Big data intelligent prospecting operation method based on multi-source geological relationship data

Abstract

The application provides a big data intelligent ore prospecting operation method based on multi-source geological relationship data, applies artificial intelligence technology, big data technology and geological theory research achievements to the field of mineral exploration technology, collects or constructs detailed knowledge graphs of single ore type or multiple ore types according to different situations, divides ore prospecting target areas into a plurality of unit modules with spatial information as the most basic geological relationship knowledge graph carrier unit, respectively assigns corresponding attribute contents to the unit modules according to spatial positions, description contents and geological knowledge, references existing related ore type knowledge graphs, performs knowledge fusion and disambiguation on the attribute contents in the unit modules, and constructs the geological relationship knowledge graph of the unit module, and obtains the scores of the unit modules or the scores of the different types of ore-forming possibilities of the unit modules according to weights, so that the ore prospecting prospective areas can be efficiently and accurately delineated or the dominant ore can be determined.

Description

Technical Field

[0001] This invention belongs to the field of big data intelligent mineral exploration technology, specifically relating to a big data intelligent mineral exploration operation method based on multi-source geological relationship data. Background Technology

[0002] Mineral resources are a crucial material foundation for economic and social development. In recent years, with the rapid development and widespread application of artificial intelligence and big data technologies, big data-driven intelligent mineral exploration methods—the integration of artificial intelligence and big data technologies with traditional mineral exploration techniques—have gradually become one of the most active technological frontiers in the field of mineral exploration, yielding many valuable research results. For example, the invention patent with authorization announcement number CN117557401B discloses an intelligent delineation method for uranium ore prospecting target areas based on geological big data. The main technical contents of this method are as follows:

[0003] Collect different logging curves corresponding to different mines in the past and present, and obtain historical logging data and current logging data based on the different logging curves;

[0004] A preset data-driven model is constructed. Historical logging data and current logging data are input into the preset data-driven model to obtain historical geological feature matrices and current geological feature matrices. The historical geological feature matrices and current geological feature matrices are matched to obtain approximate geological feature data of the uranium exploration target area to be delineated.

[0005] Based on the geological feature data, a three-dimensional histogram is constructed, and the matching degree of nodes in the historical geological feature matrix and the current geological feature matrix and the local matching degree between individual boreholes are calculated to obtain the best matching result of the current geological feature data.

[0006] Calculate the local approximation between the historical geological feature matrix and the current geological feature matrix, and then establish a rectangular coordinate system with the first target data value as the origin. Based on the tuples of the rectangular coordinate system, obtain the slope and distance values, and obtain the distribution of adjacent boreholes in the rectangular coordinate system based on the slope and distance values.

[0007] Obtain the coordinates of adjacent boreholes in the first target data value in the rectangular coordinate system and the minimum distance between adjacent boreholes in the second target data value. Calculate the mine matching degree of the minimum distance value and the distance from the borehole to the origin of the minimum distance value in the second target data value to obtain the local matching degree between the first and second completely undirected graphs.

[0008] Based on the best matching result, the data distribution of the current geological feature data is obtained; based on the data distribution, the hit ratio of the same geological feature data is calculated; and based on the hit ratio, the intelligent delineation of uranium exploration target areas is completed.

[0009] Graph structure data is constructed from the historical geological feature matrix and the current geological feature matrix to obtain a first completely undirected graph and a second completely undirected graph. In the first completely undirected graph and the second completely undirected graph, a single borehole is a node, and the edge weight between nodes is a binary tuple of the direction and distance values ​​between the corresponding boreholes.

[0010] The approximation of the first and second completely undirected graphs is matched to obtain the first target data value in the historical geological feature matrix corresponding to the first completely undirected graph and the second target data value in the current geological feature matrix corresponding to the second completely undirected graph. The first target data value and the second target data value are ANDed to obtain a matching degree matrix with the same number.

[0011] The matching degree of the first and second completely undirected graphs is obtained by the ratio of the sum of all values ​​in the matching degree matrix to the number of all data in the first or second target data value matrix; based on the matching degree, the approximate geological feature data of the uranium prospecting target area to be delineated is obtained.

[0012] For example, patent application CN115907151A discloses an intelligent mineralization prediction method based on geological big data. The main technical contents of this method include the following steps:

[0013] Step S1, Selection of mineral exploration prediction factors

[0014] Based on geological surveys and literature reviews of the study area, three levels of predictive factors for gold exploration in the region were summarized and used as feature data for machine learning. The first-level factors include geological, geochemical, and geophysical factors. The secondary factors corresponding to the geological factors include stratigraphy, tectonics, and igneous rocks; the secondary factors corresponding to the geochemical factors include 1:50,000 stream sedimentary geochemical data; and the secondary factors corresponding to the geophysical factors include 1:50,000 high-precision magnetic geophysical data. The tertiary factors corresponding to the stratigraphy include the Xinghongpu Formation and the Gudaoling Formation, among others. The tertiary factors corresponding to the tectonics include NW-trending brittle-ductile fractures, NE-trending brittle fractures, fracture intersections, and anticlines. The tertiary factors corresponding to the igneous rocks include the distance to the Huahongshuping pluton. The tertiary factors corresponding to the 1:50,000 stream sedimentary geochemical data include the Au-related factor F5 obtained from Au element geochemistry and R-type factor analysis. The tertiary factors corresponding to the 1:50,000 high-precision magnetic geophysical data include the regionalized polar magnetic susceptibility ΔT.

[0015] Step S2, prediction unit partitioning

[0016] Based on a survey of known mineral deposits (points), the shortest distance between each gold deposit (point) was found to be 170m. Furthermore, according to the theory of Carranza et al., 2009, the minimum scale of a prediction unit at a scale of 1:50,000 is 75m × 75m. Therefore, this study divided the study area into prediction units of 100m × 100m, covering a total area of ​​500.83 km². 2 A total of 50,083 prediction units were divided;

[0017] Step S3: Quantitative characterization and data fusion of predictors

[0018] GIS technology was used to quantitatively characterize each predictor, obtain numerical data of various predictors corresponding to all predictor units in the study area, and fuse the data to construct a dataset.

[0019] Step S4, Construction of the sample set

[0020] Prediction units containing gold deposits (points) are used as positive samples, and prediction units containing non-gold deposits (lead-zinc mines, copper mines, etc.) are used as negative samples. The positive and negative samples together form the sample set, which ensures the accuracy and representativeness of the modeling data. 70% of the sample set is randomly divided into training set, and the remaining samples are used as test set.

[0021] Step S5: Training and tuning the machine learning prediction model

[0022] Five algorithms from machine learning were selected for modeling: Random Forest (RF), Support Vector Machine (SVM), Gradient Boosting Decision Tree (GBDT), Naive Bayes (NB), and Logistic Regression (LR). The models were trained using training set samples, and the hyperparameters of the prediction models were tuned using methods such as random search and grid search.

[0023] Step S6: Quantitative evaluation and optimization of machine learning prediction models

[0024] The model performance was quantitatively evaluated using test set samples. The evaluation metrics included accuracy, precision, recall, F1 score, and AUC score. Based on the quantitative evaluation of the above five prediction models, the random forest model was selected as the prediction model for gold mineralization in the study area. The random forest model achieved an accuracy of over 95% on both the training and test sets, indicating that the model has excellent predictive ability.

[0025] Step S7, Calculation of mineralization probability in the study area

[0026] Using a pre-trained random forest model, the probability of gold mineralization in 50,083 prediction units in the study area was calculated, and the mineralization probability of the study area was displayed using GIS technology.

[0027] Step S8, Delineation of the mineralization target area

[0028] Based on the "success rate curve" method proposed by Rodriguez (2015), the study area was divided into a gold mineralization area with extremely high potential, a high potential, and a general potential. The extremely high potential and high potential areas can be used as target areas for gold exploration in the area.

[0029] Besides the aforementioned invention patents or patent applications, there are many other research and development achievements related to big data intelligent mineral exploration methods, which will not be elaborated upon here. In general, compared with traditional mineral exploration technologies, existing big data intelligent mineral exploration methods have significantly improved the efficiency of mineral exploration work. However, existing big data intelligent mineral exploration methods, including the technical solutions described in invention patents CN117557401B and CN115907151A, also have some shortcomings. These mainly lie in the insufficient coupling between the ore-forming information reflected in multi-source geological data and the geological relationships in the field, as well as the inadequate full exploration, rational selection, and application of various ore-forming models and basic geological research results, especially the insufficient emphasis on relevant ore-forming theoretical research results. Furthermore, existing big data intelligent mineral exploration methods are only applicable to specific mineral types and ore-forming types, and have limited application value in finding unknown ore deposit types, thus exhibiting significant limitations.

[0030] Currently, my country has accumulated abundant geological data, and there are numerous research results and literature on metallogenic models for various mineral deposits. Basic geological research related to mineralization is also relatively comprehensive. Faced with massive amounts of geological and mineral information, traditional mineral exploration techniques struggle to cope, while the application of artificial intelligence and big data technologies can effectively compensate for the shortcomings of traditional mineral exploration techniques in information processing. However, in the process of applying artificial intelligence and big data technologies, neglecting relevant geological theoretical research results—such as metallogenic theoretical research results represented in the form of geological knowledge maps—can lead to blind spots in the work and even cause significant deviations in the final results, such as the delineated prospective exploration areas. Deviations in prospective exploration areas often result in unnecessary investment and ineffective work.

[0031] It is important to emphasize that while geological theoretical research findings, including metallogenic theories, need continuous practical testing and refinement, and may even be superseded by new theoretical research, existing geological theoretical research findings play a crucial guiding role, both in uncovering key information about deep mineralization and in understanding the interrelationships between seemingly unrelated deep geological information. How to fully and effectively utilize existing geological theoretical research findings while employing artificial intelligence and big data technologies is a pressing technical issue that needs to be addressed in the field of mineral exploration. Summary of the Invention

[0032] The purpose of this invention is to apply artificial intelligence technology, big data technology, and related geological theories (especially metallogenic theories) to the field of mineral exploration technology, so as to fully explore the metallogenic geological background information and deep-level key metallogenic information. It can not only organically integrate mineral exploration information from various disciplines horizontally to reduce human interference, but also link seemingly unrelated deep-level geological information vertically to uncover more mineral exploration clues. It can also reduce unnecessary investment and ineffective work, thereby significantly improving mineral exploration efficiency.

[0033] To achieve the above objectives, the present invention employs the following technical solution:

[0034] A big data intelligent mineral exploration method based on multi-source geological relationship data includes the following steps:

[0035] S01: Based on the project's objectives and tasks, determine the target mineral exploration area and identify the target mineral type and category;

[0036] S02: Divided into two categories, a and b: Category a is where the target mineral type and type have been identified, and the prospective exploration area for that type of mineral can be delineated; Category b is where the target mineral type and / or type are uncertain, and the distribution range of different advantageous target mineral types and / or types within the target exploration area needs to be delineated.

[0037] S03: Collect relevant geological data for the target area, including published papers, monographs, research reports, intelligence information, and other relevant data in geology, geophysics, geochemistry, and remote sensing. Process the data according to two categories, a and b. For category a, collect or construct a detailed knowledge map of a single mineral type; for category b, collect or construct a detailed knowledge map of multiple mineral types.

[0038] S04: Divide the mineral exploration target area into several unit modules with spatial information, which serve as the most basic geological relationship knowledge map carrier units;

[0039] S05: Based on spatial location, overlap the planar map with spatial information with the unit module, and assign geological features and geological relationship attributes to the corresponding unit module;

[0040] S06: Based on the description, combine the reports, attached figures, tables, etc. that correspond to the attributes within the unit module to uncover deeper geological information that cannot be reflected on the plan view and assign it to the unit module as the corresponding attribute content.

[0041] S07: Based on geological knowledge, extract specific metallogenic geological information related to a specific mineral deposit from published scientific papers, monographs, research reports, intelligence information and other relevant materials, and assign it as the corresponding attribute content to the unit module;

[0042] S08: Referring to existing knowledge graphs of relevant mineral types, knowledge fusion and disambiguation are performed on the attribute content in the unit module to ensure consistency of description content. A geological relationship knowledge graph of the unit module is constructed, and each triple (entity-relationship-entity) is assigned a weight. The basic principle of weight assignment is that direct mineral exploration indicators that are closely related to mineral exploration have a high weight, while indirect mineral exploration clues have a low weight.

[0043] S09: Divided into two categories, a and b: Category a involves calculating the similarity between the geological relationship knowledge graph and the mineral exploration knowledge graph of the target type, and obtaining the score for each unit module based on the weight; Category b involves calculating the similarity between the geological relationship knowledge graph and the mineral exploration knowledge graph of the series of types, and obtaining the mineralization probability score for different types of minerals in each unit module based on the weight.

[0044] S10: Divided into two categories, a and b: Category a determines the minimum score for potential mineralization based on the different scores of unit modules and with reference to metallogenic geological theories, and delineates the prospective exploration area for the target type of mineral; Category b determines the distribution range of each dominant type of mineral in the target exploration area based on the different scores of unit modules among different types of minerals.

[0045] Based on the above technical solutions, the present invention may employ the following additional technical means to better or more specifically solve the technical problems to be solved by the present invention:

[0046] The geological features and geological relationship attributes include the spatiotemporal distribution characteristics of geological bodies, structures, alteration, geophysical and geochemical anomalies, and remote sensing anomalies at the unit module, the relationships between them, the distance to each feature point, whether there is feature superposition, and various geological phenomena that have been observed in the field.

[0047] Furthermore, the deep geological information includes the basic geological characteristics, genesis, evolution, influencing factors, and macroscopic and microscopic relationships of specific geological elements.

[0048] Furthermore, the specific geological elements include geological elements such as rocks, minerals, strata, structures, geological ages, alteration, geophysics, geochemistry, and remote sensing.

[0049] Furthermore, when performing step S02, if it meets the condition of category b, that is, when the target mineral type and / or type is uncertain, category b is first selected to determine the mineral exploration target of the dominant type in the mineral exploration target area, and then category a is selected to delineate the prospective area of ​​the dominant type of mineral for the determined dominant type of mineral.

[0050] Furthermore, referring to the geological relationship knowledge graph, the detailed knowledge graph of a single mineral type collected or constructed in step S03, or the detailed knowledge graph of multiple mineral types collected or constructed in step S03, is modified and improved to have as many identical ternary units as possible with the geological relationship knowledge graph.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] This invention applies artificial intelligence, big data, and related geological theories (especially metallogenic theories) to the field of mineral exploration. For different situations, it collects or constructs detailed knowledge graphs for single or multiple mineral types, dividing the prospecting target area into several spatially informational unit modules as the most basic geological relationship knowledge graph carrier units. Based on this, corresponding attribute content is assigned to the unit modules according to spatial location, descriptive content, and geological knowledge. Referring to existing knowledge graphs of related mineral types, knowledge fusion and disambiguation are performed on the attribute content in the unit modules to ensure consistency of descriptive content, and a geological relationship knowledge graph for each unit module is constructed, assigning weights to each triple. Based on the weights, the score of each unit module or the metallogenic probability score of different types of minerals in each unit module is obtained, thereby efficiently and accurately delineating prospective areas or identifying dominant mineral types. Attached Figure Description

[0053] Figure 1 This is a flowchart of the big data intelligent mineral exploration operation method based on multi-source geological relationship data of the present invention;

[0054] Figure 2 It is a structural diagram of a detailed knowledge map of porphyry copper deposits;

[0055] Figure 3 This is a schematic diagram of the unit module division;

[0056] Figure 4 This is a schematic diagram of a geological body after it has been assigned unit modules;

[0057] Figure 5 This is a schematic diagram showing the geochemical anomalies assigned to the unit module;

[0058] Figure 6 This is a schematic diagram of the technical route for extracting the common rare earth element characteristics of porphyry copper deposits from literature.

[0059] Figure 7 It is a schematic diagram of geological relationship knowledge map. Detailed Implementation

[0060] To facilitate a thorough understanding of the technical solution and working principle of the present invention by those skilled in the art, two embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0061] Example 1

[0062] like Figure 1 As shown (this embodiment belongs to) Figure 1 (Case a) A big data intelligent mineral exploration operation method based on multi-source geological relationship data, comprising the following steps:

[0063] S01: Based on the project's objectives and tasks, the target mineral exploration area is determined to be the Xiongmei area of ​​Tibet (longitude: 89°01′50″-89°08′40″; latitude: 31°23′00″-31°28′00″). Based on existing geological data, the target mineral is preliminarily identified as copper, and its type is porphyry.

[0064] S02: The target mineral type and type has been identified as porphyry copper deposit. The purpose is to specifically delineate the prospective exploration area for this type of mineral in the Xiongmei area of ​​Tibet.

[0065] S03: Collect relevant geological data for the target area, the Xiongmei region of Tibet, and surrounding areas with similar geological backgrounds, including published papers, monographs, research reports, and intelligence information in geology, geophysics, geochemistry, and remote sensing. Collect or construct a detailed knowledge map of porphyry copper deposits. The main contents of this knowledge map are as follows: Figure 2 As shown.

[0066] S04: The prospecting target area is divided into several spatially informational unit modules, serving as the most basic geological relationship knowledge map carrier units. The division of unit modules is closely related to the prospecting target area, target mineral type, and type, and a comprehensive selection is made. Generally, when the prospecting target area is large, the unit module density can be appropriately reduced; when the prospecting target mineral type and type is small-scale or vein-like, the unit module density should be increased. In this embodiment, the prospecting target is a large-scale porphyry deposit, with a target area of ​​approximately 100 km². 2 The unit modules were set to 100m × 100m, and a total of 10341 unit modules were divided. The division of unit modules is as follows: Figure 3 As shown.

[0067] S05: Based on spatial location, overlay planar maps containing spatial information with unit modules, and assign geological features and geological relationship attributes to the corresponding unit modules. The spatial information in this step includes geological bodies, geochemical anomalies, etc. The content displayed after assigning geological bodies to unit modules is as follows: Figure 4 As shown. Figure 4 Different colors in the diagram represent different geological bodies. For example, red represents Late Cretaceous granodiorite porphyry, pink represents Late Cretaceous monzogranite, and white represents contact zones. For other colors representing geological bodies, see [link to relevant documentation]. Figure 4Legend in the diagram. The content displayed after assigning geochemical anomalies to a unit module is as follows: Figure 5 As shown, Figure 5 The dark areas in the diagram represent geochemical anomaly zones.

[0068] S06: Based on the descriptive content, combine the reports, attached maps, and tables corresponding to the attributes within the unit module to uncover deeper geological information that cannot be reflected on the planar maps, and assign it as the corresponding attribute content to the unit module. Regarding the descriptive content in this step, for example, the geological map shows γδπK2 (Late Cretaceous granodiorite porphyry). Apart from its distribution location and contact relationship with surrounding geological elements, no further information can be obtained from the geological map. However, the report corresponding to the geological map contains a detailed description of γδπK2 (Late Cretaceous granodiorite porphyry), including geological and petrographic characteristics, rock formation age, rock geochemical characteristics, rock classification, tectonic setting, and a series of related descriptions. All of this descriptive content can be used as attribute content for the rock mass and assigned to the corresponding unit module.

[0069] S07: Based on geological knowledge, extract specific metallogenic geological information related to porphyry copper deposits from published scientific papers, monographs, research reports, intelligence information, and other relevant materials (e.g., specific trace element and isotopic characteristics of the ore-forming rock body of the porphyry copper deposit), and assign it as corresponding attribute content to the unit module. The basic technical idea of ​​this step is as follows: Figure 6 As shown. Regarding the geological knowledge involved in this step, let's take the academic paper "Discussion on Source Rocks or Source Regions of Porphyry Copper Deposits" published by Rui Zongyao et al. in the *Acta Petrologica Sinica* in 2004 as an example. This paper mentions several common petrogeochemical characteristics of porphyry copper deposits: such as the common REE (rare earth element) characteristics of porphyry copper deposits being "right-cline, relatively enriched in light rare earth elements, relatively depleted in heavy rare earth elements, and with insignificant Eu anomalies." Therefore, porphyry bodies with similar REE characteristics can serve as a clue for mineral exploration, expanding the exploration basis and incorporating some seemingly unrelated indirect exploration evidence into the exploration clues.

[0070] S08: Referring to existing knowledge graphs of porphyry copper deposits, perform knowledge fusion and disambiguation on the attribute content in the unit module to ensure consistency in description. Construct a geological relationship knowledge graph for the unit module and assign weights to each triple (in a triple, i.e., in entity-relationship-entity, the two entities refer to specific objects or concepts that are related through a certain relationship). The basic principle for weight assignment is that direct prospecting indicators closely related to mineral exploration have high weights (e.g., the discovery of malachite in the field is a direct prospecting clue and has a high weight), while indirect prospecting clues have low weights (e.g., granodiorite porphyry has a minor Eu anomaly and has a low weight). Furthermore, disambiguation in this step ensures that identical descriptions convey the same meaning. For example, "enrichment of the ore-forming element copper" and "relative enrichment of copper in the trace elements of the surrounding rock" can both be simply expressed as "enrichment of Cu," but the copper content they refer to is not on the same order of magnitude, leading to ambiguity. The expression needs to be modified to ensure uniqueness.

[0071] S09: Integrate geological relationship knowledge graphs (e.g.) Figure 7 The similarity between the geological relationship knowledge graph shown and the mineral exploration knowledge graph of the target type is calculated, and the score of each unit module is obtained according to the weight.

[0072] S10: Referring to the metallogenic geological theory of porphyry copper deposits, determine the minimum score for possible mineralization, delineate the specific prospective area of ​​porphyry copper deposits in the Xiongmei area of ​​Tibet, and provide a basis for further work.

[0073] It should also be noted that, in this embodiment, the geological features and geological relationship attributes include geological bodies, structures, ages, alterations, and geophysical and geochemical anomalies. The deep geological information includes the geochemical characteristics and structural occurrence of specific rock bodies. Furthermore, in this embodiment, when performing step S03, the detailed knowledge map of porphyry copper deposits collected or constructed in step S03 is modified and improved with reference to the geological relationship knowledge map, so that it shares as many identical ternary units as possible with the geological relationship knowledge map.

[0074] Example 2

[0075] like Figure 1 As shown, this embodiment (belonging to) Figure 1 The basic steps for case b in this embodiment are the same as in Example 1, except that a knowledge graph of multiple mineral types and / or categories needs to be collected or constructed in the target mineral exploration area. Specifically, this embodiment differs from Example 1 in the following four aspects:

[0076] First, in this embodiment, when performing step S02, the target mineral type and / or type of the prospecting is not determined, but the distribution range of different advantageous target mineral types and / or types needs to be delineated within the prospecting target area.

[0077] Second, in this embodiment, during step S03, detailed knowledge maps of various mineral types are collected or constructed. The determination of target mineral types and / or types should be based on research into the geological background of the target area and its surroundings, making the target mineral types and / or types more targeted and avoiding blind mineral exploration. For example, taking the mineral exploration target area mentioned in Embodiment 1 as an example, its fault structures and intrusive rocks are relatively well-developed. Through previous geological work, there are clues for hydrothermal deposits. Therefore, the possible target mineral types and / or types in this area include porphyry type, hydrothermal vein type, skarn type copper, molybdenum, lead, zinc, silver, etc.

[0078] Third, the similarity between the geological relationship knowledge graph and the mineral exploration knowledge graph of the series of mineral types is calculated. Based on the weights, the mineralization probability scores for different types of minerals in each unit module are obtained. For example, in a certain unit module, the score for porphyry copper deposit is 9.8 points, for porphyry molybdenum deposit is 7.6 points, for skarn copper deposit is 6.8 points, for skarn lead-zinc deposit is 6.5 points, and for quartz vein silver deposit is 2.3 points. This indicates that among the target mineral types, porphyry copper deposit has the highest mineralization probability in this module, followed by porphyry molybdenum deposit, skarn copper deposit, and skarn lead-zinc deposit, while quartz vein silver deposit has the lowest mineralization probability. In other unit modules, quartz vein silver deposit may have the highest mineralization probability, while porphyry copper deposit may have the lowest.

[0079] Fourth, based on the different scores of each unit module among different types of minerals, the distribution of each target mineral type or / and type in the prospecting target area is determined, and one or more target mineral types or / and types with the most prospecting potential are identified, providing a basis for further work.

[0080] It should also be noted that, in addition to the four differences mentioned above, in this embodiment, the deep geological information includes the basic geological characteristics, genesis, evolution, and influencing factors of specific geological elements such as rocks, minerals, strata, structures, geological ages, alteration, geophysics, geochemistry, and remote sensing, as well as the relationship between the macroscopic and microscopic levels. In other words, compared with Embodiment 1, this embodiment involves more specific geological elements when performing step S06. Of course, this difference is relative. In fact, when performing step S06 in Embodiment 1, if those skilled in the art deem it necessary, all of the above-mentioned specific geological elements can be comprehensively involved.

Claims

1. A big data intelligent prospecting operation method based on multi-source geological relationship data, characterized in that, Comprising the following steps: S01: According to the project target task, determine the prospecting target area, judge the target mineral and type; S02: Divided into a, b two kinds of cases: a kind of case is to determine the target mineral and type, and the range of the target mineral is determined; b kind of case is not to determine the target mineral or / and type, and the distribution range of different advantage target mineral or / and type in the target area is determined; S03: Collect relevant geological data of the target area, including published papers, monographs, research reports, intelligence information in the aspects of geology, geophysical prospecting, geochemical prospecting, remote sensing, and process according to a, b two kinds of cases respectively, a kind of case collects or constructs detailed knowledge graph of single mineral type; b kind of case collects or constructs detailed knowledge graph of multiple mineral types; S04: Divide the prospecting target area into a plurality of unit modules with spatial information as the most basic geological relationship knowledge graph carrier unit; S05: Based on spatial position, overlap the planar drawing with spatial information and the unit module, and attribute the geological characteristics and geological relationship to the corresponding unit module; S06: Based on description content, combine the reports, drawings and tables corresponding to the attributes in the unit module, and mine the deep geological information that cannot be reflected on the planar drawing, and attribute the corresponding attribute content to the unit module; S07: Based on geological knowledge, mine the specific ore-forming geological information related to specific deposits in the published scientific papers, monographs, research reports and intelligence information, and attribute the corresponding attribute content to the unit module; S08: Refer to the existing related mineral type knowledge graph, fuse and disambiguate the attribute content in the unit module, ensure the consistency of the description content, construct the geological relationship knowledge graph of the unit module, attribute the weight of each triple, and the basic principle of weight attribution is that the direct prospecting marks with close relationship with prospecting have high weight, and the indirect prospecting clues have low weight; S09: Divided into a, b two kinds of cases: a kind of case is to calculate the similarity between the geological relationship knowledge graph and the target type mineral prospecting knowledge graph, and obtain the score of each unit module according to the weight; b kind of case is to calculate the similarity between the geological relationship knowledge graph and the series type mineral prospecting knowledge graph, and obtain the mineralization possibility score of each unit module of different types of minerals according to the weight; S10: Divided into a, b two kinds of cases: a kind of case is to determine the minimum score of possible mineralization according to the score difference of the unit module, and determine the target type mineral prospecting perspective area; b kind of case is to determine the distribution range of each advantage type mineral in the target area according to the score difference of the unit module among different types of minerals; When performing step S02, if it meets the b kind of case, that is, the target mineral or / and type is not determined, first select the b kind of case to determine the advantage type mineral target of the target area, and then select the a kind of case to determine the advantage type mineral prospecting perspective area of the advantage type mineral. Referring to the geological relationship knowledge graph content, the detailed knowledge graph of a single ore type collected or constructed in step S03, or the detailed knowledge graphs of multiple ore types collected or constructed in step S03, are modified and improved to have as many same triple element units as possible with the geological relationship knowledge graph. 2.The method for intelligent prospecting operation of big data based on multi-source geologic relationship data according to claim 1, characterized in that: The geological features and geological relationship attributes include the spatial and temporal distribution characteristics of geological bodies, structures, alterations, geophysical and geochemical exploration anomalies, remote sensing anomalies at the unit module, the relationships therebetween, the distances from each feature point, whether there is feature superposition, and various geological phenomena that have been observed in the field. 3.The method of claim 1, wherein the method is characterized in that: The deep geological information includes the basic geological features, genesis, evolution, and influencing factors of specific geological elements, as well as the connection between macro and micro.

4. The method of claim 3, wherein the method is characterized by: The specific geological elements include geological elements of rock, mineral, stratum, structure, geological era, alteration, geophysics, geochemistry, and remote sensing.

Images