Method for rapidly delineating gold mine prospecting target area

Abstract

The invention relates to the technical field of geological science, in particular to a method for quickly delineating a gold ore prospecting target area, which comprises the following steps of: 1, constructing an ore deposit geological feature library; step 2, multi-source heterogeneous data integration and basic drawing dynamic compilation; 3, reconstructing a construction-metallogenic event sequence and dynamically dividing a metallogenic system; 4, regional ore deposit pedigree construction and representative ore deposit combination multi-method collaborative dating are carried out; 5, multi-stage abnormal superposition analysis and dynamic metallogenic geological necessary condition extraction are carried out; step 6, dynamic element combinatorial analysis and nonlinear integrated prediction modeling; according to the method, under the constraint of a tectonic-thermal event sequence, geologic-tectonic domains are divided, exclusive element combinations of ore deposits in different periods in the geologic-tectonic domains are recognized, a machine learning algorithm is adopted to train a nonlinear gold mineralization strength prediction model, a continuous mineralization favor index is generated, and the target classification dynamic verification is carried out. And the prediction precision of the mineralization intensity space change is obviously improved.

Description

Technical Field

[0001] This invention relates to the field of geological science and technology, and in particular to a method for rapidly delineating gold prospecting target areas. Background Technology

[0002] Gold, as a critical global strategic resource, faces challenges in exploration from traditional single-method approaches (such as geological mapping, geochemical exploration, and geophysical exploration), including low efficiency, high cost, long cycles, and difficulties in identifying complex geological conditions. By integrating multi-source heterogeneous data and combining artificial intelligence algorithms for feature extraction, correlation analysis, and pattern recognition of multi-dimensional information, this approach enables rapid target area screening, precise location of anomalies, quantitative assessment of mineralization potential, and intelligent optimization of exploration paths. This improves mineral exploration efficiency, reduces exploration costs, minimizes ineffective drilling, and increases the probability of gold discovery. Through the integration of multiple technologies, this approach overcomes the limitations of traditional single-method approaches, accelerates the discovery of concealed ore bodies, optimizes the allocation of exploration resources to support national resource security and local economic development, and further promotes the transformation of geological exploration models from experience-driven to data-intelligent-driven, fostering the deep cross-integration of geology, computer science, and big data technologies.

[0003] In existing techniques, anomalous elements in gold deposits within the study area are statistically analyzed, and their correlation coefficients with Au are calculated. Then, a comprehensive predicted value is calculated from the correlation coefficients. Following geochemical element data processing methods, the lower limit for predicted Au anomalies is calculated for geochemical anomaly delineation. The correlation coefficients are calculated using R-type cluster analysis. However, R-type cluster analysis assumes that elemental correlations are stable in the temporal-spatial dimension. But multi-phase tectonic activity can cause elemental combinations to change with tectonic phases. For example, Au is strongly correlated with As during the Early Paleozoic mineralization period, while the correlation between Au and Hg increases during the Mesozoic alteration period. Traditional methods cannot dynamically capture this change. Therefore, a method for rapidly delineating gold prospecting target areas is proposed. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a method for rapidly delineating gold prospecting target areas.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A method for rapidly delineating gold prospecting target areas includes the following steps: Step 1: Construction of the geological feature database of ore deposits: Through field geological surveys, core logging, systematic sampling and rock and mineral identification, geochemical and isotopic analysis of all known gold deposits and mineral occurrences in the study area, the geological features of each object are obtained and structured to form a feature database containing information on metallogenic epoch, ore-bearing lithology, alteration assemblages, mineral paragenesis sequences and elemental assemblages. Step 2: Integration of multi-source heterogeneous data and dynamic compilation of basic maps: Integrate high-precision geological maps of the study area, multi-period isotope dating data, tectonic deformation measurement data, and original geophysical, geochemical, remote sensing, and heavy mineral data. Use 3D geological modeling software to construct a spatiotemporally unified multidimensional data base, compile dynamic geological and mineral maps and tectonic-thermal event sequence maps, and form basic maps for gold deposit prediction that support the analysis of multi-period tectonic-metallogenic events. Step 3: Reconstruction of tectonic-metallogenic event sequence and dynamic classification of metallogenic systems: Based on the multidimensional data foundation, the sequence of multiple tectonic-thermal events after the orogenic period in the study area is reconstructed, and dynamic metallogenic system categories are classified accordingly. The spatial superposition and transformation relationship of metallogenic systems at different stages is analyzed. Step 4: Construction of regional ore deposit series and multi-method collaborative dating of representative ore deposit combinations: Based on the tectonic-metallogenic event sequence and ore deposit geological characteristics, a clustering algorithm is used to divide the ore deposit series, select representative ore deposit combinations of each series, and use multiple isotope dating techniques for cross-dating to determine the main metallogenic period and the later transformation period, and construct a dynamic metallogenic model with superimposed metallogenic elements of multiple periods. Step 5: Analysis of superimposed anomalies in multiple phases and extraction of necessary conditions for dynamic metallogenic geology: Based on the ore deposit series and the dynamic metallogenic model, machine learning image segmentation technology is used to identify and distinguish remote sensing alteration anomalies of different phases. Combined with geochemical element combinations and geophysical anomaly characteristics, phased anomaly identification markers are established, and necessary conditions for dynamic metallogenic geology coupling the three elements of "favorable metallogenic geological bodies, multi-phase tectonic superimposed structural surfaces, and fluid migration channel network" are extracted. Step Six: Dynamic Element Combination Analysis and Nonlinear Integrated Prediction Modeling: Under the constraints of the tectonic-thermal event sequence, geological-tectonic domains are divided and the exclusive element combinations of different periods of ore deposits within each geological-tectonic domain are identified. A nonlinear gold mineralization intensity prediction model is trained using machine learning algorithms to generate a continuous mineralization favorability index and adaptively delineate geochemical anomalies (unsupervised learning models such as deep autoencoders are used to learn the spatial distribution of the mineralization favorability index, automatically identify outlier features in the data distribution, and thus non-parametrically determine the anomaly threshold to achieve adaptive delineation of anomaly boundaries). Step 7: Target Area Classification and Dynamic Verification: Based on the mineralization favorability index and dynamic mineralization geological necessity conditions, the target area is delineated and classified using three-dimensional geological modeling technology. Differentiated verification projects are implemented to obtain verification data, and the verification results are fed back into the model of the previous steps to form a dynamic closed loop of prediction-verification-feedback-optimization, and finally the mineral exploration target area is determined.

[0006] The above further includes: Furthermore, the formation process of the gold mine prediction base map is as follows: Multi-source heterogeneous data preprocessing: performing coordinate system standardization, format standardization, and quality checks on multi-source heterogeneous data; Construction of a multidimensional data base: The pre-processed multi-source heterogeneous data is imported into the 3D geological modeling software. Based on geographic coordinates and stratigraphic age, spatial registration and temporal attribute association are performed to establish a multidimensional data base that includes spatial geometry, stratigraphic age, rock properties, tectonic elements, geophysical field, geochemical field and thermal event age labels. Dynamic geological and mineral map compilation: With the support of the multi-dimensional data base, based on the time slice or tectonic-thermal event period set by the user, the stratigraphic boundaries, magmatic bodies, tectonic features and distribution information of known mineral deposits or mineralization points within the corresponding spatiotemporal range are dynamically extracted and rendered. Through the map section and attribute filtering functions of the three-dimensional geological modeling software, dynamic geological and mineral maps reflecting the geological and mineral features of a specific geological historical period are automatically generated. Interpretation and generation of tectonic-thermal event sequence diagrams: Based on the multi-stage isotopic dating data and tectonic deformation data in the multi-dimensional data base, the spatiotemporal analysis module of the 3D geological modeling software is used to identify the concentrated intervals of age data using density clustering algorithm to identify the main thermal event stages. Combined with the spatial intersection relationship of tectonic deformation patterns, the sequence of their order is determined, and a tectonic-thermal event sequence diagram is automatically drawn with time as the vertical axis, which intuitively displays the age range, tectonic properties, and spatial influence range of each thermal event stage. Multi-phase tectonic-metallogenic event correlation analysis and basic map synthesis: The dynamic geological and mineral map is correlated with the tectonic-thermal event sequence map. In the multi-dimensional data base, through spatial overlay and attribute linking, the causal relationship between specific deposits or mineralization points and specific tectonic-thermal event periods is identified. A gold deposit prediction basic map that simultaneously carries spatial geological and mineral information and temporal evolution sequence information is synthesized. This map supports dynamic switching of geological scenes based on time axis sliding or event period selection and the overlay display of multi-phase tectonic-metallogenic events.

[0007] Furthermore, the reconstruction process of the construction-thermal event sequence is as follows: Paleotectonic framework restoration: The study area is placed in the context of the tectonic unit to which it belongs. Using plate reconstruction models, the paleogeographic location, plate movement direction and rate change trajectory after the orogenic period of the study area are restored, and a regional tectonic dynamic background evolution framework is established. Event spatiotemporal visualization and clustering: Isotopic age data of the study area are visualized in a 3D geological modeling environment according to spatial location, rock unit, and measured mineral type. Spatiotemporal sequence matching algorithm and cluster analysis method are used to identify the concentrated distribution intervals (peak periods) of age data in the time domain. These concentrated distribution intervals represent important magmatic, metamorphic, or hydrothermal events. Combined with the period division of tectonic deformation data, the concentrated distribution intervals are associated with specific tectonic movements (such as collision compression, post-collision extension, and large-scale strike-slip), defining clear tectonic-thermal events and arranging them in chronological order to form a preliminary tectonic-thermal event sequence.

[0008] Furthermore, in step four, the classification of ore deposit series specifically involves: based on the genetic type of the ore deposit, the host rocks, the alteration mineral assemblage, and the spatiotemporal correlation with specific tectonic-thermal events, a hierarchical clustering algorithm is used to classify all known ore deposit points in the study area, forming a genetic-stage coupled classification system such as "Early Paleozoic orogenic gold deposit series" or "Mesozoic porphyry-skarn copper-molybdenum-gold deposit series".

[0009] Furthermore, the specific steps for constructing a dynamic metallogenic model based on the superposition of multiple metallogenic elements are as follows: dividing the ore deposit series. Within each of the aforementioned ore deposit series, based on the ore deposit scale, the degree of geological research, the completeness of data, and the representativeness of spatial distribution, several (no fewer than two) ore deposits are selected as representative ore deposit combinations for that series. At the same time, the spatial clustering characteristics of ore deposit points within the same series are analyzed using a DBSCAN-based spatial clustering algorithm to ensure that the selected combinations can reflect the dominant spatial distribution patterns and ore-controlling tectonic background of the ore deposits in that series under the influence of major tectonic-thermal events. Multi-mineral-multi-method isotopic cross-dating and age determination: For selected representative deposit assemblages, mineral samples from different generations are collected for dating, including but not limited to: zircon from magmatic rocks closely related to mineralization, hydrothermal zircon or sphene from hydrothermal veins, metallic sulfides (such as molybdenite and pyrite), and hydrothermal alteration minerals (such as mica and potassium feldspar). By comparing and analyzing the dating results of different minerals within the same deposit, cross-validation is performed to determine the main mineralization period and identify possible later hydrothermal alteration periods or tectonic thermal disturbance events. Constructing a dynamic metallogenic model with superimposed metallogenic elements across multiple phases: Integrating ore deposit series, spatial distribution characteristics, and the time sequence of metallogenesis and alteration, and for each ore deposit series, constructing a flowchart of superimposed metallogenic elements across multiple phases in chronological order.

[0010] Furthermore, in step five, machine learning image segmentation technology is used to identify and distinguish remote sensing alteration anomalies of different periods. This means that the U-Net convolutional neural network model is used to train multispectral or hyperspectral remote sensing images to extract silicification and sericitization anomalies related to early deep-seated hydrothermal activity and mudstone and claystone anomalies related to late-stage shallow-seated activity.

[0011] Furthermore, in step six, a nonlinear gold mineralization intensity prediction model is trained using a machine learning algorithm, specifically as follows: Using known ore deposits and mineralization points as label samples, and phased anomaly identification markers as input features, the model is trained using XGBoost or support vector regression algorithms, and the output is a mineralization favorability index, which is a continuous variable.

[0012] Furthermore, in step seven, the target area is classified into at least three levels based on the complexity of the mineralization system, the intensity and scale of multiple superimposed anomalies, and the three-dimensional spatial resource potential estimation results: high-confidence deep target area, medium-shallow verification target area, and prospective exploration area.

[0013] The present invention has the following beneficial effects: In this invention, under the constraint of tectonic-thermal event sequences, geological-tectonic domains are divided and the unique element combinations of different mineral deposits in each geological-tectonic domain are identified. The unique element combinations of different mineralization periods are identified, breaking the assumption of global static correlation. A nonlinear gold mineralization intensity prediction model is trained using machine learning algorithms to generate a continuous mineralization favorability index, which can handle complex coupling relationships. The output continuous mineralization favorability index significantly improves the prediction accuracy and reliability of spatial variation of mineralization intensity. Attached Figure Description

[0014] Figure 1 This is a flowchart illustrating the steps of a method for rapidly delineating gold prospecting target areas proposed in this invention. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] Please see Figure 1 As shown, this invention provides a method for rapidly delineating gold prospecting target areas, comprising the following steps: Step 1: Construction of the geological feature database of ore deposits: Through field geological surveys, core logging, systematic sampling and rock and mineral identification, geochemical and isotopic analysis of all known gold deposits and mineral occurrences in the study area, the geological features of each object are obtained and structured to form a feature database containing information on metallogenic epoch, ore-bearing lithology, alteration assemblages, mineral paragenesis sequences and elemental assemblages. Step Two: Integration of Multi-Source Heterogeneous Data and Dynamic Compilation of Basic Maps: Integrate high-precision geological maps, multi-period isotope dating data, tectonic deformation measurement data, and original geophysical, geochemical, remote sensing, and heavy mineral data of the study area. Use 3D geological modeling software to construct a spatiotemporally unified multidimensional data base, compile dynamic geological and mineral maps and tectonic-thermal event sequence maps, and form basic maps for gold deposit prediction that support multi-period tectonic-metallogenic event analysis. More specifically, the integrated multi-source heterogeneous data includes at least high-precision geological maps with scales between 1:50,000 and 1:250,000, zircon U-Pb and mica 40Ar-39Ar isotope dating data, tectonic deformation foliation and lineation measurement data, as well as geophysical exploration data, stream sediment geochemical data, multispectral and hyperspectral remote sensing data, and natural heavy mineral measurement data. Spatiotemporal registration and fusion are performed using 3D geological modeling software. Step 3: Reconstruction of tectonic-metallogenic event sequence and dynamic classification of metallogenic systems: Based on the multidimensional data foundation, the sequence of multiple tectonic-thermal events after the orogenic period in the study area is reconstructed, and dynamic metallogenic system categories are classified accordingly. The spatial superposition and transformation relationship of metallogenic systems at different stages is analyzed. It should be noted that the specific analytical process for analyzing the spatial superposition and transformation relationships of ore-forming systems at different stages is as follows: For each major tectonic-thermal event phase in the tectonic-thermal event sequence (such as the "Early Paleozoic Main Collision Phase" and the "Mesozoic Late Triassic Post-Collision Extension Phase"), analyze the common characteristics of all known mineral deposits (points) formed under that phase. If the ore-forming materials of a certain period of deposit mainly originate from the mantle or a mixture of crust and mantle, and the ore body is located in the contact zone between the inner and outer parts of a magmatic rock body of the same period, and its genesis is directly related to clear magmatic hydrothermal activity, then the mineralization of this period is classified as a magmatic-hydrothermal mineralization system and given the prefix of tectonic period. If a certain period of deposit is clearly controlled by a large shear zone of the same period, and the ore-forming fluid is mainly metamorphic fluid or tectonic water, and the ore body is hosted in a specific metamorphic rock layer or tectonic rock, then it is classified as a tectonic-metamorphic hydrothermal mineralization system. For mineral deposits that are formed in sedimentary basins and are closely related to specific sedimentary facies and paleogeographic environments, they are classified as sedimentary metallogenic systems. This classification is dynamic, and the same region may develop completely different types of metallogenic systems under different tectonic-thermal events. The spatial extent of the mineralization systems belonging to different tectonic-thermal events is reconstructed and overlaid in three dimensions.

[0017] Spatial overlay identification: Using three-dimensional spatial analysis tools, examine the spatial overlap, intersection, or superposition of different mineralization systems; Determining the nature of superposition relationships based on geological evidence, including: Inheritance and enrichment: Later mineralization is based on the early mineralized body or ore source layer, and further enrichment into mineralization is manifested as the superposition of element combinations; Modification and destruction: Later tectonic activities cause displacement, dismemberment, or dilution of early ore bodies; Shielding and preservation: Later sedimentary basins overlying earlier mineralization systems act as a protective layer. These distinctions require comprehensive confirmation in conjunction with microscopic evidence from typical deposits.

[0018] Step 4: Construction of regional ore deposit series and multi-method collaborative dating of representative ore deposit combinations: Based on the tectonic-metallogenic event sequence and ore deposit geological characteristics, a clustering algorithm is used to divide the ore deposit series, select representative ore deposit combinations of each series, and use multiple isotope dating techniques for cross-dating to determine the main metallogenic period and the later transformation period, and construct a dynamic metallogenic model with superimposed metallogenic elements of multiple periods. Step 5: Analysis of superimposed anomalies in multiple phases and extraction of necessary conditions for dynamic metallogenic geology: Based on the ore deposit series and the dynamic metallogenic model, machine learning image segmentation technology is used to identify and distinguish remote sensing alteration anomalies of different phases. Combined with geochemical element combinations and geophysical anomaly characteristics, phased anomaly identification markers are established, and necessary conditions for dynamic metallogenic geology coupling the three elements of "favorable metallogenic geological bodies, multi-phase tectonic superimposed structural surfaces, and fluid migration channel network" are extracted. Step Six: Dynamic Element Combination Analysis and Nonlinear Integrated Prediction Modeling: Under the constraints of the tectonic-thermal event sequence, geological-tectonic domains are divided and the exclusive element combinations of different periods of ore deposits within each geological-tectonic domain are identified. A nonlinear gold mineralization intensity prediction model is trained using machine learning algorithms to generate a continuous mineralization favorability index and adaptively delineate geochemical anomalies (unsupervised learning models such as deep autoencoders are used to learn the spatial distribution of the mineralization favorability index, automatically identify outlier features in the data distribution, and thus non-parametrically determine the anomaly threshold to achieve adaptive delineation of anomaly boundaries). Step 7: Target Area Classification and Dynamic Verification: Based on the mineralization favorability index and dynamic mineralization geological necessity conditions, the target area is delineated and classified using three-dimensional geological modeling technology. Differentiated verification projects are implemented to obtain verification data, and the verification results are fed back into the model of the previous steps to form a dynamic closed loop of prediction-verification-feedback-optimization, and finally the mineral exploration target area is determined.

[0019] In one embodiment, the process of forming the gold mine prediction base map is as follows: Multi-source heterogeneous data preprocessing: performing coordinate system standardization, format standardization, and quality checks on multi-source heterogeneous data; Construction of a multidimensional data base: The pre-processed multi-source heterogeneous data is imported into the 3D geological modeling software. Based on geographic coordinates and stratigraphic age, spatial registration and temporal attribute association are performed to establish a multidimensional data base that includes spatial geometry, stratigraphic age, rock properties, tectonic elements, geophysical field, geochemical field and thermal event age labels. Dynamic geological and mineral map compilation: With the support of the multi-dimensional data base, based on the time slice or tectonic-thermal event period set by the user, the stratigraphic boundaries, magmatic bodies, tectonic features and distribution information of known mineral deposits or mineralization points within the corresponding spatiotemporal range are dynamically extracted and rendered. Through the map section and attribute filtering functions of the three-dimensional geological modeling software, dynamic geological and mineral maps reflecting the geological and mineral features of a specific geological historical period are automatically generated. Interpretation and generation of tectonic-thermal event sequence diagrams: Based on the multi-stage isotopic dating data and tectonic deformation data in the multi-dimensional data base, the spatiotemporal analysis module of the 3D geological modeling software is used to identify the concentrated intervals of age data using density clustering algorithm to identify the main thermal event stages. Combined with the spatial intersection relationship of tectonic deformation patterns, the sequence of their order is determined, and a tectonic-thermal event sequence diagram is automatically drawn with time as the vertical axis, which intuitively displays the age range, tectonic properties, and spatial influence range of each thermal event stage. Multi-phase tectonic-metallogenic event correlation analysis and basic map synthesis: The dynamic geological and mineral map is correlated with the tectonic-thermal event sequence map. In the multi-dimensional data base, through spatial overlay and attribute linking, the causal relationship between specific deposits or mineralization points and specific tectonic-thermal event periods is identified. A gold deposit prediction basic map that simultaneously carries spatial geological and mineral information and temporal evolution sequence information is synthesized. This map supports dynamic switching of geological scenes based on time axis sliding or event period selection and the overlay display of multi-phase tectonic-metallogenic events.

[0020] In one embodiment, the reconstruction process of the construction-thermal event sequence is as follows: Paleotectonic framework restoration: The study area is placed in the context of the tectonic unit to which it belongs. Using plate reconstruction models, the paleogeographic location, plate movement direction and rate change trajectory after the orogenic period of the study area are restored, and a regional tectonic dynamic background evolution framework is established. Event spatiotemporal visualization and clustering: Isotopic age data of the study area are visualized in a 3D geological modeling environment according to spatial location, rock unit, and measured mineral type. Spatiotemporal sequence matching algorithm and cluster analysis method are used to identify the concentrated distribution intervals (peak periods) of age data in the time domain. These concentrated distribution intervals represent important magmatic, metamorphic, or hydrothermal events. Combined with the period division of tectonic deformation data, the concentrated distribution intervals are associated with specific tectonic movements (such as collision compression, post-collision extension, and large-scale strike-slip), defining clear tectonic-thermal events and arranging them in chronological order to form a preliminary tectonic-thermal event sequence.

[0021] It should be noted that the specific analysis process of event spatiotemporal visualization and clustering is as follows: Age Database Construction and Preprocessing: A dedicated isotopic age database will be established. All multi-period isotopic dating data extracted from the multidimensional data base (including but not limited to zircon U-Pb ages, mica 40Ar-39Ar ages, and monazite or sphene U-Th-Pb ages) will be structurally entered. Each record must include: age value, ±1σ or ±2σ error range, the mineral being measured, the corresponding lithosphere, the geographic coordinates of the sampling point, and its tectonic location. Data preprocessing will include removing obviously unreasonable or excessively erroneous data and converting all age values ​​to the same time unit (e.g., millions of years, Ma). Age peak period identification based on kernel density estimation: The probability density distribution of the preprocessed age dataset of the entire study area is calculated in the time domain. Specifically, Gaussian kernel and bandwidth parameters are selected, and the probability density value of dating data at each time point (t) on the continuous time axis is calculated by kernel density estimation, thereby generating a smooth age probability density curve. The obvious peaks on the age probability density curve, i.e. the probability density peaks, represent the concentrated distribution intervals of age data in the time domain in a statistical sense. These intervals are initially identified as the peak periods of potential magmatic, metamorphic, or hydrothermal events. Each peak period is defined by its peak age and a certain width. Structural deformation data is periodized and spatially correlated: In parallel, structural deformation measurement data is divided into periods and spatially coded. Based on field observations and microstructural analysis, different types of structural deformation (such as penetrating foliation, ductile shear bands, conjugate joints, and fault striations) are classified into structural periods with a relative chronological order according to their superposition and cutting relationships. In GIS or 3D modeling software, each period of structural deformation data (such as shear band strike lines and foliation poles) is associated with its inferred structural motion type (such as collision compression, post-collision extension, and large-scale strike-slip) and assigned a preliminary, relative temporal code (such as D1, D2, D3…), forming a structural deformation period-spatial distribution layer. Spatiotemporal sequence matching algorithm to associate peak periods with tectonic movements: The identified temporal peak periods are causally correlated with the divided spatial tectonic periods. The spatiotemporal sequence matching algorithm is implemented, specifically: In the three-dimensional geological model, the spatiotemporal sequence matching algorithm searches and verifies whether the age data of a specific peak period (and the rock units to which it belongs, such as granite bodies and metamorphic strata) have a significant coexistence or control relationship with the tectonic deformation data of a certain period (and the tectonic movements it represents) in terms of spatial distribution. For example, if an age probability density peak of ~420 Ma is identified, and this peak mainly comes from zircon U-Pb and mica 40Ar-39Ar ages of mylonitized granites distributed within a regional ductile shear zone, and this shear zone is identified as a product of a contemporaneous compressional environment, the spatiotemporal sequence matching algorithm strongly correlates this time peak period (~420 Ma) with the collisional compressional tectonic movement and the corresponding tectonic deformation period (such as D1) through spatial overlay analysis and statistical testing; conversely, if a peak of ~130 Ma mainly comes from a series of undeformed porphyry bodies occurring in the form of stocks, and its spatial distribution is controlled by a set of high-angle normal faults, the algorithm will associate it with post-collisional extensional tectonic movements; Tectonic-thermal event definition and sequence generation: Based on the above matching results, each age peak period successfully associated with a specific tectonic movement is formally defined as a tectonic-thermal event. Each tectonic-thermal event must clearly record its peak age, age range, dominant tectonic movement nature, and main geological manifestations. Finally, all defined tectonic-thermal events are arranged from oldest to newest according to their peak age, and the relative time intervals between them are marked, thereby generating a tectonic-thermal event sequence for the study area.

[0022] In one embodiment, step four, classifying the ore deposit series, specifically involves: classifying all known ore deposit points in the study area using a hierarchical clustering algorithm based on the genetic type of the ore deposit, the host rocks, the alteration mineral assemblage, and the spatiotemporal correlation with specific tectonic-thermal events, forming a genetic-stage coupled classification system such as "Early Paleozoic orogenic gold deposit series" or "Mesozoic porphyry-skarn copper-molybdenum-gold deposit series".

[0023] More specifically, based on a multidimensional data foundation and a tectonic-thermal event sequence, the attribute characteristics of all known mineral deposits (points) in the study area are extracted in a structured manner and quantified and encoded. The attribute features include at least: spatial coordinates, metallogenic epoch (isotopic age data), host rock type, ore mineral assemblage, hydrothermal alteration type, tectonic-thermal event sequence, and spatial relationship with specific tectonic deformation patterns. The above attribute features are converted into numerical or categorical feature vectors that can be processed by clustering algorithms; Using the generated ore deposit feature vector as input, and Euclidean distance or Jaccard distance as the similarity metric, the feature similarity between different ore deposits (points) is calculated. By generating a tree diagram of mineral deposit phylogenetics through systematic clustering, and based on a preset similarity threshold or the natural breakpoints of the tree diagram, the mineral deposits (points) in the study area are initially divided into several mineral deposit phylogenetics with similar genesis and spatiotemporal attributes.

[0024] In one embodiment, the specific steps for constructing a dynamic metallogenic model with superimposed multi-stage ore-forming elements are: dividing the ore deposit series. Within each of the aforementioned ore deposit series, based on the ore deposit scale, the degree of geological research, the completeness of data, and the representativeness of spatial distribution, several (no fewer than two) ore deposits are selected as representative ore deposit combinations for that series. At the same time, the spatial clustering characteristics of ore deposit points within the same series are analyzed using a DBSCAN-based spatial clustering algorithm to ensure that the selected combinations can reflect the dominant spatial distribution patterns and ore-controlling tectonic background of the ore deposits in that series under the influence of major tectonic-thermal events. Multi-mineral-multi-method isotopic cross-dating and age determination: For selected representative deposit assemblages, mineral samples from different generations are collected for dating, including but not limited to: zircon from magmatic rocks closely related to mineralization, hydrothermal zircon or sphene from hydrothermal veins, metallic sulfides (such as molybdenite and pyrite), and hydrothermal alteration minerals (such as mica and potassium feldspar). By comparing and analyzing the dating results of different minerals within the same deposit, cross-validation is performed to determine the main mineralization period and identify possible later hydrothermal alteration periods or tectonic thermal disturbance events. More specifically, the uranium-lead age of zircon / hydrothermal zircon is determined by LA-ICP-MS U-Pb method to constrain the age of diagenesis or hydrothermal activity, the rhenium-osmium age of molybdenite is determined by Re-Os method to directly obtain the mineralization age, and the argon-argon age of mica or potassium feldspar is determined by 40Ar-39Ar method to record the hydrothermal alteration or cooling age. Constructing a dynamic metallogenic model with superimposed metallogenic elements across multiple phases: Integrating ore deposit series, spatial distribution characteristics, and metallogenic and alteration time sequences, and for each ore deposit series, constructing a flowchart of superimposed metallogenic elements across multiple phases in chronological order; This flowchart dynamically displays the sequential relationships of the following elements: Dominant tectonic-thermal events, accompanying magmatic activity sequences, fluid property evolution, changes in mineralization types and elemental combinations, and superimposed wall rock alteration combinations; This model clearly identifies the key ore-controlling elements and ore-forming products during the main mineralization period, and visualizes the superposition, modification, destruction, or re-enrichment processes of the early ore bodies during the later alteration period.

[0025] In one embodiment, in step five, machine learning image segmentation technology is used to identify and distinguish remote sensing alteration anomalies of different periods. This means that the U-Net convolutional neural network model is used to train multispectral or hyperspectral remote sensing images to extract silicification and sericitization anomalies related to early deep-seated hydrothermal activity and mudstone and claystone anomalies related to late-stage shallow-seated activity. In one embodiment, step six involves training a nonlinear gold mineralization intensity prediction model using a machine learning algorithm, specifically as follows: Using known ore deposits and mineralization points as label samples, and phased anomaly identification markers as input features, the model is trained using XGBoost or support vector regression algorithms, and the output is a mineralization favorability index, which is a continuous variable.

[0026] In one embodiment, in step seven, the target area is classified into at least three levels: high-confidence deep target area, mid-shallow verification target area, and prospective exploration area, based on the complexity of the mineralization system, the intensity and scale of multiple superimposed anomalies, and the three-dimensional spatial resource potential estimation results.

[0027] Differential validation of the target area: For high-confidence deep target areas, priority should be given to deploying large-scale integrated geophysical exploration and structural geochemical profiles, and verification should be carried out in conjunction with drilling. For the shallow and medium-depth verification target area, high-precision remote sensing alteration mapping and surface engineering exposure were used for verification. All new data obtained from the verification were fed back in real time to optimize and update the tectonic-metallogenic event sequence model, ore deposit series classification and nonlinear prediction model parameters.

[0028] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for rapidly delineating gold prospecting target areas, characterized in that, Includes the following steps: Step 1: Construction of the Geological Feature Database of Mineral Deposits: Obtain and structure the geological features of all known gold deposits and mineral occurrences in the study area to form a feature database; Step 2: Integration of multi-source heterogeneous data and dynamic compilation of basic maps: Integrate geological maps, multi-period isotope dating data, tectonic deformation measurement data, and original geophysical, geochemical, remote sensing, and heavy mineral data of the study area to construct a multi-dimensional data base, compile dynamic geological and mineral maps and tectonic-thermal event sequence maps, and form basic maps for gold deposit prediction; Step 3: Reconstruction of tectonic-metallogenic event sequence and dynamic classification of metallogenic systems: Based on the multidimensional data foundation, the sequence of multiple tectonic-thermal events after the orogenic period in the study area is reconstructed, and dynamic metallogenic system categories are classified accordingly. The spatial superposition and transformation relationship of metallogenic systems at different stages is analyzed. Step 4: Construction of regional ore deposit series and multi-method collaborative dating of representative ore deposit combinations: Based on the tectonic-metallogenic event sequence and ore deposit geological characteristics, a clustering algorithm is used to divide the ore deposit series, select representative ore deposit combinations of each series, and use multiple isotope dating techniques for cross-dating to determine the main metallogenic period and the later transformation period, and construct a dynamic metallogenic model with superimposed metallogenic elements of multiple periods. Step 5: Analysis of multiple phases of anomaly superposition and extraction of necessary conditions for dynamic metallogenic geology: Based on the ore deposit series and the dynamic metallogenic model, machine learning image segmentation technology is used to identify and distinguish remote sensing alteration anomalies of different phases. Combined with geochemical element combinations and geophysical anomaly characteristics, phased anomaly identification markers are established, and necessary conditions for dynamic metallogenic geology are extracted. Step Six: Dynamic Element Combination Analysis and Nonlinear Integrated Prediction Modeling: Under the constraints of the tectonic-thermal event sequence, the geological-tectonic domains are divided and the exclusive element combinations of different periods of ore deposits in each geological-tectonic domain are identified. A nonlinear gold mineralization intensity prediction model is trained using machine learning algorithms to generate a continuous mineralization favorability index. Step 7: Target area classification and dynamic verification: Based on the mineralization favorability index and the necessary dynamic mineralization geological conditions, the target area is delineated and classified and predicted using three-dimensional geological modeling technology.

2. The method for rapidly delineating gold prospecting target areas according to claim 1, characterized in that, The formation process of the gold mine prediction base map is as follows: Multi-source heterogeneous data preprocessing: performing coordinate system standardization, format standardization, and quality checks on multi-source heterogeneous data; Construction of a multidimensional data base: The pre-processed multi-source heterogeneous data is imported into the 3D geological modeling software. Based on geographic coordinates and stratigraphic age, spatial registration and temporal attribute association are performed to establish a multidimensional data base that includes spatial geometry, stratigraphic age, rock properties, tectonic elements, geophysical field, geochemical field and thermal event age labels. Dynamic geological and mineral map compilation: With the support of the multi-dimensional data base, based on the time slice or tectonic-thermal event period set by the user, the stratigraphic boundaries, magmatic rock bodies, tectonic features and known mineral deposit distribution information within the corresponding spatiotemporal range are dynamically extracted and rendered to generate a dynamic geological and mineral map that reflects the geological and mineral features of a specific geological historical period. Interpretation and generation of tectonic-thermal event sequence diagram: Based on the multi-period isotopic dating data and tectonic deformation data in the multidimensional data base, density clustering algorithm is used to identify the concentrated intervals of age data to mark the main thermal event periods, and the spatial intersection relationship of tectonic deformation patterns is combined to determine their sequence, and a tectonic-thermal event sequence diagram with time as the vertical axis is drawn. Multi-phase tectonic-metallogenic event correlation analysis and basic map synthesis: The dynamic geological and mineral map is correlated with the tectonic-thermal event sequence map. Through spatial overlay and attribute linking in the multi-dimensional data base, the causal relationship between specific deposits and specific tectonic-thermal event periods is identified, and basic maps for gold deposit prediction are synthesized.

3. The method for rapidly delineating gold prospecting target areas according to claim 1, characterized in that, The reconstruction process of the construction-thermal event sequence is as follows: Paleotectonic framework restoration: The study area is placed in the context of the tectonic unit to which it belongs. Using plate reconstruction models, the paleogeographic location, plate movement direction and rate change trajectory after the orogenic period of the study area are restored, and a regional tectonic dynamic background evolution framework is established. Event spatiotemporal visualization and clustering: Isotopic age data of the study area are visualized in a three-dimensional geological modeling environment according to spatial location, rock unit and measured mineral type. Spatiotemporal sequence matching algorithm and cluster analysis method are used to identify the concentrated distribution intervals of age data in the time domain. Combined with the period division of tectonic deformation data, the concentrated distribution intervals are associated with specific tectonic movements, tectonic-thermal events are defined, and they are arranged in chronological order to form a preliminary tectonic-thermal event sequence.

4. The method for rapidly delineating gold prospecting target areas according to claim 1, characterized in that, In step four, the classification of ore deposit series specifically involves: based on the genetic type of the ore deposit, the host rocks, the alteration mineral assemblage, and the spatiotemporal correlation with specific tectonic-thermal events, a hierarchical clustering algorithm is used to classify all known ore deposit points in the study area, forming a genetic-stage coupled classification system.

5. The method for rapidly delineating gold prospecting target areas according to claim 4, characterized in that, The specific steps for constructing a dynamic metallogenic model with superimposed multi-stage metallogenic elements are as follows: dividing the ore deposit series. Within each of the aforementioned mineral deposit series, several mineral deposits are selected as a representative combination of mineral deposits for that series; Multi-mineral-multi-method isotopic cross-dating and age determination: For a selected representative deposit assemblage, mineral samples from different generations are collected for dating. By comparing and analyzing the dating results of different minerals within the same deposit, cross-validation is performed to determine the main mineralization period and identify the later hydrothermal alteration period or tectonic thermal disturbance events. Constructing a dynamic metallogenic model with superimposed metallogenic elements across multiple phases: Integrating ore deposit series, spatial distribution characteristics, and the time sequence of metallogenesis and alteration, and for each ore deposit series, constructing a flowchart of superimposed metallogenic elements across multiple phases in chronological order.

6. The method for rapidly delineating gold prospecting target areas according to claim 1, characterized in that, In step five, machine learning image segmentation technology is used to identify and distinguish remote sensing alteration anomalies of different periods. This means that the U-Net convolutional neural network model is used to train multispectral or hyperspectral remote sensing images to extract silicification and sericitization anomalies related to early deep-seated hydrothermal activity and mudstone and clay alteration anomalies related to late-stage shallow-seated activity.

7. The method for rapidly delineating gold prospecting target areas according to claim 1, characterized in that, In step six, a nonlinear gold mineralization intensity prediction model is trained using a machine learning algorithm, specifically as follows: Using known ore deposits and mineralization points as label samples, and phased anomaly identification markers as input features, the model is trained using XGBoost or support vector regression algorithms, and the output is a mineralization favorability index, which is a continuous variable.

8. The method for rapidly delineating gold prospecting target areas according to claim 1, characterized in that, In step seven, the target area is classified into at least three levels based on the complexity of the mineralization system, the intensity and scale of multiple superimposed anomalies, and the three-dimensional spatial resource potential estimation results: high-confidence deep target area, medium-shallow verification target area, and prospective exploration area.

Images