..., f, f represents the total number of all possible different rock types within the i-th grid cell, forming a rock type coding list. The distribution area of ​​the k-th rock type within the i-th grid cell is set to A. ik The total area is S i The formula for the area proportion of the k-th rock type within this grid cell is expressed as: In the formula, p ik Let A represent the area percentage of the k-th rock type within the i-th grid cell. ik S represents the distribution area of ​​the k-th rock type within the i-th grid cell. i Represented as the total area of ​​the i-th grid cell, Within each grid cell, the area percentage p of different rock types is compared. ik Extract the rock type with the highest proportion and use the code of that rock type as the rock type code feature value of this grid cell.

3. The method for three-dimensional quantitative prediction of concealed mineral deposits according to claim 1, characterized in that: Step 2, using the model data as the training set, includes the following steps: Using the presence or absence of mineral deposits as the stratification variable, and with a total number of mineralized and non-mineralized samples of e, a random number generator is used to generate random numbers within the range of 1 to e for all mineralized samples. Use a unique random number generator to add the corresponding numbered mineralized samples to the training set. For non-mineralized samples, use a random number generator to generate a random number within the range of 1 to e. Use a unique random number to add all the corresponding unmineralized samples to the training set. When the model accurately predicts the proportion of mineralized and unmineralized samples in the training set, i.e., the accuracy rate, is higher than 90%, training stops.

4. The method for three-dimensional quantitative prediction of concealed mineral deposits according to claim 1, characterized in that: Step 2, setting the maximum depth, the minimum number of samples required for node splitting, and the minimum number of samples for leaf nodes includes the following steps: Initially, the maximum depth was set to 8. Using the pre-divided training set, a random forest model was constructed. The minimum number of samples required for node splitting was set to 8, and the minimum number of samples for leaf nodes was set to 5.

5. The method for three-dimensional quantitative prediction of concealed mineral deposits according to claim 1, characterized in that: In step 2, the method for evaluating and optimizing the random forest model using accuracy is as follows: In each iteration of the random forest model training, sample data from the training set is input into the current random forest model for prediction. Let TP be the number of correctly predicted mineralized samples and TN be the number of correctly predicted non-mineralized samples in the training set, and let e be the total number of mineralized and non-mineralized samples. After each training iteration, the accuracy is calculated according to the formula: In the formula, Accuracy represents the accuracy rate, e represents the total number of mineralized and non-mineralized samples, TP represents the number of correct predictions for mineralized samples, and TN represents the number of correct predictions for non-mineralized samples. Training should stop when the accuracy rate is above 90%.

6. The method for three-dimensional quantitative prediction of concealed mineral deposits according to claim 1, characterized in that: In step 3, the method for constructing the three-dimensional geological model is as follows: The names, thicknesses, strikes, dips, and dip angles of each stratum in the stratigraphic system data are saved as a geological feature data table in a unified text format. This data is then imported into geological modeling software, where different rock types are distinguished by different colors or textures. This allows the model to visually reflect the distribution characteristics of rock types. Structural feature data is integrated into the model to show the location, shape, and intersection relationship of faults with strata in three-dimensional space, thus constructing a complete three-dimensional geological model that includes stratigraphic structure, rock types, and structural features.

7. The method for three-dimensional quantitative prediction of concealed mineral deposits according to claim 1, characterized in that: Step 4, the location prediction of concealed ore in three-dimensional space includes the following steps: From the constructed 3D model of the area to be detected, feature data is extracted in the same way as when training the model, including stratum thickness, strike, dip, dip angle, rock type encoding, and fault strike, dip, dip angle, and displacement. The 3D space of the area to be detected is divided into cubic meshes, and each mesh cell is regarded as an independent sample point. All sample data are input into the trained random forest model. Based on the input feature data, the relationship between geological features and mineralization learned during training is used. After the model prediction results are output, a probability threshold μ is set. When the probability value of the predicted result of a certain mesh cell corresponding to the existence of hidden minerals is higher than the probability threshold μ, it is determined that there are hidden minerals at the location of this mesh cell; otherwise, if the probability value does not reach the probability threshold μ, it is determined that there are no hidden minerals at this location.

8. A three-dimensional quantitative prediction system for concealed mineral deposits, characterized in that: The system is used to execute a three-dimensional quantitative prediction method for concealed mineral deposits as described in any one of claims 1-7, comprising: The feature extraction module is used to collect an equal number of mineralized and non-mineralized model samples, divide all model samples into grid cells, and use the thickness, strike, dip, and dip angle of the stratigraphic system of each grid cell as stratigraphic feature values, classify and encode rock types as rock feature values, and use the fault strike, dip, dip angle, and fault displacement of structural features as structural feature values. The model building module is used to take the presence or absence of minerals as the target variable, associate the target variable with its corresponding stratigraphic, rock, and structural feature values ​​to form a set of geological feature data, label the set of all geological feature data as model data, use the model data as the training set, build a random forest model, and after setting the maximum depth, the minimum number of samples required for node splitting, and the minimum number of samples for leaf nodes, the accuracy is used to evaluate and optimize the random forest model. The 3D modeling module is used to extract geological feature data of the area to be detected by querying geological logs, save it as a geological feature data table, and import it into geological modeling software to build a 3D geological model. The geological feature data includes stratigraphic structure, rock type, and structural feature data. The concealed ore location module is used to divide the constructed 3D geological model into cubic meshes. Based on the geological feature data table, the geological feature data corresponding to each mesh unit is input into the optimized random forest model. The optimized random forest model predicts the mesh units that may contain concealed ore, thereby realizing the location prediction of concealed ore in 3D space.

Images