Multi-source heterogeneous geological data fusion method, processing system and device, and storage medium

Abstract

The application discloses a multi-source heterogeneous geological data fusion method and processing system, device and storage medium, and comprises the following steps: defining a semantic constraint framework condition; acquiring a plurality of heterogeneous geological data; analyzing each of the heterogeneous geological data to obtain vector content in each of the heterogeneous geological data; using an artificial intelligence model to extract a plurality of candidate feature data from the vector content obtained in each of the heterogeneous geological data based on the semantic constraint framework condition; performing a similarity judgment fusion step on the candidate feature data to obtain result data; in the similarity judgment fusion step, the similarity degree of any two candidate feature data from different heterogeneous geological data is calculated, and when the similarity degree of the two candidate feature data satisfies a similarity fusion condition, the two candidate feature data are combined into one; and the result data is output and stored. The design breaks the fusion barrier of different modal data and realizes efficient storage of data.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence and big data processing technology, and in particular to a method, processing system, device, and storage medium for fusing multi-source heterogeneous geological data. Background Technology

[0002] As my country's mineral resource exploration expands into deep, concealed, and overgrown areas, and with the rapid penetration of big data and artificial intelligence technologies, geological prospecting is undergoing a profound transformation from traditional "experience-driven" to "data-driven" methods. In the field of geological exploration, the location of ore bodies is strictly controlled by the spatiotemporal coupling relationship between specific ore-hosting strata and ore-guiding / distributing structures. However, the massive amount of geological data accumulated over a long period has resulted in a complex situation where unstructured text (exploration reports), semi-structured graphics (CAD engineering drawings), and structured tables (drilling data) coexist, leading to semantic fragmentation between data and low information utilization. This poses new challenges to the efficient integration and comprehensive utilization of multi-source heterogeneous geological data.

[0003] Due to limitations in geological map drawing standards and storage formats, a significant semantic gap exists between CAD and other graphic data (geometric morphology) and textual descriptions (geological connotations). Without a unified correlation mechanism, geological entities remain isolated across different modal data, and crucial ore-controlling structural information cannot be effectively transferred between text and graphics. This results in poor knowledge correlation and makes it difficult to construct complete metallogenic geological models for in-depth analysis. Summary of the Invention

[0004] This invention aims to address at least one of the technical problems existing in the prior art. To this end, this invention proposes a method, processing system, device, and storage medium for fusing multi-source heterogeneous geological data, breaking down the barriers to fusing different modal data, achieving unified representation of the same object from different modal data, and enabling efficient data storage.

[0005] A multi-source heterogeneous geological data fusion method according to a first aspect of the present invention includes: defining semantic constraint framework conditions, wherein the semantic constraint framework conditions include an entity type set, a relation type set, and an attribute type set, the entity type set including multiple entity names for representing extracted objects, the relation type set including multiple relation features for representing logical connection relationships between extracted objects, and the attribute type set including multiple attribute features for representing descriptive attributes of extracted objects; acquiring multiple heterogeneous geological data, wherein the multiple heterogeneous geological data include at least a combination of text data and graphic data; analyzing each of the heterogeneous geological data to obtain vector content in each heterogeneous geological data, the vector content including text data and graphic data; and utilizing artificial intelligence... The model performs semantic extension based on the semantic constraint framework and extracts multiple candidate feature data from the vector content acquired from various heterogeneous geological data. The candidate feature data includes entity names and at least one of the corresponding relational features and attribute features. A similarity judgment and fusion step is performed on the candidate feature data to obtain the result data. In the similarity judgment and fusion step, the similarity degree between any two candidate feature data from different heterogeneous geological data is calculated. When the similarity degree between two candidate feature data satisfies the similarity fusion condition, the two candidate feature data are merged into one. The similarity degree is derived from at least one of the entity names, relational features, and attribute features in the two candidate feature data. The result data is then output and stored.

[0006] The multi-source heterogeneous geological data fusion method according to embodiments of the present invention has at least the following beneficial effects: This invention presents a multi-source heterogeneous geological data fusion method. First, it acquires and analyzes the vector content from various heterogeneous geological data sets. Geological experts define semantic constraint framework conditions, inputting multiple keywords representing entity names, relational features, and attribute features of the extracted object into entity type sets, relation type sets, and attribute type sets, respectively. This allows the artificial intelligence model to be constrained by keywords from different type sets, extracting candidate feature data. Then, a pairwise similarity judgment fusion step is performed between the candidate feature data extracted from different heterogeneous geological data sets to calculate the similarity degree. When the similarity degree meets the similarity fusion condition, the two candidate feature data sets are merged into one. Finally, the obtained candidate feature data are appropriately fused to form the result data and stored. This design breaks down the fusion barriers between different modalities of data, achieving a unified representation of the same object from different modalities and enabling efficient data storage.

[0007] According to some embodiments of the present invention, the similarity fusion condition includes a similarity fusion threshold and a suspected association threshold, wherein the suspected association threshold is less than the similarity fusion threshold. The similarity judgment fusion step includes: calculating a weighted comprehensive value based on the similarity of corresponding factors in two candidate feature data; when the weighted comprehensive value is greater than or equal to the similarity fusion threshold, merging the two candidate feature data into one; when the weighted comprehensive value is less than the similarity fusion threshold and the weighted comprehensive value is greater than or equal to the suspected association threshold, placing the two candidate feature data in a manual review queue and obtaining a manual review instruction; when the manual review instruction represents the same entity object, merging the two candidate feature data into one; when the manual review instruction represents two independent entity objects, calculating the similarity of the two candidate feature data with other candidate feature data respectively; when the weighted comprehensive value is less than the suspected association threshold, calculating the similarity of the two candidate feature data with other candidate feature data respectively.

[0008] According to some embodiments of the present invention, the step of calculating the weighted comprehensive value based on the similarity weight of corresponding factors in two candidate feature data includes: extracting entity names, relationship features, and attribute features from the candidate feature data; calculating the similarity values ​​and weighted values ​​of multiple corresponding factors respectively; calculating the weight coefficient of each factor based on the weighted value of each factor; and calculating the weighted comprehensive value using the weight coefficient and the similarity value of each factor. The weighted composite value is: ; in, This is the weighted composite value. The similarity value for the entity names. The weighted value of the factors for entity name. The similarity value of factors representing relational characteristics. The weighted values ​​of the factors that define the relationship characteristics. The similarity value of factors representing attribute features. The weighted value of the factors that represent the attribute characteristics.

[0009] According to some embodiments of the present invention, the factors of entity names in candidate features include name similarity, and the similarity judgment fusion step includes: obtaining entity names in two candidate feature data, and calculating the similarity value of the two entity names based on the edit distance algorithm.

[0010] According to some embodiments of the present invention, the factors of relational features among candidate features include coordinate similarity, and the similarity judgment fusion step includes: Obtain the spatial coordinates from two candidate feature data sets, and calculate the Euclidean distance between the two candidate feature data sets as the similarity value of the coordinates; The Euclidean distance between the two candidate feature data is: ; in, Let be the Euclidean distance between two candidate feature data points a and b. The spatial coordinates of candidate feature data point a are ( ). The spatial coordinates of candidate feature data b are ( ); The similarity value of the coordinates of two candidate feature data is: .

[0011] According to some embodiments of the present invention, the factors of the attribute features in the candidate features include directional similarity, and the similarity judgment fusion step includes: obtaining the orientation information in the two candidate feature data, wherein the orientation information includes the inclination angle or orientation range; calculating the cosine similarity based on the orientation deviation value of the orientation information in the two candidate feature data as the similarity value of the directional similarity.

[0012] According to some embodiments of the present invention, each heterogeneous geological data has a fusion priority value. The process of merging two candidate feature data into one includes: performing a heterogeneous incremental step and a homogeneous merging step on the two candidate feature data; in the heterogeneous incremental step, if the type of a feature in one of the candidate feature data does not exist in the other candidate feature data, then the content of that feature is incrementally added to the result data; in the homogeneous merging step, if the type of a feature in the two candidate feature data is the same, then the fusion priority values ​​of the two candidate feature data are compared, and the content of that feature in the candidate feature data with the higher fusion priority value is selected to be added to the result data, while the content of that feature in the candidate feature data with the lower fusion priority value is discarded.

[0013] According to a second aspect of the present invention, a processing system is used to execute the multi-source heterogeneous geological data fusion method disclosed in any of the above embodiments. The processing system includes: a definition module for defining semantic constraint framework conditions; an acquisition module for acquiring multiple heterogeneous geological data; an analysis module for analyzing each of the heterogeneous geological data to obtain vector content in each heterogeneous geological data; an extraction processing module for using an artificial intelligence model to perform semantic expansion based on the semantic constraint framework conditions, and extracting multiple candidate feature data from the vector content acquired from each heterogeneous geological data; a fusion processing module for performing a similarity judgment fusion step on the candidate feature data to obtain result data, wherein in the similarity judgment fusion step, the similarity degree between any two candidate feature data from different heterogeneous geological data is calculated, and when the similarity degree between two candidate feature data satisfies the similarity fusion condition, the two candidate feature data are merged into one; and a storage module for outputting and storing the result data.

[0014] The processing system according to embodiments of the present invention has at least the following beneficial effects: The processing system of this invention applies the multi-source heterogeneous geological data fusion method disclosed in any of the above embodiments, breaks down the fusion barriers of different modal data, realizes unified representation of the same object from different modal data, and achieves efficient data storage.

[0015] According to a third aspect of the present invention, the control device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the multi-source heterogeneous geological data fusion method disclosed in any of the above embodiments.

[0016] According to a fourth aspect of the present invention, a computer-readable storage medium stores a computer program, characterized in that, when executed by a processor, the computer program implements the multi-source heterogeneous geological data fusion method disclosed in any of the above embodiments.

[0017] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0018] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is a first flowchart of one embodiment of the multi-source heterogeneous geological data fusion method of the present invention; Figure 2 This is a second flowchart of one embodiment of the multi-source heterogeneous geological data fusion method of the present invention; Figure 3 This is a schematic diagram of the principle structure of one embodiment of the processing system of the present invention.

[0019] Figure label: Definition module 310; Acquisition module 320; Analysis module 330; Extraction and processing module 340; Fusion and processing module 350; Storage module 360. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0021] It should be noted that although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first," "second," etc., in the specification, claims, and the aforementioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0022] In the description of this invention, "several" means one or more, "multiple" means two or more, "greater than", "less than", "exceeding" etc. are understood to exclude the number itself, and "above", "below", "within" etc. are understood to include the number itself.

[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0024] like Figure 1 As shown, the multi-source heterogeneous geological data fusion method according to a first aspect embodiment of the present invention includes: S110. Define semantic constraint framework conditions, wherein the semantic constraint framework conditions include an entity type set, a relation type set, and an attribute type set. The entity type set includes multiple entity names used to characterize the extracted objects. The relation type set includes multiple relation features used to characterize the logical connection relationship between the extracted objects. The attribute type set includes multiple attribute features used to characterize the descriptive attributes of the extracted objects. S120. Acquire multiple heterogeneous geological data, wherein the multiple heterogeneous geological data include at least a combination of text data and graphic data; S130. Analyze each of the heterogeneous geological data to obtain the vector content in each heterogeneous geological data; S140. Using an artificial intelligence model to perform semantic extension under the conditional constraints of a semantic constraint framework, and extracting multiple candidate feature data from the vector content obtained from each heterogeneous geological data, wherein the candidate feature data includes entity names and at least one of relational features and attribute features corresponding to the entity names. S150. Perform a similarity judgment and fusion step on the candidate feature data to obtain the result data. In the similarity judgment and fusion step, calculate the similarity degree of any two candidate feature data from different heterogeneous geological data. When the similarity degree of the two candidate feature data meets the similarity fusion condition, merge the two candidate feature data into one. The similarity degree is obtained from at least one of the entity name, relation features and attribute features in the two candidate feature data. S160. Output and store the result data.

[0025] Heterogeneous geological data can include unstructured geological report text, semi-structured CAD geological maps, structured tabular data, etc. The vector content of unstructured geological report text and structured tabular data is usually mainly text data, and contains some tabular and graphic data.

[0026] The vector content of semi-structured CAD geological maps is mainly graphic data, with some text data such as text annotations. When acquiring heterogeneous geological data, exemplified by CAD format files, the graphic data is presented as primitives. These primitives can be directly read by the system in steps S120 and S130. A primitive is the smallest geometric unit constituting the graphic data; it is a vector object stored in the underlying data table of the CAD file, containing precise geometric coordinate information but often lacking direct geological semantics. Text annotations in heterogeneous geological data in CAD format can record object entities, and the graphic data maps to these object entities in the text annotations, recording logical connections between them, such as intersection and containment. Essentially, it is a precise sequence of geometric coordinates for primitives (such as polylines, closed polygons, and text insertion points) in the underlying data table of the CAD file. For intersection relationships, the system uses computational geometry algorithms (such as Boolean operations) to determine whether the vertex trajectory of linear primitives (faults) crosses the closed boundary of planar primitives (strata), thereby identifying geological topological logic such as "faults cutting strata". For inclusion relationships, the system uses spatial topological operations or density-based spatial clustering algorithms (DBSCAN) to determine whether borehole coordinates are located inside a specific stratum polygon, thereby identifying geological topological relationships. In addition, attribute features can be extracted from graphic shapes and text annotations.

[0027] As can be seen from the above, the vector content in heterogeneous geological data includes both text data and graphic data.

[0028] For the definition of semantic constraint framework conditions in step S110, the system can provide a configuration interface or JSON configuration file, allowing users to define the ontology in the geological field.

[0029] For example, when defining an entity type set, you can input terms such as "Devonian dolomite" (related to ore-hosting strata), "tensional-torsional fault" (related to ore-conducting structures), "low-temperature silicification" and "dolomitization" (related to hydrothermal alteration), and "sphalerite," "galena," and "pyrite" (related to mineral assemblages) as entity names representing the extracted objects.

[0030] Define a set of relation types. You can input terms such as "cut through", "integrated contact", "fill in", and "accompany" into the set of relation types as relation features that characterize the logical connection relationship between the extracted objects.

[0031] Define an attribute type set. You can input terms such as mechanical properties like "compression and torsion" and "tension" into the attribute type set, as well as terms such as ore body attributes like "activity period", "lead-zinc equivalent grade", and "thickness" as descriptive attributes representing the extracted object.

[0032] In step S140, the artificial intelligence model can adopt the GeoGPT model. The GeoGPT model is a large language model that can quickly process text data, improve efficiency, and reduce manual input. The staff can input the following into the GeoGPT model: Please read the vector content of the following data, and under the adaptive semantic expansion provided in the semantic constraint framework, extract the geological entity names, corresponding relationship features, and attribute features mentioned therein, and output them in JSON format. The output results include entity name, entity type, location description, attribute description, etc.

[0033] This invention presents a multi-source heterogeneous geological data fusion method. First, it acquires and analyzes the vector content from various heterogeneous geological data sets. Geological experts define semantic constraint framework conditions, inputting multiple keywords representing entity names, relational features, and attribute features of the extracted object into entity type sets, relation type sets, and attribute type sets, respectively. This allows the artificial intelligence model to be constrained by keywords from different type sets, extracting candidate feature data. Then, a pairwise similarity judgment fusion step is performed between the candidate feature data extracted from different heterogeneous geological data sets to calculate the similarity degree. When the similarity degree meets the similarity fusion condition, the two candidate feature data sets are merged into one. Finally, the obtained candidate feature data are appropriately fused to form the result data and stored. This design breaks down the fusion barriers between different modalities of data, achieving a unified representation of the same object from different modalities and enabling efficient data storage.

[0034] Therefore, by using the above methods, the artificial intelligence model can easily understand the content that needs to be extracted, ensuring the correctness of the results obtained by the model and guaranteeing the consistency of the results.

[0035] It is understandable that multiple candidate feature data can be extracted from different heterogeneous geological data. These heterogeneous geological data may exist in different formats, such as CAD format, text format, and tabular format. For a given entity object, there may be content records related to it, and some content may be duplicated. During the extraction process, three candidate feature data are generated. To improve the efficiency of data storage and the convenience of subsequent user queries, this design requires a similarity judgment and fusion process in step S150 to fuse these candidate feature data that relate to the same entity object. Specifically, in some embodiments of this invention, the similarity fusion conditions include a similarity fusion threshold and a suspected association threshold, where the suspected association threshold is less than the similarity fusion threshold. The similarity judgment and fusion step includes: The weighted composite value is calculated by weighting the similarity of corresponding factors in the two candidate feature data. If the weighted composite value is greater than or equal to the similarity fusion threshold, the two candidate feature data will be merged into one. If the weighted composite value is less than the similarity fusion threshold and the weighted composite value is greater than or equal to the suspected association threshold, then the two candidate feature data are placed in the manual review queue and a manual review instruction is obtained. If the manual review instruction represents the same entity object, then the two candidate feature data are merged into one. If the manual review instruction represents two independent entity objects, then the two candidate feature data are used to calculate the similarity with other candidate feature data respectively. If the weighted composite value is less than the suspected association threshold, then the similarity between the two candidate feature data and other candidate feature data is calculated respectively.

[0036] It should be noted that since the artificial intelligence model extracts candidate feature data from different heterogeneous geological data, even for the same entity, the content of the same factor in different candidate feature data may differ. For example, the entity name of the same entity may not be exactly the same in different candidate feature data, and the content recorded about spatial location may also differ, as may the content recorded about extension or relative direction. Therefore, the similarity weight calculation is performed on corresponding (i.e., the same type) factors in two candidate feature data. For example, if two candidate feature data have two different entity names, the similarity of the corresponding factor of entity name is calculated. Then, since multiple similarity calculation values ​​of different factors are generated, these similarity calculation values ​​are weighted according to the weight of different factors to obtain a weighted comprehensive value. For example, whether the entity name is similar is a more important comparison feature, so the entity name has a higher weight, while other factors are relatively lower.

[0037] After calculating the weighted composite value, if the weighted composite value is greater than or equal to the similarity fusion threshold, it can be characterized that the two candidate feature data have a high probability of referring to the same entity object, and the two candidate feature data can be merged.

[0038] When the weighted composite value is less than the similarity fusion threshold and greater than or equal to the suspected association threshold, it indicates that it is difficult for the system to directly determine whether the two candidate feature data refer to the same entity object. In this case, manual intervention can be requested, and the two candidate feature data can be placed in the manual review queue. Specifically, the data can be output on the user's display device, allowing the user to make a manual judgment and input the manual review instruction representing the judgment result through the input device. If the manual judgment determines that they refer to the same entity object, the candidate feature data can be merged. If the manual judgment determines that they are two independent entity objects, or if the weighted composite value is less than the suspected association threshold, the two candidate feature data will be used to calculate the similarity of each other with other candidate feature data.

[0039] It is understandable that each candidate feature data extracted from heterogeneous geological data needs to be similar to each candidate feature data in other heterogeneous geological data, until candidate feature data with similarity that meets the similarity fusion condition is found and merged, or until no candidate feature data with similarity that meets the similarity fusion condition is found after traversing each candidate feature data. For candidate feature data that is not merged, it exists as separate data in each result data.

[0040] In some embodiments of the present invention, such as Figure 2 As shown, the weighted comprehensive value calculated based on the similarity of corresponding factors in the two candidate feature data includes: S210. Extract entity names, relational features, and factors from attribute features in the candidate feature data; S220. Calculate the similarity value and weighted value of multiple corresponding factors respectively; S230. Calculate the weight coefficient of each factor based on the weighted value of each factor; S240. Calculate the weighted composite value using the weighting coefficients and the similarity values ​​of each factor. The weighted composite value is: ; in, This is the weighted composite value. The similarity value for the entity names. The weighted value of the factors for entity name. The similarity value of factors representing relational characteristics. The weighted values ​​of the factors that define the relationship characteristics. The similarity value of factors representing attribute features. This is the weighted value of the attribute feature factors. Examples of entity name factors, relation feature factors, and attribute feature factors are described below.

[0041] It is understandable that for two candidate feature data from different heterogeneous geological data, there may be corresponding factors such as entity name, relation features, and attribute features, or there may only be a few combinations of these. When performing weighted calculations, if a fixed weight coefficient is used to calculate the weighted composite value for two candidate feature data with few corresponding factor combinations, even if the corresponding factors are highly similar, the calculated weighted composite value will be low and insufficient to be greater than or equal to the similarity fusion threshold, which will lead to misjudgment.

[0042] Therefore, the weight coefficients of each factor in this design are calculated from the weighted values ​​of the corresponding factors. The weight coefficients will dynamically change depending on the number of combinations of the corresponding factors. The weighted values ​​of each category of factors can be preset by the user according to the actual situation; for example, the weighted value of the entity name factor... It can be 0.5, the weighted value of the factors representing the relationship characteristics. It can be 0.3, the weighted value of the attribute features. It can be 0.2. When both candidate feature data have corresponding factors for entity name, relation feature, and attribute feature, the weight coefficient of the entity name factor is... The weight coefficients of the factors representing relational characteristics are: The weight coefficients of the attribute features are: When two candidate feature data only have corresponding factors for entity name and relation feature, the weight coefficient of the entity name factor is: The weight coefficients of the factors representing relational characteristics are: This ensures the rationality of the weighted composite value and improves the accuracy of the similarity judgment fusion step.

[0043] In some embodiments of the present invention, the entity name factor among the candidate features includes name similarity, and the similarity judgment fusion step includes: Obtain the entity names from two candidate feature data sets, and calculate the similarity value between the two entity names based on the edit distance algorithm.

[0044] The edit distance algorithm quantifies the similarity between two strings by calculating the minimum number of edit operations required to transform one string into another. The fewer the number of operations, the higher the similarity; the more the number of operations, the lower the similarity.

[0045] For example, suppose the length of the first string is m and the length of the second string is n, calculate the minimum number of edits required to convert the first string to the second string. Then reduce the minimum number of edits. Normalized to normalized per unit value The similarity value between the two entity names is obtained by dividing the range of the two entity names.

[0046] For example, the entity name extracted from the heterogeneous geological data in the report text is "F1 extensional-torsional fault," and the entity name extracted from the heterogeneous geological data in the CAD format file is also "F1." A similarity value can be calculated using the edit distance algorithm. .

[0047] In some embodiments of the present invention, the factors of relational features among candidate features include coordinate similarity, and the similarity judgment fusion step includes: Obtain the spatial coordinates from two candidate feature data sets, and calculate the Euclidean distance between the two candidate feature data sets as the similarity value of the coordinates; The Euclidean distance between the two candidate feature data is: ; in, Let be the Euclidean distance between two candidate feature data points a and b. The spatial coordinates of candidate feature data point a are ( ). The spatial coordinates of candidate feature data b are ( ); The similarity value of the coordinates of two candidate feature data is: .

[0048] The degree of positional overlap can be determined by the Euclidean distance between the spatial coordinates of two candidate feature data. The smaller the Euclidean distance, the higher the degree of spatial overlap and the higher the similarity of coordinates.

[0049] In some embodiments of the present invention, the attribute features among the candidate features include directional similarity, and the similarity judgment fusion step includes: Obtain the orientation information from two candidate feature data, where the orientation information includes the tilt angle or the range of orientation; The cosine similarity is calculated based on the orientation deviation value of the orientation information in the two candidate feature data and used as the similarity value of the orientation similarity.

[0050] Specifically, in the calculation of cosine similarity, if one of the directional information is a directional range, the center direction of that directional range is first calculated, and the maximum deviation angle Δθmax of the directional range is also calculated. The directional deviation value Δθactual is calculated using the center direction and the center direction or the tilt angle of the other directional information. Finally, the cosine value is calculated using the directional deviation value and the maximum deviation angle to obtain the cosine similarity. .

[0051] Furthermore, the obtained cosine similarity can be normalized to a normalized per-unit value. Interval.

[0052] For example, if the text data in the report records the directional information as "southeast," specifically within the range of [118°, 146°], and the directional information extracted from the graphic data in the CAD format file is 135°, then the cosine similarity can be calculated. It is 0.943.

[0053] When the similarity between two candidate feature data satisfies the similarity fusion condition, how should two contents with deviations in the same type of feature factors be selected during the merging process of the two candidate feature data? In some embodiments of the present invention, this design sets each heterogeneous geological data to have a fusion priority value. Specifically, different fusion priority values ​​can be set according to different feature factors in each heterogeneous geological data. For example, factors involving entity name features are set with a first fusion priority value, while factors involving location coordinate features are set with a second fusion priority value, etc.

[0054] The process of merging two candidate feature data into one includes: Perform a heterogeneous incremental step and a homogeneous merging step on two candidate feature data. In the heterogeneous incremental step, if the type of a feature in one candidate feature data does not exist in the other candidate feature data, then the content of that feature is incrementally added to the result data. In the homogeneous merging step, if the type of a feature in two candidate feature data is the same, then compare the fusion priority values ​​of the two candidate feature data, select the content of that feature in the candidate feature data with the higher fusion priority value to add to the result data, and discard the content of that feature in the candidate feature data with the lower fusion priority value.

[0055] It is understandable that, in any candidate feature data, if the type of one feature has not appeared in another candidate feature data, the feature content of this type can be directly incremented into the result data of the corresponding entity object. However, if the type of one feature in two candidate feature data is the same, but the feature content is different, the content of that feature in the candidate feature data with the higher fusion priority value is selected and added to the result data.

[0056] For example, the text data in the report states that F1 fault has a tensional-torsional mechanical property, associated alteration of strong silicification / breccia, and a measured width of 5m. Meanwhile, the graphic data extracted from the CAD file shows F1 with an attitude of 135°∠75° and a measured width of 4.8m. Since the characteristics of the mechanical property (tensional-torsional), associated alteration of strong silicification / breccia, and attitude (135°∠75°) do not conflict, the heterogeneous incremental step is executed. However, if the characteristic types of the entity name and the measured width of F1 fault conflict, the same-type merging step is executed. In this step, the user sets the merging priority value of the entity name in the report text to be higher than that in the CAD file, while the merging priority value of the measured width in the report text is lower than that in the CAD file. Therefore, the merged result is: F1 fault, mechanical property: tensional-torsional, associated alteration of strong silicification / breccia, attitude: 135°∠75°, measured width: 4.8m.

[0057] A processing system according to a second aspect of the present invention is used to execute the multi-source heterogeneous geological data fusion method disclosed in any of the above embodiments, such as... Figure 3 As shown, the processing system includes: Define module 310, which is used to define semantic constraint framework conditions; Module 320 is used to acquire multiple heterogeneous geological data; Analysis module 330 is used to analyze each of the heterogeneous geological data to obtain vector content in each heterogeneous geological data; The extraction and processing module 340 is used to perform semantic expansion based on the semantic constraint framework using an artificial intelligence model, and to extract multiple candidate feature data from the vector content obtained from each heterogeneous geological data. The fusion processing module 350 is used to perform a similarity judgment fusion step on the candidate feature data to obtain the result data. In the similarity judgment fusion step, the similarity degree of any two candidate feature data from different heterogeneous geological data is calculated. When the similarity degree of the two candidate feature data meets the similarity fusion condition, the two candidate feature data are merged into one. Storage module 360 ​​is used to output and store the result data.

[0058] It should be noted that the specific implementation process of this embodiment can be found in the specific implementation process described in the above method embodiments, and will not be repeated here.

[0059] The processing system of this invention applies the multi-source heterogeneous geological data fusion method disclosed in any of the above embodiments, breaks down the fusion barriers of different modal data, realizes unified representation of the same object from different modal data, and achieves efficient data storage.

[0060] After storing the various result data, the result data can be imported into a graph database to construct a geological knowledge graph.

[0061] Specifically, the result data contains entity objects and related relational and attribute features. Topological relationships can be constructed using this data. For example, based on whether the coordinates of borehole ZK01 are located inside a closed polygon in the K1 stratum, if so, a connection relationship is automatically created. Users can also input query commands on smart devices, and the system will retrieve related paths from the geological knowledge graph, returning a complete geological knowledge sub-graph containing text descriptions, spatial locations, and attribute data.

[0062] After the results data are integrated and organized, the system no longer performs black-box queries on various heterogeneous geological data. Instead, based on the constructed high-precision geological knowledge graph, it provides users with cross-modal data joint retrieval, multi-hop logical traversal and statistical analysis services through the query interface of the graph database, assisting geological experts in data verification and pattern summarization.

[0063] For example, geologists hope to find all fault structures in the mining area that "cut through Devonian strata and are accompanied by sphalerite mineralization".

[0064] Command Conversion: The user inputs the query conditions: {Entity Type: "Fault", Attribute: "Extensional", Associated Strata: "Devonian", Associated Mineral: "Sparkling Zinc"}. The system's backend query parser converts the above semantic conditions into a standard graph database query statement, for example: MATCH (f:Fault {property: 'Extensional'})-[:CUTS]->(s:Strata{name:'Devonian'}), (f)-[:ACCOMPANIES]->(m:Mineral {name: 'Sparkling Zinc'}) RETURN f,s, m.

[0065] The system performs quantitative statistical analysis on the data and presents it in a combined "figure-text-table" view to support geological experts in decision-making.

[0066] The system automatically calculates the statistical characteristics of the retrieved fault set. For example, it calculates the average strike and dip distribution histograms of all ore-bearing faults that meet the criteria, as well as the average lead-zinc grades of the strata on both sides of the fault.

[0067] The geological atlas view can display the topological connection diagram of "fault-strata-mineralization," intuitively showing the logical relationships between entities. Furthermore, through the GIS / CAD mapping view, the highlighted geometric trajectories of the retrieved F1 and F5 faults can be overlaid on the original engineering plan, allowing users to zoom in and out to view their specific spatial locations. Additionally, clicking on a specific fault in the view interface simultaneously displays a summary of its original report, borehole analysis data, and a screenshot of its CAD occurrence in the sidebar, achieving "one search, complete view." Through these displays, geological experts can quickly verify geological hypotheses such as "whether extensional faults universally control ore deposits," thereby formulating the next exploration plan based on accurate statistical data, rather than relying on the system's automatic predictions.

[0068] According to a third aspect of the present invention, the control device includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the multi-source heterogeneous geological data fusion method disclosed in any of the above embodiments.

[0069] The control device can be any intelligent terminal, including a central computer, a remote equipment terminal computer, or any other intelligent terminal.

[0070] Another embodiment of the control device may further include: a processor, which may be implemented in the form of a general-purpose central processing unit (CPU), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, for executing relevant programs to implement the technical solutions provided in the embodiments of this application; The memory can be implemented in the form of read-only memory (ROM), static storage device, dynamic storage device, or random access memory (RAM). The memory can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory and called and executed by the processor to execute the multi-source heterogeneous geological data fusion method of the embodiments of this application. Input / output interfaces are used to implement information input and output; The communication interface is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.). A bus transmits information between various components of a device (such as processors, memory, input / output interfaces, and communication interfaces), and can also be used to access the smart Internet of Things. The processor, memory, input / output interfaces, and communication interfaces communicate with each other within the device via a bus.

[0071] According to a fourth aspect of the present invention, a computer-readable storage medium stores a computer program, characterized in that, when executed by a processor, the computer program implements the multi-source heterogeneous geological data fusion method disclosed in any of the above embodiments.

[0072] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0073] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0074] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0075] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0076] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0077] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0078] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

[0079] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0080] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for fusing multi-source heterogeneous geological data, characterized in that, include: Define semantic constraint framework conditions, wherein the semantic constraint framework conditions include an entity type set, a relation type set, and an attribute type set. The entity type set includes multiple entity names used to characterize the extracted objects. The relation type set includes multiple relation features used to characterize the logical connection relationships between the extracted objects. The attribute type set includes multiple attribute features used to characterize the descriptive attributes of the extracted objects. Acquire multiple heterogeneous geological data, wherein the heterogeneous geological data includes at least a combination of text data and graphic data; The heterogeneous geological data are analyzed to obtain vector content in each heterogeneous geological data, including text data and graphic data. The semantic extension is performed using an artificial intelligence model under the conditional constraints of a semantic constraint framework, and multiple candidate feature data are extracted from the vector content of each heterogeneous geological data. The candidate feature data includes entity names and at least one of relational features and attribute features corresponding to the entity names. A similarity judgment and fusion step is performed on the candidate feature data to obtain the result data. In the similarity judgment and fusion step, the similarity degree of any two candidate feature data from different heterogeneous geological data is calculated. When the similarity degree of the two candidate feature data meets the similarity fusion condition, the two candidate feature data are merged into one. The similarity degree is obtained by at least one factor among entity name, relation feature and attribute feature in the two candidate feature data. Output and store the results data.

2. The multi-source heterogeneous geological data fusion method according to claim 1, characterized in that, The similarity fusion conditions include a similarity fusion threshold and a suspected association threshold, wherein the suspected association threshold is less than the similarity fusion threshold. The similarity judgment fusion step includes: The weighted composite value is calculated by weighting the similarity of corresponding factors in the two candidate feature data. If the weighted composite value is greater than or equal to the similarity fusion threshold, the two candidate feature data will be merged into one. If the weighted composite value is less than the similarity fusion threshold and the weighted composite value is greater than or equal to the suspected association threshold, then the two candidate feature data are placed in the manual review queue and a manual review instruction is obtained. If the manual review instruction represents the same entity object, then the two candidate feature data are merged into one. If the manual review instruction represents two independent entity objects, then the two candidate feature data are used to calculate the similarity with other candidate feature data respectively. If the weighted composite value is less than the suspected association threshold, then the similarity between the two candidate feature data and other candidate feature data is calculated respectively.

3. The multi-source heterogeneous geological data fusion method according to claim 2, characterized in that, The weighted composite value calculated based on the similarity of corresponding factors in the two candidate feature data includes: Extract entity names, relationship features, and attribute features from candidate feature data; Calculate the similarity scores and weighted scores for multiple corresponding factors separately; The weight coefficient of each factor is calculated based on the weighted value of each factor; The weighted composite value is calculated using the weighting coefficients and the similarity values ​​of each factor. The weighted composite value is: ； in, This is the weighted composite value. The similarity value for the entity names. The weighted value of the factors for entity name. The similarity value of factors representing relational characteristics. The weighted values ​​of the factors that define the relationship characteristics. The similarity value of factors representing attribute features. The weighted value of the factors that represent the attribute characteristics.

4. The multi-source heterogeneous geological data fusion method according to claim 2, characterized in that, The entity name factor in the candidate features includes name similarity, which is included in the similarity judgment fusion step: Obtain the entity names from two candidate feature data sets, and calculate the similarity value between the two entity names based on the edit distance algorithm.

5. The multi-source heterogeneous geological data fusion method according to claim 2, characterized in that, The factors of relational features among candidate features include coordinate similarity, and the similarity judgment fusion step includes: Obtain the spatial coordinates from two candidate feature data sets, and calculate the Euclidean distance between the two candidate feature data sets as the similarity value of the coordinates; The Euclidean distance between the two candidate feature data is: ； in, Let be the Euclidean distance between two candidate feature data points a and b. The spatial coordinates of candidate feature data point a are ( ). The spatial coordinates of candidate feature data b are ( ); The similarity value of the coordinates of two candidate feature data is: 。 6. The multi-source heterogeneous geological data fusion method according to claim 2, characterized in that, The factors of attribute features among candidate features include directional similarity, and the similarity judgment fusion step includes: Obtain the orientation information from two candidate feature data, where the orientation information includes the tilt angle or the range of orientation; The cosine similarity is calculated based on the orientation deviation value of the orientation information in the two candidate feature data and used as the similarity value of the orientation similarity.

7. The multi-source heterogeneous geological data fusion method according to claim 1, characterized in that, Each heterogeneous geological data set has a fusion priority value, which includes merging two candidate feature data sets into one: Perform heterogeneous incremental step and homogeneous merging step on two candidate feature data; In the heterogeneous incremental step, if the type of a feature in one of the candidate feature data does not exist in another candidate feature data, then the content of that feature is incrementally added to the result data. In the same merging step, when one of the features in two candidate feature data is of the same type, the fusion priority values ​​of the two candidate feature data are compared. The content of that feature in the candidate feature data with the higher fusion priority value is added to the result data, while the content of that feature in the candidate feature data with the lower fusion priority value is discarded.

8. A processing system for executing the multi-source heterogeneous geological data fusion method as described in any one of claims 1 to 7, characterized in that, The processing system includes: Define the module, which is used to define semantic constraint framework conditions; The acquisition module is used to acquire multiple heterogeneous geological data. The analysis module is used to analyze the various heterogeneous geological data to obtain the vector content in each heterogeneous geological data. The extraction and processing module is used to perform semantic expansion based on the semantic constraint framework using an artificial intelligence model, and to extract multiple candidate feature data from the vector content obtained from each heterogeneous geological data. The fusion processing module is used to perform a similarity judgment fusion step on the candidate feature data to obtain the result data. In the similarity judgment fusion step, the similarity degree of any two candidate feature data from different heterogeneous geological data is calculated. When the similarity degree of two candidate feature data meets the similarity fusion condition, the two candidate feature data are merged into one. The storage module is used to output and store the result data.

9. A control device, characterized in that, The control device includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the multi-source heterogeneous geological data fusion method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the multi-source heterogeneous geological data fusion method according to any one of claims 1 to 7.

Images