A method and equipment for constructing and using geological data indexes

Abstract

This application discloses a method and apparatus for constructing and using a geological data index, relating to the field of data management. The method includes: acquiring a geological data feature vector library and a geological knowledge topic library; the geological data feature vector library is a set of high-dimensional feature vectors corresponding to several geological object data groups; the geological knowledge topic library includes several geological knowledge topics; using a dimensionality reduction algorithm, the geological data feature vector library is dimensionality reduced to obtain a two-dimensional coordinate set; the two-dimensional coordinate set includes several two-dimensional coordinates; the two-dimensional coordinates and the geological object data groups have a one-to-one correspondence; using the RBush algorithm, a spatial index tree structure corresponding to the two-dimensional coordinate set is constructed; and a semantic index structure is constructed based on the geological knowledge topic library and the topic attributes of the geological objects. This application enables efficient retrieval based on global topic association and fine-grained local regional positioning.

Description

Technical Field

[0001] This application relates to the field of data management technology, and in particular to a method and device for constructing and using geological data indexes. Background Technology

[0002] In the field of geological exploration and informatization, with the continuous advancement of data acquisition technology, geological survey data are becoming increasingly complex and their scale is constantly growing. Traditional data management and retrieval methods are no longer sufficient to meet the demands for efficient and accurate applications. Geological survey data is typically high-dimensional data, including spatial, physical attributes, chemical composition, and other multi-dimensional data. When this high-dimensional data is mapped to a two-dimensional index, traditional indexes cannot dynamically adapt to the data distribution. Cross-data source geological knowledge retrieval lacks an effective association mechanism, making it difficult to balance global topic associations with precise local area positioning. Furthermore, existing index architectures are not designed in conjunction with the characteristics of geological survey data, resulting in poor adaptability. These problems lead to poor geological data retrieval efficiency, insufficient accuracy, and low content matching. Summary of the Invention

[0003] The purpose of this application is to provide a method and device for constructing and using geological data indexes, which can achieve efficient retrieval based on global topic association and fine local regional positioning.

[0004] To achieve the above objectives, this application provides the following solution: Firstly, this application provides a method for constructing a geological data index, including: A geological data feature vector library and a geological knowledge topic library are acquired; the geological data feature vector library is a collection of high-dimensional feature vectors corresponding to several geological object data groups; the geological knowledge topic library includes several geological knowledge topics. The geological data feature vector library is dimensionality reduced using a dimensionality reduction algorithm to obtain a two-dimensional coordinate set; the two-dimensional coordinate set includes several two-dimensional coordinates; the two-dimensional coordinates and the geological object data group have a one-to-one correspondence. Using the RBush algorithm, construct a spatial index tree structure corresponding to the two-dimensional coordinate set; A semantic index structure is constructed based on the topic attributes of each geological object data group corresponding to the geological knowledge topic database and the geological data feature vector database.

[0005] Secondly, this application provides a method for using geological data indexing, including: The query request is obtained and parsed to obtain spatial and semantic conditions; the spatial conditions include the query area; the semantic conditions include the query geological knowledge topic. The spatial index tree structure and the semantic index structure are retrieved based on spatial and semantic conditions, respectively, to obtain a first set of geological object identifiers and a second set of geological object identifiers; the spatial index tree structure and the semantic index structure are determined according to the geological data index construction method described in the first aspect; The final geological object identifier set is determined based on the intersection of the first geological object identifier set and the second geological object identifier set. Based on the final set of geological object identifiers, retrieve the corresponding geological object data from the geological object database.

[0006] Thirdly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the geological data index construction method described in the first aspect, or the geological data index usage method described in the second aspect.

[0007] According to the specific embodiments provided in this application, this application has the following technical effects: This application provides a method and device for constructing and using geological data indexes. By utilizing a dimensionality reduction algorithm, the geological data feature vector library is dimensionality reduced to obtain a set of two-dimensional coordinates (points). Then, the RBush algorithm is used to construct a spatial index tree structure corresponding to the two-dimensional coordinate set. Compared with traditional static grid indexes, the dynamic spatial index based on the RBush algorithm can dynamically adapt to data distribution and can adaptively adjust as data is added or deleted without reconstruction. This overcomes the shortcomings of traditional static grid indexes, which perform poorly in both sparse and dense areas of geological data, and significantly improves retrieval efficiency and resource utilization. Furthermore, based on the geological knowledge subject base and the subject attributes of geological objects, a semantic index structure is constructed, which can support the simultaneous processing of spatial and semantic information, achieving both global subject association and fine local positioning. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is an application environment diagram of a geological data index construction method according to an embodiment of this application; Figure 2 A flowchart illustrating a geological data index construction method provided in an embodiment of this application; Figure 3A flowchart illustrating a method for using a geological data index according to an embodiment of this application; Figure 4 A flowchart illustrating a geological data indexing method provided in an embodiment of this application; Figure 5 A schematic diagram of the functional modules of a geological data indexing system provided in another embodiment of this application; Figure 6 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0010] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0011] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0012] Example 1 The geological data indexing construction method provided in this application embodiment can be applied to, for example... Figure 1In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be set up independently, integrated into server 104, or placed in the cloud or on another server. Terminal 102 can send the geological data feature vector library and geological knowledge topic library to be processed to server 104. After receiving the geological data feature vector library and geological knowledge topic library, server 104 uses a dimensionality reduction algorithm to perform dimensionality reduction processing on the geological data feature vector library to obtain a two-dimensional coordinate set. The two-dimensional coordinate set includes several two-dimensional coordinates; the two-dimensional coordinates have a one-to-one correspondence with the geological object data groups. Using the RBush algorithm, a spatial index tree structure corresponding to the two-dimensional coordinate set is constructed. Based on the topic attributes of each geological object data group corresponding to the geological knowledge topic library and geological data feature vector library, a semantic index structure is constructed. Server 104 can feed back the obtained spatial index tree structure and semantic index structure to terminal 102. In addition, in some embodiments, the geological data index construction method can also be implemented by the server 104 or the terminal 102 separately. For example, the terminal 102 can directly construct the index for the geological data feature vector library and the geological knowledge topic library, or the server 104 can obtain the geological data feature vector library and the geological knowledge topic library from the data storage system and construct the index.

[0013] The terminal 102 can be, but is not limited to, various desktop computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. The server 104 can be implemented using a standalone server or a server cluster consisting of multiple servers, or it can be a cloud server.

[0014] like Figure 2 As shown, this embodiment provides a method for constructing a geological data index. This method is executed by a computer device, specifically by a terminal or server alone, or by both a terminal and a server. In this embodiment, the method is applied to... Figure 1 Taking server 104 as an example, the explanation includes the following steps 201 to 204. Wherein: Step 201: Obtain the geological data feature vector library and the geological knowledge topic library; the geological data feature vector library is a collection of high-dimensional feature vectors corresponding to several geological object data groups; the geological knowledge topic library includes several geological knowledge topics (hereinafter referred to as topics). There is a one-to-one correspondence between the geological object data groups and the high-dimensional feature vectors.

[0015] Step 202: Using a dimensionality reduction algorithm, the geological data feature vector library is dimensionality reduced to obtain a two-dimensional coordinate set; the two-dimensional coordinate set includes several two-dimensional coordinates; the two-dimensional coordinates and the geological object data group have a one-to-one correspondence.

[0016] Step 203: Use the RBush algorithm to construct a spatial index tree structure corresponding to the two-dimensional coordinate set.

[0017] Step 204: Construct a semantic index structure based on the topic attributes of each geological object data group corresponding to the geological knowledge topic base and the geological data feature vector base. The topic attributes include relevant topics and corresponding weights.

[0018] As an optional implementation, the dimensionality reduction algorithm is the t-SNE algorithm; the specific process of using the t-SNE algorithm to perform dimensionality reduction processing on the geological data feature vector library is as follows: Step 202.1: Standardize each high-dimensional feature vector in the geological data feature vector library to obtain a standardized feature vector library.

[0019] Step 202.2: Calculate the conditional probability matrix corresponding to the standardized feature vector library, and perform symmetry processing on the conditional probability matrix to obtain the joint probability matrix.

[0020] Step 202.3: Randomly select a point set in the two-dimensional space and calculate the joint probability matrix of similarity corresponding to the point set; the number of points in the point set is equal to the number of high-dimensional feature vectors in the geological data feature vector library.

[0021] Step 202.4: With the goal of minimizing the KL divergence between the joint probability matrix corresponding to the standardized feature vector library and the joint probability matrix of similarity corresponding to the point set, the coordinates of the points in the point set are iteratively optimized to obtain the optimal set of points; the optimal set of points is determined as the two-dimensional coordinate set.

[0022] For any row in the conditional probability matrix, the elements of that row represent the conditional probabilities of a normalized eigenvector and itself or any other normalized eigenvector; where: A standardized feature vector has a conditional probability of 0 for itself. The conditional probability of a standardized eigenvector corresponding to any other standardized eigenvector is calculated using a conditional probability calculation model; the mathematical expression of the conditional probability calculation model is: in, Represented by standardized feature vectors Standardized feature vectors Standardized feature vectors Standardized feature vectors Standardized feature vector; symbol Indicates Euclidean distance; Represented by standardized feature vectors The standard deviation of a Gaussian distribution centered at .

[0023] For any row in the joint similarity probability matrix, the elements of the row represent the joint similarity probability between a point and itself or any other point; where: The joint probability of a point's similarity to itself is 0. The joint probability of similarity between a point and any other point is calculated using a joint similarity probability calculation model; the mathematical expression of the joint similarity probability calculation model is: in, Point and points The joint probability of similarity; Point The coordinates; Point The coordinates; Point The coordinates; Point Coordinates of the point; ,point ,point and points All are points in the set of points; symbol This represents the Euclidean distance.

[0024] As an optional implementation, step 203, "constructing a spatial index tree structure corresponding to the two-dimensional coordinate set using the RBush algorithm," specifically includes: Step 203.1: Generate the corresponding minimum bounding rectangle for each point in the two-dimensional coordinate set.

[0025] Step 203.2: Using the minimum bounding rectangle corresponding to each point in the two-dimensional coordinate set as input, the RBush algorithm is used to construct a spatial index tree structure according to a preset insertion strategy and preset node splitting rules. The spatial index tree structure includes leaf nodes and internal nodes. The internal nodes include the root node and several intermediate nodes between the root node and any leaf node. Each leaf node includes the identifier of several geological object data groups and the minimum bounding rectangle. Each internal node includes the corresponding minimum bounding rectangle and the addresses of all its next-level nodes. The minimum bounding rectangle corresponding to any internal node is the smallest rectangle that includes the minimum bounding rectangles corresponding to all its next-level nodes.

[0026] As an optional implementation, the preset insertion strategy specifically includes: 1-1) For a minimum bounding rectangle to be inserted, starting from the root node, recursively select the insertion path with the goal of minimizing the area expansion of the minimum bounding rectangle corresponding to the node after insertion, until the leaf node is reached.

[0027] 1-2) Insert the minimum outer rectangle into the leaf node and update all internal nodes on the path to which the leaf node belongs.

[0028] As an optional implementation, the preset node splitting rules specifically include: 2-1) When the number of identifiers in the inserted leaf node exceeds the maximum capacity, all identifiers in the inserted leaf node are divided into two groups according to the preset fill degree.

[0029] 2-2) Create corresponding leaf nodes for each group to replace the inserted leaf nodes, and update all internal nodes on the path. If the number of next-level node addresses in any internal node exceeds the maximum capacity, divide all next-level node addresses in the internal node into two groups according to the preset fill degree. Create corresponding internal nodes for each group to replace the internal nodes.

[0030] As an optional implementation, step 204, "based on the theme attributes of each geological object data group corresponding to the geological knowledge theme base and the geological data feature vector base," specifically includes: Step 204.1: Based on the theme attributes of the geological object data group, for each geological knowledge theme in the geological knowledge theme database, retrieve the identifier of the geological object data group that includes the corresponding geological knowledge theme and whose corresponding weight exceeds a preset threshold; wherein, the weight represents the degree of correlation between the geological object data group and the geological knowledge theme.

[0031] In essence, weights are quantified values ​​of correlation. In geological data indexing, weights are used to filter data groups strongly related to the theme (e.g., identifiers with weights exceeding a threshold are retrieved), ensuring the index focuses on key geological information. Due to the multi-source heterogeneity, spatial complexity, and uncertainty of geological data, determining weights may involve professional knowledge such as geological ontology, spatial topology, and grey system theory. The general process follows "feature extraction - model matching - professional correction - threshold verification," such as: extracting feature vectors from geological object data groups; constructing feature models for geological knowledge themes; calculating basic correlation weights; correcting weights using geological professional rules; and normalizing and verifying weights.

[0032] Step 204.2: Using each geological knowledge topic as the key and the list of identifiers of all geological object data groups retrieved based on a geological knowledge topic as the value, construct a semantic index structure.

[0033] Example 2 This embodiment provides a method for using geological data indexing. (See attached document.) Figure 3 This includes the following steps: Step 301: Obtain the query request and parse it to obtain spatial conditions and semantic conditions; the spatial conditions include the query area; the semantic conditions include the query geological knowledge topic.

[0034] Step 302: The spatial index tree structure and the semantic index structure are retrieved according to the spatial conditions and semantic conditions, respectively, to obtain the first set of geological object identifiers and the second set of geological object identifiers; the spatial index tree structure and the semantic index structure are determined according to the geological data index construction method in Example 1.

[0035] Step 303: Determine the final geological object identifier set based on the intersection of the first geological object identifier set and the second geological object identifier set.

[0036] Step 304: Obtain the corresponding geological object data set from the geological object database based on the final geological object identifier set.

[0037] Example 3 To aid understanding by those skilled in the art, the conceptual design of this application is further illustrated below through this embodiment. This embodiment includes the construction and use of geological data indexes, referred to as the geological data index construction and use method, or simply the geological data indexing method. This method comprises five main steps, see reference... Figure 4 .

[0038] (a) Data preparation.

[0039] Access high-dimensional feature vectors of the geological object data set to be indexed from the feature vector library. ,in, The total number of geological object data sets, high-dimensional feature vectors , It serves as a feature dimension, integrating multi-dimensional attributes such as spatial, physical, and chemical properties.

[0040] This vector originates from the preprocessing and integrates the spatial, physical, chemical, and other multidimensional attributes of the geological object.

[0041] (ii) High-dimensional feature mapping (spatialization mapping module).

[0042] 2.1) Dimensionality reduction algorithm selection and initialization.

[0043] The t-SNE algorithm was chosen for dimensionality reduction. Based on the amount of data and dimensionality complexity of the geological survey, the algorithm was supplemented and improved with features such as high-dimensional conditional probability calculation and symmetric probability matrix transformation. Then, a series of rules or algorithm hyperparameters were preset, such as: perplexity ∈ [5, 50]; learning rate ∈ [10, 1000]; number of iterations (n_iter) ∈ [1000, 3000]; exaggeration factor = 12 or 14; momentum = 0.8.

[0044] The full name of the t-SNE algorithm is t-distributed random neighborhood embedding.

[0045] The core objective of the t-SNE algorithm is to map high-dimensional data to a low-dimensional space while preserving the "neighborhood relationships" between data points as much as possible. Simply put, it's like "drawing" a bunch of high-dimensional points on a piece of paper, so that points that are "similar" in the original high-dimensional space are close together on the paper, while "dissimilar" points are far apart.

[0046] 2.2) Feature standardization processing.

[0047] To ensure effective dimensionality reduction, the high-dimensional feature vectors are first standardized using Z-scores to eliminate the influence of dimensions and improve the stability and effectiveness of the dimensionality reduction algorithm. The following formula can be used: in, The feature vector after standardization; This is the original feature vector; The mean; The standard deviation is denoted as .

[0048] 2.3) Similarity matrix calculation.

[0049] Calculating the pairwise similarity (or distance) between the high-dimensional feature vectors corresponding to all geological object data sets in high-dimensional space provides a basis for low-dimensional mapping. For the t-SNE algorithm, it calculates the conditional probability distribution to represent the object similarity in high-dimensional space; see the conditional probability calculation model for details.

[0050] Then, after symmetry processing, the joint probability matrix is ​​obtained.

[0051] 2.4) Low-dimensional space initialization and optimization.

[0052] A point set is randomly initialized in a two-dimensional space. Then, the positions of the points in the two-dimensional space are iteratively adjusted using an optimization algorithm (such as gradient descent) to minimize the Kullback-Leibler divergence between the probability distributions of the high-dimensional and low-dimensional spaces.

[0053] In two-dimensional space, the joint probability of similarity is calculated directly; see the joint probability calculation model for similarity.

[0054] Objective: To adjust the position of the midpoint in the low-dimensional space so that the probability distribution Q in the low-dimensional space is as similar as possible to the probability distribution P in the high-dimensional space.

[0055] A tool for measuring differences: KL divergence. This is a mathematical tool for measuring the difference between two probability distributions. The smaller the KL divergence, the more similar the two distributions are.

[0056] Optimization method: Use gradient descent. Calculate the gradient of the low-dimensional point positions with respect to the KL divergence (i.e., "in which direction should we move to reduce the difference"), and then gradually move these points along the gradient direction, iterating several times until the distribution is basically stable.

[0057] 2.5) Two-dimensional coordinate output.

[0058] After optimization and convergence, the final low-dimensional coordinate (point) set is output. ,in And establish mapping relationships , A unique identifier (or simply identifier) ​​for each geological object data set, linking each geological object data set to its two-dimensional coordinates. Establish strong associations.

[0059] This step "spatializes" the high-dimensional geological object data set, providing "indexable spatial units" for the subsequent RBush 2D index construction. These units are direct inputs for the MBR calculations in subsequent steps; without them, the coordinates and... The association is such that the index results cannot be mapped back to the original geological object.

[0060] (III) Construction of two-dimensional dynamic spatial index based on RBush algorithm (spatial index engine).

[0061] 3.1) Calculation of minimum bounding rectangle (MBR).

[0062] For each bound unique identifier Geological object data set (or, for each point) The system generates an MBR for it. Since the object is a point feature, its MBR is a minimal rectangle centered at that point (or degenerates into a single point), meaning the coordinates of its top-left and bottom-right corners are infinitely close. Typically, for the stability of numerical calculations, it is defined as a rectangle centered at that point with a side length of a minimal positive value. A rectangle.

[0063] Setting points The coordinates are Then the MBR corresponding to that point is represented as: MBR : Coordinates of the geological object data set in two-dimensional space (known quantities, from the output of the high-dimensional feature mapping step).

[0064] A very small positive number (e.g.) ).

[0065] Note: This is a necessary input for constructing the RBush tree in subsequent steps. Without MBR, node insertion and splitting of the tree cannot be performed.

[0066] 3.2) RBush tree construction (insertion).

[0067] The system calls the RBush algorithm, taking the MBR of all objects as input, to start building an R-Tree, including node selection (strategy) and node splitting.

[0068] The construction process is briefly described below.

[0069] 3.2.1) Node Definition: Each internal node contains several entries, and each entry is an MBR.

[0070] Set the maximum node capacity M (usually M=4 or 9) and the minimum fill degree m=0.4M.

[0071] All leaf nodes store the actual geological object data group identifier and its MBR.

[0072] 3.2.2) Insertion strategy: For the new MBR, recursively select the insertion path starting from the root to minimize the area expansion of the MBR node after insertion.

[0073] 3.2.3) Node splitting: If the insertion causes the number of items in a node to exceed M, it needs to be split into two new nodes. .

[0074] Explanation: This is the foundation of tree balancing. The splitting strategy directly affects the hierarchical structure of the tree and provides the initial tree framework for subsequent balancing.

[0075] 3.3) Tree structure balance.

[0076] The RBush algorithm automatically maintains a highly balanced tree structure through the aforementioned insertion strategy and splitting mechanism, ensuring that all leaf nodes are at the same depth, thereby guaranteeing the stability of query efficiency.

[0077] (iv) Construction of a two-level mapping mechanism between the topic and the knowledge base (semantic indexing engine).

[0078] 4.1) Global topic index construction.

[0079] 4.1.1) Inverted list creation: Traverse each topic in the knowledge subject base For each topic Retrieve a list of identifiers for all geological object data groups that contain the topic and have a weight exceeding a certain threshold. .

[0080] 4.1.2) Index Storage: The system is organized by topic. As the key, a list of identifiers For the value, construct a global inverted index (i.e., a semantic index structure, also known as a global topic index) and persist it.

[0081] 4.1.3) Construct the subject-object inverted index formula: .

[0082] Indicating in the index item The corresponding index structure; : indicates in Geological object data sets associated under conditions Sets, these It refers to a specific data entity (such as a spatial object, a time series data point, etc.).

[0083] 4.2) Local grid knowledge base association.

[0084] 4.2.1) Leaf Node Traversal: Traverse all leaf nodes of the RBush tree. Each leaf node represents a final dynamic mesh cell. .

[0085] 4.2.2) Object-Mesh Association: For each leaf node To obtain the unique identifier of all the geological object data groups contained therein.

[0086] 4.2.3) Knowledge base encapsulation: [This section likely refers to a process involving encapsulation or knowledge base encapsulation.] The spatial scope (i.e., the MBR of the leaf node) and its contained list of identifiers are encapsulated to form a local knowledge base unit.

[0087] Bind RBush leaf nodes to knowledge base units using formulas: ; : indicates the first Individual data units (such as grids in a spatial index, R-tree nodes, or intervals in a time series).

[0088] Minimum bounding rectangle, defined The spatial range. For example, in two-dimensional coordinates. .

[0089] Stored in Within range List.

[0090] (v) Execute hierarchical index queries (query execution engine).

[0091] 5.1) Receive and parse mixed query requests: Receive mixed query requests submitted by users, such as: find geological objects within “Region A” that are related to the topic “porphyry copper mineralization”. The parser breaks down the request into spatial conditions (“Region A”) and semantic conditions (“‘porphyry copper mineralization’ topic”).

[0092] 5.2) Spatial Query (Parallel): Using the constructed RBush index, quickly retrieve the set of identifiers S_space for all geological object data groups located within the “Region A” range.

[0093] 5.3) Semantic Query (Parallel): Utilize the constructed global inverted index to quickly retrieve the set of identifiers S_topic for all geological objects related to the topic of "porphyry copper mineralization".

[0094] 5.4) Union of result sets: After the two threads have finished executing, the engine performs an intersection operation on the obtained sets S_space and S_topic to obtain the final candidate set S_result that satisfies both spatial and semantic conditions.

[0095] 5.5) Result verification: To ensure accuracy, the specific coordinates and theme weights of the geological object data group can be retrieved from the database based on S_result to perform a precise secondary filtering of the query conditions.

[0096] The above (1)-(5) constitute a complete process from data processing (dimensionality reduction, mapping, etc.), index construction to efficient querying, which is logically progressive and interconnected.

[0097] High-dimensional feature mapping is fundamental: it maps abstract high-dimensional geological features into specific two-dimensional spatial coordinates, providing a unified spatiotemporal foundation for all subsequent spatial location-based indexes and queries.

[0098] The two-dimensional dynamic spatial index forms the skeleton: it receives two-dimensional coordinates output by the spatialization mapping module and uses the RBush algorithm to construct an efficient and balanced two-dimensional spatial index tree. This forms the spatial skeleton of the entire system, enabling rapid spatial range filtering and refined spatial management of massive geological object data.

[0099] The core of the topic-knowledge base two-level mapping is that it injects semantic information onto the spatial skeleton. By associating the global topic inverted index with the local grid knowledge base, semantics (topics) and space (grid) are tightly coupled, forming a two-level mapping mechanism of "spatial positioning of semantics and semantic enrichment of space", which is the core of realizing cross-modal hybrid retrieval.

[0100] Executing hierarchical index queries is the goal and validation: it serves as system output, directly addressing user needs. By executing spatial and semantic queries in parallel and merging the results, it effectively utilizes the constructed index system.

[0101] To address the shortcomings of existing technologies, this method utilizes an optimized RBush algorithm for two-dimensional gridded indexing. By dynamically partitioning the space, it overcomes the key challenge of efficiently mapping high-dimensional geological data to a two-dimensional plane, thus improving index response speed. Through a two-level mapping mechanism of "topic-knowledge base index map," it achieves global topic association across data sources and fine-grained local grid management, resolving the contradiction between balancing global retrieval and local positioning. Furthermore, a hierarchical indexing strategy is employed to enhance the index's adaptability to geological survey scenarios, simultaneously ensuring retrieval accuracy and efficiency, and perfecting the indexing solution that integrates high-dimensional and two-dimensional data.

[0102] Example 4 Based on the same inventive concept, this embodiment provides a geological data indexing system for implementing the geological data indexing method described above. The solution provided by this system is similar to the implementation scheme described in the above method; therefore, the specific limitations of the one or more geological data indexing system embodiments provided below can be found in the limitations of the geological data indexing method described above, and will not be repeated here.

[0103] This embodiment provides a geological data indexing system, see reference. Figure 5 It includes: a data input module, a high-dimensional feature mapping module, a spatial index construction module, a semantic index construction module, a two-level mapping coordination module, an index storage module, a query interface module, a query parsing module, a parallel query execution module, a result processing module, and a knowledge model output module.

[0104] (a) Functions of each module.

[0105] 1. Data Input Module: This module utilizes the feature vector library interface, the geological knowledge topic library interface, and data validation and preprocessing components. Function Description: It reads high-dimensional feature vectors from geological object data sets from the feature vector library and obtains quantified topic tags from the geological knowledge topic library, ensuring data quality and format consistency.

[0106] 2. High-dimensional feature mapping module: Through feature standardization, similarity matrix calculation, and low-dimensional space optimization, high-dimensional feature vectors are mapped to a two-dimensional coordinate space.

[0107] 3. Spatial Index Construction Module: A dynamic two-dimensional spatial index is constructed based on the RBush algorithm using R-Tree structure, dynamic spatial partitioning, and balancing algorithm.

[0108] 4. Semantic Index Building Module: Constructs a global topic index through inverted index, weight threshold filtering, and index compression.

[0109] 5. Two-level mapping coordination module: Through mapping relationship establishment, consistency maintenance, and association optimization, it coordinates spatial indexes and semantic indexes to establish a two-level mapping relationship.

[0110] 6. Index storage module: Persistently stores various index structures and related data, supporting efficient read and write, data compression, and fast retrieval.

[0111] 7. Query Interface Module: Provides a unified query access point and receives user query requests. 8. Query parsing module: Parses mixed query requests and breaks them down into executable query components by parsing spatial and semantic conditions and identifying logical relationships.

[0112] 9. Parallel Query Execution Module: Through thread pool management, load balancing, and query optimization, it concurrently executes spatial and semantic queries.

[0113] 10. Result Processing Module: Processes query results by merging result sets, sorting by relevance, and formatting results, and performs aggregation and validation.

[0114] 11. Knowledge Model Output Module: Integrates query results, constructs a complete semantic knowledge model of geological objects, and performs structured knowledge representation to support further analysis and application.

[0115] (ii) Collaboration between modules.

[0116] 1. Data flow during the construction phase: Data input → Feature mapping → (Spatial index + Semantic index) → Two-level mapping → Index storage.

[0117] 2. Data flow during the query phase: Query interface → Query parsing → Parallel query → Index storage → Result processing → Knowledge output.

[0118] 3. Feedback optimization mechanism: Query results can be fed back to the index building process to achieve adaptive optimization of the index structure.

[0119] The beneficial effects of this application are: The advantage of this application lies in its use of three core technologies—high-dimensional spatialization, dynamic indexing, and two-level mapping—to construct an efficient, flexible, and deeply integrated knowledge and data indexing architecture for geological data. This addresses the key shortcomings of existing technologies, such as limited retrieval dimensions, low efficiency, and inability to effectively link implicit knowledge. Specifically: 1. It achieves "global second-level positioning and local millisecond-level query".

[0120] Dynamic spatial indexing based on the RBush algorithm avoids the uneven data distribution problem of traditional static grids. It supports parallel query engines to process spatial and semantic conditions simultaneously, solving the efficiency bottleneck caused by sequentially querying multiple independent indexes in traditional methods.

[0121] 2. Supports hybrid search based on "spatial location + semantic topic".

[0122] In response to the high-dimensional nature of geological survey data, a two-level mapping mechanism of "topic-knowledge base" is used to associate the global topic index with the local spatial grid, thereby enabling joint retrieval of space and content.

[0123] The RBush algorithm's dynamic partitioning and tree balancing mechanism can dynamically adapt to complex and ever-changing geological data, and enable the spatial index to adaptively adjust as data is added or deleted without reconstruction. This overcomes the shortcomings of fixed grid indexes, which perform poorly in both sparse and dense areas of geological data, resulting in more efficient resource utilization.

[0124] 4. Unify "quantitative data characteristics" and "qualitative document knowledge" under the same architecture.

[0125] By using high-dimensional feature spatialization mapping, abstract attributes are transformed into indexable spatial points; through two-level mapping, document topics are bound to spatial objects; the barriers between data and knowledge are broken down, realizing the leap from "data retrieval" to "knowledge retrieval", and providing a structured knowledge foundation for intelligent analysis.

[0126] Example 5 This embodiment provides a computer device, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 6As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores geological object data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a geological data index construction method or a geological data index usage method.

[0127] Those skilled in the art will understand that Figure 6 The structures shown are merely block diagrams of some structures related to the present application and do not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than shown in the figures, or combine certain components, or have different component arrangements. In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0128] Example 6 This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0129] Example 7 This embodiment provides a computer program product, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0130] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Moreover, the collection, use and processing of the relevant data are carried out in compliance with the relevant data protection laws and policies of the country where the location is located, and with the authorization granted by the owner of the corresponding device.

[0131] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0132] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0133] 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.

[0134] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for constructing a geological data index, characterized in that, include: A geological data feature vector library and a geological knowledge topic library are acquired; the geological data feature vector library is a collection of high-dimensional feature vectors corresponding to several geological object data groups; the geological knowledge topic library includes several geological knowledge topics. The geological data feature vector library is dimensionality reduced using a dimensionality reduction algorithm to obtain a two-dimensional coordinate set; the two-dimensional coordinate set includes several two-dimensional coordinates; the two-dimensional coordinates and the geological object data group have a one-to-one correspondence. Using the RBush algorithm, construct a spatial index tree structure corresponding to the two-dimensional coordinate set; A semantic index structure is constructed based on the topic attributes of each geological object data group corresponding to the geological knowledge topic database and the geological data feature vector database.

2. The geological data indexing construction method according to claim 1, characterized in that, The dimensionality reduction algorithm is the t-SNE algorithm; the specific process of using the t-SNE algorithm to perform dimensionality reduction on the geological data feature vector database is as follows: Each high-dimensional feature vector in the geological data feature vector library is standardized to obtain a standardized feature vector library. Calculate the conditional probability matrix corresponding to the standardized feature vector library, and perform symmetry processing on the conditional probability matrix to obtain the joint probability matrix; A point set is randomly selected in a two-dimensional space, and the joint probability matrix of similarity corresponding to the point set is calculated; the number of points in the point set is equal to the number of high-dimensional feature vectors in the geological data feature vector library. With the goal of minimizing the KL divergence between the joint probability matrix corresponding to the standardized feature vector library and the joint probability matrix of similarity corresponding to the point set, the coordinates of the points in the point set are iteratively optimized to obtain the optimal set of points; the optimal set of points is determined as the two-dimensional coordinate set.

3. The geological data indexing construction method according to claim 2, characterized in that, For any row in the conditional probability matrix, the elements in that row represent the conditional probabilities of a standardized eigenvector and itself or any other standardized eigenvector; where... A standardized feature vector has a conditional probability of 0 for itself. The conditional probability of a standardized eigenvector corresponding to any other standardized eigenvector is calculated using a conditional probability calculation model; the mathematical expression of the conditional probability calculation model is: in, Represented by standardized feature vectors When centered, the standardized feature vector The conditional probability of being selected as a neighbor; standardized feature vector To divide the standardized feature vector Any normalized feature vector other than the one mentioned above; symbol Indicates Euclidean distance; Represented by standardized feature vectors The standard deviation of a Gaussian distribution centered at .

4. The geological data indexing construction method according to claim 2, characterized in that, For any row in the joint similarity probability matrix, the elements in the row represent the joint similarity probability between a point and itself or any other point; where... The joint probability of a point's similarity to itself is 0. The joint probability of similarity between a point and any other point is calculated using a joint similarity probability calculation model; the mathematical expression of the joint similarity probability calculation model is: in, Point and points The joint probability of similarity; Point The coordinates; Point The coordinates; Point The coordinates; Point Coordinates of the point; ,point ,point and points All are points in the set of points; symbol This represents the Euclidean distance.

5. The geological data indexing construction method according to claim 1, characterized in that, Using the RBush algorithm, a spatial index tree structure corresponding to a set of two-dimensional coordinates is constructed, specifically including: Generate the minimum bounding rectangle for each point in the two-dimensional coordinate set; Using the minimum bounding rectangle corresponding to each point in the two-dimensional coordinate set as input, a spatial index tree structure is constructed using the RBush algorithm according to a preset insertion strategy and preset node splitting rules. The spatial index tree structure includes leaf nodes and internal nodes. The internal nodes include the root node and several intermediate nodes between the root node and any leaf node. Each leaf node includes the identifier of several geological object data groups and the minimum bounding rectangle. Each internal node includes the corresponding minimum bounding rectangle and the addresses of all its next-level nodes. The minimum bounding rectangle corresponding to any internal node is the smallest rectangle that includes the minimum bounding rectangles corresponding to all its next-level nodes.

6. The geological data indexing construction method according to claim 5, characterized in that, The preset insertion strategy specifically includes: For a minimum bounding rectangle to be inserted, starting from the root node, recursively select the insertion path with the goal of minimizing the area expansion of the minimum bounding rectangle corresponding to the node after insertion, until the leaf node is reached. Insert the minimum outer rectangle into the leaf node, and update all internal nodes on the path to which the leaf node belongs.

7. The geological data indexing construction method according to claim 6, characterized in that, The preset node splitting rules specifically include: When the number of identifiers in the inserted leaf node exceeds the maximum capacity, all identifiers in the inserted leaf node are divided into two groups according to the preset fill degree. For each group, a corresponding leaf node is created to replace the inserted leaf node, and all internal nodes on the corresponding path are updated. If the number of next-level node addresses in any internal node exceeds the maximum capacity, all next-level node addresses in the internal node are divided into two groups according to a preset fill factor. For each group, a corresponding internal node is created to replace the internal node.

8. The geological data indexing construction method according to claim 1, characterized in that, Based on the topic attributes of each geological object data group corresponding to the geological knowledge topic database and the geological data feature vector database, a semantic index structure is constructed, specifically including: Based on the theme attributes of the geological object data group, for each geological knowledge theme in the geological knowledge theme database, the identifier of the geological object data group that includes the corresponding geological knowledge theme and whose corresponding weight exceeds a preset threshold is retrieved; wherein, the weight represents the degree of correlation between the geological object data group and the geological knowledge theme; A semantic index structure is constructed by using each geological knowledge topic as the key and the list of identifiers of all geological object data groups retrieved based on a geological knowledge topic as the value.

9. A method for using geological data indexing, characterized in that, include: The query request is obtained and parsed to obtain spatial and semantic conditions; the spatial conditions include the query area; the semantic conditions include the query geological knowledge topic. The spatial index tree structure and the semantic index structure are retrieved based on spatial and semantic conditions, respectively, to obtain a first set of geological object identifiers and a second set of geological object identifiers; the spatial index tree structure and the semantic index structure are determined by the geological data indexing construction method according to any one of claims 1-8; The final geological object identifier set is determined based on the intersection of the first geological object identifier set and the second geological object identifier set. Based on the final set of geological object identifiers, retrieve the corresponding geological object data set from the geological object database.

10. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that the processor executes the computer program to implement the geological data index construction method of any one of claims 1-8, or the geological data index usage method of claim 9.

Images