Methods, systems, media and equipment for the full lifecycle management of multi-source geoscience data

Abstract

This application belongs to the field of information management technology and discloses a method, system, medium, and device for full lifecycle management of multi-source geoscientific data. The method includes: adaptively acquiring raw multi-source geoscientific data based on the geological structural complexity of the target area; preprocessing the raw multi-source geoscientific data to generate a standard dataset; constructing an adaptive grid system, achieving spatial adaptive management through grid splitting and merging; constructing a three-dimensional hybrid coding system based on the grid, assigning a unique identifier to each data entry; implementing dual-scale hierarchical storage and constructing a hybrid index based on the coding system; and performing value assessment, consistency verification, and archiving migration on this basis, feeding back repair experience to the acquisition strategy optimizer to form a closed-loop adaptive management system. This application can achieve closed-loop adaptive management of multi-source geoscientific data from acquisition to archiving, thereby improving the overall efficiency and intelligence level of data acquisition, storage, retrieval, and optimization.

Description

Technical Field

[0001] This application belongs to the field of information management technology, specifically involving methods, systems, media and equipment for the full lifecycle management of multi-source geoscientific data. Background Technology

[0002] Existing multi-source geoscientific data management methods typically lack a systematic consideration of the entire data lifecycle. In particular, during the data acquisition phase, they fail to adaptively differentiate sampling based on the spatial heterogeneity of geological structural complexity, resulting in insufficient sampling in high-complexity areas and redundant sampling in low-complexity areas. At the data storage and management level, existing methods mostly adopt fixed grid or uniform partitioning strategies, making it difficult to balance local details with global efficiency. At the same time, they lack a unified multi-dimensional coding system and hybrid indexing mechanism, which limits the rapid retrieval and fusion analysis of cross-source data. Furthermore, existing methods do not pay enough attention to post-management stages such as data value assessment, consistency verification, and archiving migration, making it difficult to form a closed-loop optimization from acquisition to archiving. As a result, the quality and efficiency of data management cannot be continuously improved with the accumulation of experience. Summary of the Invention

[0003] To address the shortcomings of existing technologies, the purpose of this application is to provide a method, system, medium, and device for full lifecycle management of multi-source geoscientific data. This application can realize closed-loop adaptive management of multi-source geoscientific data from acquisition to archiving, and improve the overall efficiency and intelligence level of acquisition, storage, retrieval, and optimization while ensuring data representativeness.

[0004] To achieve the above objectives, this application provides the following technical solution: A method for full lifecycle management of multi-source geoscientific data, comprising: adaptively and differentially collecting multi-source geoscientific data based on the geological structural complexity of the target area to obtain an original multi-source geoscientific dataset matching the geological structural complexity of the target area; preprocessing the multi-source geoscientific data in the original multi-source geoscientific dataset to generate a standard multi-source geoscientific dataset; constructing an adaptive grid system for the target area based on the standard multi-source geoscientific dataset, and achieving spatial adaptive management of the preprocessed multi-source geoscientific data through grid splitting and merging; and constructing a three-dimensional adaptive grid system based on the adaptive grid system for the target area. A 3D hybrid coding system is used to assign a unique identifier to each piece of multi-source geoscience data in the standard multi-source geoscience dataset. Based on the 3D hybrid coding system, dual-scale hierarchical storage is implemented for each piece of multi-source geoscience data in the standard multi-source geoscience dataset, and a hybrid index is constructed. On the basis of hierarchical storage and hybrid index, multi-dimensional value assessment is performed on each piece of multi-source geoscience data in the standard multi-source geoscience dataset, and hierarchical storage and archiving migration are implemented based on the value score. At the same time, consistency verification between adjacent grids is performed, inconsistency problems are repaired in a hierarchical manner, and the repair experience is fed back to the collection stage to form a closed-loop adaptive management from collection to archiving.

[0005] Optionally, the step of adaptively and differentially collecting multi-source geoscientific data based on the geological structural complexity of the target area to obtain an original multi-source geoscientific dataset that matches the geological structural complexity of the target area includes: loading prior geological maps, satellite gravity and magnetic data of the target area, and calculating local geological structural entropy distribution maps; dividing the target area into regions of different complexity according to the local geological structural entropy distribution maps, and assigning differentiated initial sampling densities to each complexity region; collecting multi-source geoscientific data in real time according to the assigned differentiated initial sampling densities, and calculating the local coefficient of variation for each sampling point; determining whether the local coefficient of variation exceeds a set threshold, and if it exceeds the set threshold, recommending the next optimal sampling point location and sending its coordinates to the sampling stage for data supplementation until the local coefficient of variation falls back below the set threshold; summarizing all collected data and outputting an original multi-source geoscientific dataset with LSE labels and sampling density levels.

[0006] Optionally, the set threshold is an adaptive threshold, calculated using the following formula:

[0007] in, This indicates that a threshold value has been set. This represents the average local geological tectonic entropy of the target area; , This represents the empirical coefficient.

[0008] Optionally, the preprocessing of the multi-source geoscientific data in the original multi-source geoscientific dataset includes the following steps: format unification, field mapping, and missing value interpolation filling of the multi-source geoscientific data to output a standard multi-source geoscientific dataset; physical consistency check of the standard multi-source geoscientific dataset; statistical anomaly screening of the multi-source geoscientific data in the standard multi-source geoscientific dataset that passed the physical consistency check to identify statistical outliers; and marking the data in the standard multi-source geoscientific dataset that failed the physical consistency check as suspicious physical anomalies and performing in-depth checks.

[0009] Optionally, the step of constructing an adaptive grid system for the target region based on a standard multi-source geoscientific dataset, and achieving spatial adaptive management of the preprocessed multi-source geoscientific data through grid splitting and merging, includes: performing edge detection on the preprocessed multi-source geoscientific data to identify geoscientific boundaries including lithological abrupt change zones, fault lines, and unconformities; embedding the identified geoscientific boundaries as hard constraints into the initial grid partitioning framework and generating an initial constraint grid; calculating the multi-source information entropy of each grid cell within the initial constraint grid to assess the degree of joint variation of multiple variables such as lithology, geochemical exploration, and geophysical exploration; splitting grid cells with multi-source information entropy exceeding the splitting threshold using a quadtree or octree, and merging grid cells with multi-source information entropy below the merging threshold and consistent with the geological semantics of adjacent grids; outputting a non-uniform adaptive grid system and establishing a grid metadata table to record the level, boundary coordinates, geological labels, and entropy value of each grid.

[0010] Optionally, the calculation of the multi-source information entropy of each grid cell within the initial constraint grid includes: extracting multivariate data of all sampling points contained in the grid cell, including discrete variables and continuous variables; calculating discrete information entropy based on discrete variables, and calculating continuous information entropy based on continuous variables; and obtaining multi-source information entropy based on discrete information entropy and continuous information entropy.

[0011] Optionally, the construction of a three-dimensional hybrid coding system based on the target region adaptive grid system to assign a unique identifier to each piece of multi-source geoscience data in the standard multi-source geoscience dataset includes: constructing a geological time-level coding system and assigning a fixed-width code to each level; establishing a standard lithology coding table based on the geological time-level coding system and assigning fixed-length codes to igneous rocks, sedimentary rocks, metamorphic rocks and their subclasses; and generating a variable-length prefix code for each grid cell using a space-filling curve, based on the target region adaptive grid system, as the spatial grid code. Each piece of multi-source geoscientific data is assigned a version number and source code to distinguish the data source, collection batch, and processing history; the above four types of codes are combined to generate a three-dimensional hybrid code, which is assigned a unique identifier to each piece of data and stored in a unified index table.

[0012] This application also provides a multi-source geoscientific data lifecycle management system, the system comprising: a data acquisition module, used to adaptively and differentially acquire multi-source geoscientific data based on the geological structural complexity of the target area, to obtain an original multi-source geoscientific dataset matching the geological structural complexity of the target area; a preprocessing module, used to preprocess the multi-source geoscientific data in the original multi-source geoscientific dataset to generate a standard multi-source geoscientific dataset; a grid construction module, used to construct an adaptive grid system for the target area based on the standard multi-source geoscientific dataset, and to achieve spatial adaptive management of the preprocessed multi-source geoscientific data through grid splitting and merging; and a coding system construction module, used to construct a coding system based on the target area. The adaptive grid system constructs a three-dimensional hybrid coding system to assign a unique identifier to each piece of multi-source geoscience data in the standard multi-source geoscience dataset. The hybrid index construction module is used to implement dual-scale hierarchical storage and construct a hybrid index for each piece of multi-source geoscience data in the standard multi-source geoscience dataset based on the three-dimensional hybrid coding system. The closed-loop adaptive management module is used to perform multi-dimensional value assessment on each piece of multi-source geoscience data in the standard multi-source geoscience dataset based on hierarchical storage and hybrid index, and implement hierarchical storage and archiving migration based on the value score. At the same time, it performs consistency verification between adjacent grids, repairs inconsistencies in a hierarchical manner, and feeds back the repair experience to the acquisition stage to form a closed-loop adaptive management from acquisition to archiving.

[0013] This application also provides a storage medium including instructions that, when executed on a computer, cause the computer to perform the method as described in the preceding claim.

[0014] This application also provides an electronic device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the method as described in any of the preceding claims.

[0015] Compared with the prior art, the beneficial effects of this application are as follows: This application enables intelligent management of multi-source geoscientific data throughout its entire lifecycle through adaptive acquisition based on geological structural complexity, adaptive grid construction, three-dimensional hybrid coding, dual-scale hierarchical storage and hybrid indexing, as well as a closed-loop value assessment and feedback mechanism. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating a method for managing the entire lifecycle of multi-source geoscientific data according to an embodiment of this application; Figure 2 This is a schematic diagram of the structure of a multi-source geoscience data lifecycle management system provided in another embodiment of this application; Figure 3 This is a schematic diagram of the structure of a storage medium provided in another embodiment of this application; Figure 4 This is a schematic diagram of the structure of an electronic device provided in another embodiment of this application. Detailed Implementation

[0017] Specific embodiments of this application will now be described in detail with reference to the accompanying drawings. While specific embodiments of this application are shown in the drawings, it should be understood that this application can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of this application and to fully convey the scope of this application to those skilled in the art.

[0018] To facilitate understanding of the embodiments of this application, further explanations and descriptions will be provided below with reference to the accompanying drawings and specific embodiments. The accompanying drawings do not constitute a limitation on the embodiments of this application.

[0019] Figure 1 This is a flowchart illustrating a method for managing the entire lifecycle of multi-source geoscientific data according to an embodiment of this application. Figure 1 As shown, the method includes the following steps: S100: Based on the geological structural complexity of the target area, adaptively differentiate the collection of multi-source geoscience data to obtain an original multi-source geoscience dataset that matches the geological structural complexity of the target area. S200: Preprocess the multi-source geoscience data in the original multi-source geoscience dataset to generate a standard multi-source geoscience dataset; S300: Based on standard multi-source geoscience datasets, construct an adaptive grid system for the target area, and achieve spatial adaptive management of preprocessed multi-source geoscience data through grid splitting and merging; S400: A three-dimensional hybrid coding system is constructed based on the target area adaptive grid system to assign a unique identifier to each piece of multi-source geoscience data in the standard multi-source geoscience dataset; S500: Based on the three-dimensional hybrid coding system, each piece of multi-source geoscience data in the standard multi-source geoscience dataset is stored in a dual-scale hierarchical manner and a hybrid index is constructed. S600: Based on hierarchical storage and hybrid indexing, it performs multidimensional value assessment on each piece of multi-source geoscience data in the standard multi-source geoscience dataset, and implements hierarchical storage and archiving migration according to the value score; at the same time, it performs consistency verification between adjacent grids, repairs inconsistencies in a hierarchical manner, and feeds back the repair experience to the collection stage to form a closed-loop adaptive management from collection to archiving.

[0020] This embodiment first employs adaptive data acquisition based on the geological structural complexity of the target area, improving acquisition efficiency while ensuring representative data. Building upon this, an adaptive grid system is constructed using grid splitting and merging, balancing local detail representation with overall organizational efficiency. This leads to the establishment of a three-dimensional hybrid coding system integrating geological age, lithology, spatial location, and source information, achieving unique data identification and rapid retrieval. Simultaneously, the use of dual-scale hierarchical storage and hybrid indexing, combined with association rule prefetching, enhances the storage and query performance of multi-source geoscientific data. Finally, regular value assessments and consistency checks, along with feedback of restoration experience to the acquisition strategy, form a closed-loop adaptive management system, continuously improving the intelligence level and reliability of multi-source geoscientific data throughout its entire lifecycle.

[0021] In another exemplary embodiment, step S100, which involves adaptively and differentially collecting multi-source geoscientific data based on the geological structural complexity of the target area to obtain an original multi-source geoscientific dataset that matches the geological structural complexity of the target area, includes the following steps: S101: Load the prior geological map of the target area, satellite gravity and magnetic data, and calculate the local geological entropy (LSE) distribution map; In this step, this embodiment loads the prior geological map, satellite gravity data, and magnetic data of the target area as initial inputs. Based on this multi-source prior information, a sliding window is used to calculate the geological structural complexity within each local window, resulting in a local geological structural entropy (LSE) distribution map. Specifically, the LSE calculation using a sliding window includes the following steps: First, a fixed-size window (e.g., 5km x 5km, adjustable according to the regional scale) and a sliding step size (e.g., 1km or 1 / 4 of the window size) are set, and the entire target area is traversed row by row and column by column. Next, within each window, three types of geological structural complexity features are extracted: lithological diversity, fault development degree, and geophysical field variation. Then, the local geological structural entropy within the window is calculated based on the following formula. :

[0022] in, This indicates the number of different geological conditions within the window; Indicates the first The frequency of geological states appearing within the window.

[0023] Finally, the local geological structural entropy corresponding to the center point of each sliding window is assigned to the geographic coordinates of that center point (i.e., the longitude and latitude of the window center, or the X and Y coordinates in the projected coordinate system), thus forming discrete point-like LSE data. Then, based on these discrete points, spatial interpolation methods such as inverse distance weighting and kriging are used to estimate the unsampled areas, thereby generating a continuous LSE raster distribution map covering the entire target area. This distribution map can reflect the spatial heterogeneity of geological structural complexity within the target area, such as lithological changes, fault development, and tectonic superposition relationships. Among them, the higher the entropy value, the more complex the structure of the target area, and subsequent sampling strategies will be designed accordingly.

[0024] S102: Based on the Local Structural Entropy (LSE) distribution map, the target area is divided into regions of different complexity, including high complexity regions (fault intersections, rock mass contact zones), medium complexity regions, and low complexity regions (homogeneous sedimentary basins), and differentiated initial sampling densities are assigned to different complexity regions. In this step, based on the calculated LSE distribution map, an LSE threshold is set to divide the target area into high-complexity, medium-complexity, and low-complexity regions. High-complexity regions correspond to geologically strongly deformed areas such as fault junctions and rock mass contact zones, with LSE values ​​typically greater than 0.7; medium-complexity regions correspond to moderately developed tectonic areas, with LSE values ​​between 0.3 and 0.7; and low-complexity regions correspond to structurally simple areas such as homogeneous sedimentary basins, with LSE values ​​less than 0.3.

[0025] Furthermore, this embodiment automatically assigns differentiated initial sampling densities for areas of varying complexity. For example, high-complexity areas are assigned the highest initial sampling density, such as 50-100 sampling points per square kilometer, or a grid spacing of 50-100 meters, to capture fine geological changes; medium-complexity areas are assigned a medium sampling density, such as 10-50 sampling points per square kilometer, or a grid spacing of 200-500 meters; and low-complexity areas are assigned the lowest sampling density, such as 1-10 sampling points per square kilometer, or a grid spacing of 500-2000 meters, to save on data collection costs. It should be noted that the above initial sampling densities are merely examples, and specific settings can be dynamically adjusted according to actual geological survey specifications, project accuracy requirements, and resource availability.

[0026] S103: Collect multi-source geoscientific data in real time according to the allocated differentiated initial sampling density, and calculate the local coefficient of variation for each sampling point; In this step, according to the differentiated initial sampling density allocated in step S102, sampling points are arranged within the target area, and multi-source geoscientific data (including but not limited to rock samples, geochemical data, geophysical data, etc.) are collected in real time. Furthermore, during the collection process, for each collected sampling point, neighboring points are selected centered on it or a neighborhood radius is set, and the local coefficient of variation for each observed variable (such as elemental content, geophysical values, etc.) within this local area is calculated. This coefficient is used to measure the relative fluctuation of data within the neighborhood of the sampling point. The larger the coefficient of variation, the more drastic the local geological changes, and the existing sampling density may be insufficient to characterize the true changes. The specific formula for calculating this local coefficient of variation is as follows:

[0027] in, Indicates the local coefficient of variation; This represents the sample standard deviation of the observed variable within its neighborhood. This represents the sample mean of the observed variable within the neighborhood.

[0028] S104: Determine whether the local coefficient of variation exceeds the set threshold. If it exceeds the threshold, recommend the next optimal sampling point location and send its coordinates to the sampling stage to perform data supplementation until the local coefficient of variation falls back below the threshold. In this step, this embodiment dynamically calculates the adaptive threshold of the local variation coefficient based on the geological complexity of the target area. The calculation formula is as follows:

[0029] in, This indicates that a threshold value has been set. This represents the average local geological tectonic entropy of the target area; , This represents the empirical coefficient.

[0030] It should be noted that for high complexity regions ( >0.7), threshold Automatically relaxed to 0.4–0.5 to tolerate the inherently high variability within this region; for medium complexity regions (0.3 ≤ ≤0.7), threshold Set it to 0.25~0.35; for the low complexity region ( <0.3), threshold Tighten to 0.15–0.20 to ensure data accuracy in low-tectonic-activity regions.

[0031] Furthermore, for the local coefficient of variation at each sampling point, if it exceeds a set threshold, it indicates that the geological changes in the neighborhood of that point are still significant, requiring additional sampling points to capture details. At this point, based on existing sampling data and spatial distribution, the system uses optimization algorithms (such as Kriging variance minimization and maximum uncertainty sampling) to recommend the location of the next optimal sampling point and sends its coordinates to the sampling stage to guide on-site data supplementation. After supplementation is completed, the local coefficient of variation in the affected area is recalculated. If it still exceeds the threshold, new supplementation points are recommended, and the above process is repeated until the local coefficient of variation in all areas falls below the threshold, thus ensuring an adaptive match between sampling density and geological structural complexity.

[0032] S105: Summarize all collected data and output the raw multi-source geoscience dataset with LSE labels and sampling density levels.

[0033] In this step, once all local coefficients of variation within the target area meet the threshold requirements, this embodiment summarizes all collected multi-source geoscientific data and adds metadata tags to each data record according to the following process: First, add local geological structure entropy labels: Based on the spatial coordinates of the sampling points, perform spatial interpolation or nearest neighbor matching on the LSE distribution map generated in step S101 to extract the LSE value at that coordinate location, and append this value as a floating-point field "LSE_value" to the data record. Simultaneously, based on the preset LSE classification threshold (high complexity: >0.7; Medium complexity: 0.3≤ ≤0.7; Low complexity: If the value is less than 0.3, the corresponding complexity level label "LSE_class" will be automatically generated, with a value of "high", "medium" or "low".

[0034] Secondly, add sampling density level labels: based on the final sampling density implemented in the area where the sampling points are located (i.e., the ratio of the actual number of sampling points deployed to the area area), compare it with the differentiated initial sampling density range set in step S102 (high density: 50-100 points / km). 2 Medium density: 10–50 points / km 2 Low density: 1-10 points / km 2 The system determines and generates a sampling density level label "density_class", with values ​​of "high_density", "medium_density", or "low_density". If an area experiences a density jump due to resampling (e.g., an area that was originally low-density but has a density exceeding 10 points / km after resampling), the system will take appropriate action. 2If the density is not found, the level will be re-determined based on the actual density, and the comparison fields “density_original” and “density_final” between the original and final levels will be retained to record the dynamic adjustment process of the sampling strategy.

[0035] Secondly, supplementary auxiliary metadata fields are added: In order to facilitate subsequent data management and traceability, the following auxiliary tags are added in this embodiment: (1) “sampling_stage”, which marks whether the data comes from the initial collection or the Nth round of supplementary collection; (2) “local_CV”, which records the local coefficient of variation value calculated by the sampling point; (3) “neighbor_count”, which records the number of neighboring points used when calculating the local coefficient of variation; (4) “acquisition_time”, which records the timestamp of data collection; (5) “quality_flag”, which marks the data quality status (such as “normal”, “verification passed after supplementary collection”, etc.).

[0036] Finally, the raw collected data and the aforementioned metadata tags are merged in a row-column structure to form a unified relational table or a columnar storage format that supports nested fields (such as Parquet, HDF5, or GeoJSON). Each record contains the original observation fields (such as element content, geophysical values, lithology codes, etc.) as well as the newly added metadata tag fields mentioned above.

[0037] In another exemplary embodiment, step S200, the preprocessing of the multi-source geoscience data in the original multi-source geoscience dataset, includes the following steps: S201: Perform format standardization, field mapping, and missing value interpolation on multi-source geoscience data to output a standard multi-source geoscience dataset; In this step, since data from different sources often use different file formats (such as text, CSV, Excel, Shapefile, NetCDF, etc.), coordinate systems (such as latitude and longitude, projected coordinates), and units of measurement (such as meters, kilometers, percentages, ppm, etc.), this embodiment first converts all data into an internal standard format (such as Parquet or HDF5). Simultaneously, it sets the coordinate system to the WGS84 geographic coordinate system or a specified projected coordinate system, and converts the units of measurement for all observations to standard units (such as unifying elemental content to ppm or %). Next, a field mapping operation is performed, which establishes a mapping relationship between source data fields and standard data model fields. For example, different named fields in the source data, such as "Cu_ppm" and "copper content," are uniformly mapped to the target field "Cu_concentration_ppm". For fields that cannot be automatically mapped, they are marked and require manual confirmation.

[0038] After standardizing the format and mapping the fields, the missing values ​​are further filled by interpolation. For spatially continuous data (such as geophysical and geochemical data), inverse distance weighted interpolation or kriging interpolation is used to estimate the missing values. For categorical data (such as lithology labels), mode filling or majority voting based on neighborhood points is used. For fields with a missing rate of more than 40%, no filling is performed, but they are marked as "missing and unavailable" and an alarm record is generated.

[0039] S202: Perform a physical consistency check on the standard multi-source geoscience dataset; The purpose of performing a physical consistency check on the standard multi-source geoscientific dataset in this step is to identify and remove data anomalies that violate basic physical laws or geological common sense. Specifically, this includes the verification of the following three types of constraint rules: First, numerical range constraints, for example, rock density values ​​should be between 1.0 and 6.0 g / cm³. 3 Within the specified range, the magnetic susceptibility should be non-negative, and the element content should not be negative. If these physical values ​​exceed their upper and lower limits, it is considered a violation of physical consistency. Second, logical correlation constraints: for example, the silicon and magnesium contents at the same sampling point should conform to the general trend of magmatic rock evolution (e.g., MgO usually decreases when SiO2 increases). If a combination of SiO2 > 70% and MgO > 15% occurs, it is considered logically unreliable. Third, spatial continuity constraints: the variation in observed values ​​between adjacent sampling points should not exceed a geologically reasonable gradient threshold. For example, if a gravity anomaly shows a jump of more than 500 μGal over a short distance, it is considered a spatially discontinuous anomaly.

[0040] Each record in the standard multi-source geoscience dataset is scanned one by one. Data records that do not meet any of the above constraint rules are marked as "failed physical consistency check" and the specific constraint type and violation value are recorded. Data records that pass all physical consistency checks are marked as "passed" and proceed to the next step for statistical anomaly screening.

[0041] S203: Perform statistical outlier screening on multi-source geoscience data that have passed the physical consistency check in the standard multi-source geoscience dataset to identify statistical outliers; In this step, this embodiment groups the data according to geological zoning (e.g., high-complexity, medium-complexity, and low-complexity zones are processed separately) or according to lithological categories. For continuous variables (e.g., element content, density, magnetic susceptibility, etc.) within each group, multiple statistical methods are used to jointly identify outliers. Specifically, this includes: first, calculating the quartiles (Q1, Q3) and interquartile range (IQR) of each variable within each group; marking values ​​less than Q1-1.5×IQR or greater than Q3+1.5×IQR as mild outliers, and values ​​less than Q1-3×IQR or greater than Q3+3×IQR as severe outliers; second, calculating the mean and standard deviation of each variable within each group; marking sample points that deviate from the mean by more than 3 times the standard deviation as statistical outliers; for variables that conform to a normal distribution, iterative outlier detection is performed using the Grubbs test or Dixon test until no new outliers are identified. For categorical variables (such as lithology labels), the system calculates the frequency of occurrence of each lithology category and marks categories with frequencies below a preset threshold (such as 1% of the total sample size) and geologically rare as low-frequency categories.

[0042] Finally, the above statistical anomaly identification results are summarized, and a "statistical anomaly flag" field is generated for each data point (with values ​​of "normal", "mild outlier", "severe outlier" or "low-frequency category"). Detailed diagnostic information of statistical outliers (including the test method, statistic and critical value used) is recorded in the metadata for reference during subsequent manual review or automatic processing.

[0043] S204: Data in the standard multi-source geoscience dataset that fails the physical consistency check is marked as suspicious physical anomalies and subjected to in-depth inspection.

[0044] In this step, for data records marked as "failed physical consistency check" in step S202, this embodiment does not directly remove them, but marks them as "suspicious physical anomalies" and uses in-depth checks to distinguish between genuine geological anomalies and pseudo-anomalies caused by data collection. The specific steps of the in-depth check are as follows: First, the geological process simulator is invoked, taking the local area where the suspicious data point is located as the scope, and inputting information such as the geological boundaries, lithological distribution, and fault structures within the area. Forward modeling methods (such as gravity forward modeling, magnetic forward modeling, heat flow simulation, or geochemical diffusion model) are used to perform reproducibility checks on the suspicious observation values. For example, for an abnormally high density value appearing at a sampling point, the geological process simulator attempts to simulate whether there is a high-density rock mass (such as an ultramafic rock intrusion) or a mineralized enrichment zone at that location. If the anomaly value can be reproduced through simulation, it is determined to be a "genuine geological anomaly," retained, and labeled "genuine anomaly_geologically reproducible"; if the simulation cannot reproduce the anomaly value (for example, the simulated theoretical density value is much lower than the measured value), it is determined to be a "spurious data collection anomaly." For data identified as false anomalies, the system does not delete them directly. Instead, it uses valid data points in the neighborhood to perform interpolation and reconstruction along the geoscientific boundary direction (such as the strike of the fault zone or the extension direction of the lithological contact zone). The system calculates the correction value at that location using anisotropic kriging or geological-guided interpolation methods, and marks the original data as "false anomaly_reconstructed". At the same time, it records the comparison information between the original value and the reconstructed value.

[0045] After completing the above in-depth checks, the output dataset is a filtered and processed dataset. This dataset contains data that has passed the physical consistency check and is statistically normal, confirmed geological anomaly data, and pseudo-anomaly correction data that has been reconstructed by interpolation. All processing records retain complete source information so that subsequent steps can trace the data processing history.

[0046] In another exemplary embodiment, step S300, which involves constructing an adaptive grid system for the target region based on a standard multi-source geoscience dataset and achieving spatial adaptive management of the preprocessed multi-source geoscience data through grid splitting and merging, includes the following steps: S301: Perform edge detection on preprocessed multi-source geoscientific data to identify geoscientific boundaries, including lithological abrupt change zones, fault lines, and unconformities; In this step, the standard multi-source geoscientific dataset of this embodiment is used as input. Various edge detection algorithms are employed to identify geoscientific boundaries within the target area. For continuous spatial data (such as gravity anomalies, magnetic anomalies, and topographic data), the Canny edge detection operator or the Laplacian of Gaussian operator is used to extract areas with drastic gradient changes. For discrete data (such as lithological classification maps and geological time maps), the region growing method or watershed segmentation algorithm is used to identify the boundaries between different geological units. Simultaneously, prior geological map information is integrated, and known fault lines, rock mass contact zones, unconformities, and other geoscientific boundaries are used as soft constraints. These are then weighted and fused with the data-driven edge detection results to ultimately output a set of geologically significant geoscientific boundary lines, including lithological abrupt change zones, fault lines, unconformities, and tectonic contact zones.

[0047] S302: Embed the identified geoscience boundaries as hard constraints into the initial mesh generation framework and generate the initial constraint mesh; In this step, the geological boundary lines identified in step S301 are used as hard constraints and embedded into the initial mesh generation process. Specifically, this includes: first, setting the initial mesh size based on the overall range of the target area (e.g., using 1km×1km initial mesh cells); then, using constrained Delaunay triangulation or constrained quadtree meshing algorithms to generate the mesh, ensuring that all identified geological boundary lines are preserved as edges or faces of mesh cells, meaning that mesh cells on both sides of the boundary do not cross the boundary. For fault lines, discontinuous mesh cells are generated on both sides of the fault to reflect the fault's faulting effect; for unconformities, the crossing of mesh cells is restricted vertically. After completing the above constrained mesh generation, an initial constrained mesh is generated. This mesh fully respects the true location of the geological boundary in space, providing a basic framework for subsequent adaptive splitting and merging.

[0048] S303: Calculate the multi-source information entropy of each grid cell within the initial constrained grid to assess the degree of joint variation of multiple variables such as lithology, geochemistry, and geophysics; This step involves calculating the multi-source information entropy of each grid cell within the initial constrained grid. Specifically, it includes the following steps: First, extracting multivariate data from all sampling points within the grid cell, including discrete variables (such as lithology type) and continuous variables (such as geochemical element content) and geophysical attribute values ​​(such as density, magnetic susceptibility, resistivity, etc.). For discrete variables (lithology), the frequency of occurrence of each lithology type within the grid cell is calculated. And calculate the discrete information entropy:

[0049] in, For discrete information entropy; This represents the total number of lithology categories within a grid cell; For the first The frequency of lithological types occurring within the grid cell. For continuous variables, the variable values ​​are divided into several equal-frequency intervals, and the sample frequency within each interval is calculated. Then, calculate the continuous information entropy based on the sample frequency:

[0050] in, It is continuous information entropy; The number of equal-frequency intervals; For the first Sample frequencies within a single frequency interval.

[0051] Ultimately, based on discrete and continuous information entropy, multi-source information entropy can be obtained:

[0052] in, Entropy of multi-source information; These are the weighting coefficients for discrete information entropy; The weighting coefficients are the continuous information entropy.

[0053] S304: Perform quadtree or octree splits on grid cells with multi-source information entropy exceeding the splitting threshold, and merge grid cells with multi-source information entropy below the merging threshold and consistent with the geological semantics of adjacent grids. In this step, this embodiment performs adaptive grid splitting and merging operations based on the multi-source information entropy of each grid cell calculated in step S303. Specifically, for grid cells with information entropy exceeding a preset splitting threshold (e.g., ...), ... Mesh cells with an entropy greater than 0.7 are split. For 2D meshes, a quadtree split (dividing a single mesh into four sub-mesh cells) is used, while for 3D meshes, an octree split (dividing into eight sub-mesh cells) is used. After splitting, the multi-source information entropy of the newly generated sub-mesh cells is recalculated. If it still exceeds the splitting threshold, iterative splitting continues until a preset minimum mesh size (e.g., 10m × 10m) is reached or the information entropy drops below the threshold. Furthermore, for cells with an information entropy lower than a preset merging threshold (e.g., ... Grid cells with a density <0.2 and adjacent grid cells with consistent geological semantics (i.e., the same main lithology and geological age) are merged into a larger grid cell to reduce data redundancy. It is important to note that this merging operation must meet the following conditions: the merged grid cell cannot exceed the preset maximum grid size (e.g., 5km × 5km), and the merged cell will not cross any geological boundaries.

[0054] This embodiment repeats the above splitting and merging operations until the information entropy of all grid cells falls between the splitting threshold and the merging threshold, or the iterative convergence condition is met.

[0055] S305: Outputs a non-uniform adaptive grid system and establishes a grid metadata table, recording the level, boundary coordinates, geological label, and entropy value of each grid.

[0056] In this step, after completing all the adaptive splitting and merging operations described in step S304, the final non-uniform adaptive grid system is output. This grid system has the following characteristics: in high-complexity areas (such as fault zones and rock mass contact zones), the grid is dense and fine, which can accurately depict geological changes; in low-complexity areas (such as homogeneous sedimentary basins), the grid is sparse and coarse, reducing data redundancy. At the same time, a grid metadata table is established to record the following key information for each grid unit: (1) Grid unique identifier (Grid_ID); (2) Grid level (Level), reflecting the depth of the grid in the quadtree or octree; (3) Boundary coordinates (Boundary_Coords), recording the corner coordinates or the bounding rectangle range of the grid; (4) Geological label (Geo_Label), including the dominant lithology, geological age, tectonic unit affiliation, etc. of the grid; (5) Multi-source information entropy value (Entropy_Value), recording the information entropy finally calculated for the grid; (6) Parent grid and child grid pointers (Parent_Link and Child_Links), used to support dynamic query and fast positioning. This grid metadata table, together with the adaptive grid system, forms the spatial framework for subsequent steps.

[0057] In another exemplary embodiment, step S400, which involves constructing a three-dimensional hybrid coding system based on the target region adaptive grid system to assign a unique identifier to each piece of multi-source geoscience data in the standard multi-source geoscience dataset, includes the following steps: S401: Construct a geological time-level coding system and assign a fixed bit-width code to each level; In this step, this embodiment first establishes a standardized geological time-level coding system, including: dividing geological time into four levels—"Eon-System-Series-Era"—according to the International Stratigraphic Tables (ICS) standard, and assigning a fixed-width binary or integer code to each level. Specifically, the Eon level is assigned a 2-bit code (e.g., 00 for Phanerozoic Eon, 01 for Proterozoic Eon, 10 for Archean Eon, 11 for Hadean Eon); the System level is assigned a 3-bit code (e.g., 001 for Cenozoic, 010 for Mesozoic, 011 for Paleozoic); the Series level is assigned a 3-bit code (e.g., 001 for Pliocene, 010 for Miocene); and the Era level is assigned a 4-bit code (e.g., 0001 for Aptian Stage, 0010 for Albian Stage). For data that cannot be precisely graded, all zeros or all Fs are filled in the corresponding level as wildcards. Finally, a geological time code with a fixed total bit width (2+3+3+4=12 bits) is generated and used as the first dimension of the three-dimensional hybrid coding.

[0058] S402: Establish a standard lithology coding table based on the geological age hierarchical coding system, and assign fixed-length codes to igneous rocks, sedimentary rocks, metamorphic rocks and their subclasses; In this step, this embodiment uses the geological time-hierarchical coding system constructed in step S401 as the basic framework to establish a standard lithology coding table associated with it. Since the lithological assemblages of different geological time units vary considerably (for example, the Proterozoic is dominated by carbonate and clastic rocks, while the Cenozoic is dominated by loose sediments), this embodiment needs to consider the synergistic relationship between lithology coding and geological time coding when establishing lithology coding to avoid semantic conflicts. Specifically, this embodiment classifies rock types into three major categories: igneous rocks, sedimentary rocks, and metamorphic rocks, assigning a 2-digit code to each category (e.g., 01 represents igneous rocks, 10 represents sedimentary rocks, and 11 represents metamorphic rocks). Further subcategories are established, including: igneous rocks are classified according to SiO2 content into ultrabasic, basic, intermediate, and acidic rocks, each assigned a 2-digit code (01 to 04); sedimentary rocks are classified according to genesis into clastic, carbonate, evaporite, and siliceous rocks, each assigned a 2-digit code; metamorphic rocks are classified according to metamorphic degree into low-grade metamorphic, intermediate-grade metamorphic, high-grade metamorphic, and migmatite, each assigned a 2-digit code. Furthermore, for more specific rock types (such as granite, basalt, limestone, marble, etc.), a 3-digit code is assigned, ultimately forming a fixed-length lithology code of 7 digits: "major category (2 digits) + subcategory (2 digits) + specific lithology (3 digits)". Simultaneously, this embodiment establishes a compatibility constraint table between lithology codes and geological age codes. For example, the "carbonate rock" code has a high probability of correlation with the "Paleozoic" age code, while its combination with the "Archean" age code will trigger a consistency warning.

[0059] Ultimately, in this embodiment, the lithology code is used as the second dimension of the three-dimensional hybrid coding system, which together with the geological age code established in step S401 constitutes the semantic dimension of multi-source geoscience data.

[0060] S403: Based on the target region adaptive grid system, a variable-length prefix code is generated for each grid cell using a space-filling curve, which serves as the spatial grid code; In this step, based on the adaptive mesh system generated in step S300, this embodiment uses a space-filling curve (such as a Z-order curve or a Hilbert curve) combined with a tree structure encoding method to assign a unique mesh code with spatial positional relationship to each mesh cell. The specific encoding rules include: Rule 1: The code of each mesh cell is formed by concatenating the index values ​​of all nodes on its path from the root node to itself, where the root node corresponds to the entire target area and the encoding prefix is ​​empty; Rule 2: Each quadtree split generates 4 sub-mesh cells, which are assigned 2-bit binary codes (00, 01, 10, 11). For example, after the first split, the 0th sub-mesh is entered and the code is 00; after the second split, the 1st sub-mesh is entered and the code is 00-01; after the third split, the 2nd sub-mesh is entered and the code is 00-01-10; the final code is 00-01-10, the depth L=3, and the encoding bit width is 2×3=6 bits. Each octree split produces 8 subgrids, each assigned a 3-bit binary code (000 to 111). For example, after the first split, the 5th subgrid is coded as 101; after the second split, the 2nd subgrid is coded as 101-010; after the third split, the 7th subgrid is coded as 101-010-111; the final code is 101-010-111, with a depth L=3 and a code width of 3×3=9 bits. Rule 3: The code of a subgrid is formed by concatenating the code of the parent grid with the code of the current layer (codes at different levels can be connected by a separator (such as "-"). Rule 4: For a grid cell with a depth of L, its code width is 2L (quadtree) or 3L (octree), where the depth L is defined as the number of splits from the root node (L=0) to the target node. Meanwhile, this embodiment maintains a grid coding mapping table, which records the grid cell boundary coordinates, hierarchy, and parent-child relationships corresponding to each variable-length prefix code to support fast spatial query and positioning. For example, the grid coding mapping table is shown in Table 1: Table 1 Encoding Mapping Table

[0061] Ultimately, the generated variable-length prefix code serves as the third dimension of the three-dimensional hybrid coding system, directly reflecting spatial proximity and hierarchical subordination relationships in the coding process.

[0062] S404: Assign a version number and source code to each piece of multi-source geoscience data to distinguish the data source, collection batch, and processing history; In this step, this embodiment establishes a coding system for data version and source to support data traceability and multi-source fusion. The version / source code adopts a fixed bit width combination of "source institution (3 digits) + acquisition project (4 digits) + acquisition time (8-digit date YYYYMMDD) + processing version (2 digits)," with a total bit width of 17 digits. Among them, the "source institution" code is used to identify the unit or research group that produced the data (e.g., 001 represents the China Geological Survey, 002 represents a university laboratory); the "acquisition project" code is used to distinguish data from different projects or batches under the same institution (e.g., 0001 represents a 1:50,000 regional geological survey project); the "acquisition time" records the specific date of data acquisition or processing; and the "processing version" is used to identify the version number of the data after preprocessing (e.g., 01 represents the original preprocessed version, 02 represents the corrected and reconstructed version). For data acquired multiple times at the same location or fused from different sources, the system distinguishes them through this version / source code and supports data filtering and consistency comparison by source, time, and other dimensions in subsequent steps. This code serves as the fourth dimension of the three-dimensional hybrid coding system.

[0063] S405: Combine the above four types of codes to generate a three-dimensional hybrid code, assign a unique identifier to each data item, and store it in a unified index table.

[0064] In this step, this embodiment concatenates the codes of the above four dimensions in the order of "geological year code + lithology code + spatial grid code + version source code" to generate a unique three-dimensional hybrid coding identifier (UID) for each data entry. Among them, the geological year code and lithology code constitute the "semantic dimension" of the data, the spatial grid code constitutes the "spatial dimension" of the data, and the version source code constitutes the "time / origin dimension" of the data. The three together achieve a multi-dimensional unified identification of multi-source geoscientific data. Furthermore, to adapt to the query needs of different application scenarios, this embodiment supports both fixed-length and variable-length storage formats. When fast indexing and equivalent queries are required, the variable-length spatial grid code is mapped to a 32-bit fixed-length integer using a hash function, forming a fixed-length code with a fixed total bit width (12+7+32+17=68 bits). When spatial range queries and proximity retrieval are required, the mixed code in the form of variable-length prefix code is retained and stored in a unified index table. This index table uses UID as the primary key and records the original data storage location, grid unit ID, geological label, and metadata information corresponding to each data record, providing a retrieval entry point for the hierarchical storage and mixed indexing of step S500.

[0065] In another exemplary embodiment, step S500, which involves implementing dual-scale hierarchical storage and constructing a hybrid index for the preprocessed multi-source geoscience data based on a three-dimensional hybrid coding system, includes the following steps: S501: Construct a dual-scale hierarchical storage structure; In this step, the dual-scale hierarchical storage structure includes a coarse-scale layer and a fine-scale layer. The coarse-scale layer uses geological time and lithological class as the main organizational units, physically aggregating data from the same geological time (e.g., belonging to the Cenozoic) and the same lithological class (e.g., all igneous rocks). It employs a columnar storage format (e.g., Parquet) to support large-scale cross-regional statistical analysis. The coarse-scale layer has a relatively coarse storage granularity, suitable for regional or global scale data retrieval and aggregation queries. The fine-scale layer uses the adaptive grid cells generated in step S300 as the basic organizational unit. Within each grid cell, it organizes compact key-value storage (e.g., RocksDB or LevelDB) according to data access frequency, placing frequently accessed data on faster storage media and less frequently accessed data on slower media. The fine-scale layer has a finer storage granularity, supporting rapid location and local updates within a single grid cell. A mapping relationship is established between the two scale layers through the spatial grid code portion of the 3D hybrid encoding generated in step S400, allowing queries to quickly locate the target geological unit from the coarse-scale layer and then delve into the fine-scale layer for precise data retrieval.

[0066] S502: Extract data association rules from query logs and record the data patterns and their co-occurrence rates in the joint queries; In this step, this embodiment continuously collects access logs from users and historical query tasks, and records the data items involved in each query and their combination patterns. Based on these query logs, association rule mining algorithms (such as the Apriori algorithm or the FP-Growth algorithm) are used to analyze the co-occurrence relationships between data. Specifically, the set of data items accessed in each query is treated as a transaction, and the frequency of different data item combinations in all transactions is statistically analyzed. The support and confidence of each pair or multiple data item combinations are calculated. Support represents the proportion of the combination appearing in all queries, and confidence represents the conditional probability of querying one data item while simultaneously querying another data item. The system sets an association threshold (e.g., support ≥ 0.05 and confidence ≥ 0.7), retains strong association rules that meet the threshold conditions, and records these rules and their co-occurrence rates in the association rule table. This table is subsequently used to guide the generation of prefetch pointers. For example, if a rule indicates that when querying "granite" data, there is an 85% probability of simultaneously querying "quartz" data, the system marks this rule as a strong association.

[0067] S503: Materialize association rules with a co-occurrence rate exceeding a threshold into prefetch pointers and embed them into a hybrid index; In this step, based on the strong association rules mined in step S502, this embodiment generates corresponding prefetch pointers and embeds them into the hybrid index structure. Specifically, for each strong association rule (such as "when querying data item A, there is a high probability of simultaneously querying data item B"), the system adds a prefetch pointer field to the index entry of data item A. This pointer points to the storage location or cache address of data item B. The prefetch pointer adopts a lightweight reference structure (such as directly storing the key value or offset of the target data) to avoid excessively increasing the index size. When the query hits data item A, the index parsing module automatically reads the prefetch pointer, identifies data items B, C, etc., that are strongly associated with A, and triggers an asynchronous prefetch mechanism: while returning data A, the system asynchronously batch-loads strongly associated data items B and C into the cache (such as Redis or memory cache) in the background without blocking the response of the current query. This prefetch pointer embedding method can be dynamically updated: when the query log accumulates to a certain scale, the system re-mines association rules, eliminates weakened prefetch pointers, and adds newly discovered strong association prefetch pointers, thereby maintaining the adaptability of the index.

[0068] S504: When a query hits a data item, the index automatically resolves the associated prefetch pointer and asynchronously batch loads the strongly associated data into the cache; In this step, this embodiment integrates a prefetch trigger into the query execution engine. When a user or application initiates a data query request, the system first locates the target data item in the hybrid index and returns the main query result. At the same time, the trigger automatically parses the prefetch pointer embedded in the index entry of the data item, extracts all strongly related data item identifiers, and then starts an asynchronous coroutine or background thread to read these strongly related data items in batches and load them into the cache (such as a distributed cache or local memory cache) without affecting the main query response time. After loading, the system updates the cache metadata table to record the cache location, validity period, and access count of the prefetched data. When subsequent query requests access this prefetched data, the results can be returned directly from the cache, thereby avoiding repeated disk I / O or network overhead. In addition, the system also implements a cache eviction policy (such as LRU or LFU). When the cache space is insufficient, cache items with low access frequency or low prefetch hit rate are evicted first to ensure the effective utilization of cache resources.

[0069] S505: Periodically update association rules and dynamically adjust prefetch pointers to maintain index adaptability.

[0070] In this step, this embodiment sets an update cycle (such as weekly or monthly). At the end of each cycle, the recent query logs are rescanned, and the association rule mining process in step S502 is executed to generate an updated set of strong association rules. The new and old rule sets are compared: for newly emerging association rules that meet the threshold conditions, new prefetch pointers are generated and inserted into the corresponding index entries; for weak association rules in the old rules that no longer meet the threshold conditions, the corresponding prefetch pointers are removed from the index entries; for rules that maintain strong associations, their prefetch priority can be appropriately increased or the cache lifetime can be extended.

[0071] The entire update process is performed online without affecting normal query services. After the update is complete, a rule change log is recorded, and optionally, the updated rules are fed back to the query optimizer to guide the generation of the execution plan. Through these steps, this application enables efficient physical storage and logical access to multi-source geoscientific data, providing a storage and indexing foundation for subsequent full lifecycle management.

[0072] In another exemplary embodiment, in step S600, based on hierarchical storage and hybrid indexing, a multidimensional value assessment is performed on each piece of multi-source geoscience data in the standard multi-source geoscience dataset, and hierarchical storage and archiving migration are implemented according to the value score; simultaneously, adjacent grid consistency verification is performed, inconsistency issues are repaired in a hierarchical manner, and the repair experience is fed back to the acquisition stage to form a closed-loop adaptive management from acquisition to archiving, including the following steps: S601: Regularly evaluate multi-dimensional indicators and calculate the comprehensive value score of grid cells; In this step, this embodiment first sets an evaluation period (such as monthly or quarterly), and at the end of each period, it traverses each grid cell in the adaptive grid system to collect multi-dimensional indicators such as data integrity, timeliness, access frequency, and consistency. Among them, the integrity indicator counts the ratio of the actual number of data points in the grid to the theoretically expected number of points and the field missing rate; the timeliness indicator records the collection time of the latest data and calculates the data's age; the access frequency indicator counts the number of queries, the amount of data accessed, and the most recent access time from the query log; the consistency indicator calculates the degree of difference between data from different sources at the same location and the smoothness of data at the boundary of adjacent grids.

[0073] Furthermore, based on the collected multi-dimensional indicators, this embodiment constructs a nonlinear adaptive value assessment model to calculate the comprehensive value score of each grid cell. This model is expressed as follows:

[0074] in, Represents grid cells The overall value score; This represents the integrity score, with a value range of [0,1]. This represents the basic score for timeliness, with a value range of [0,1]. This indicates the time span from the acquisition of multi-source geoscientific data to the current assessment time; This indicates the frequency of data access within the grid, with a value range of [0,1], where 1 represents the highest popularity. The consistency score reflects the level of data consistency within a grid and between the grid and its adjacent grids. The value range is [0,1], where 1 indicates complete consistency. This represents the weight of the integrity dimension, with a value range of [0,1]. This represents the weight of the timeliness dimension, with a value range of [0,1]. This represents the weight of the access popularity dimension, with a value range of [0,1]. This represents the consistency dimension weight, with a value range of [0,1]. The value index represents the nonlinearity of integrity. When it is less than 1, it means that the value enhancement of low-scoring areas is greater; when it is greater than 1, it means that the marginal contribution of high-scoring areas is greater; when it is equal to 1, it is linear. This represents a non-linear index of heat intensity; This represents the aging decay coefficient, with a value range of [0.05, 0.5]. Represents the exponential decay factor, when When it equals 0, it is 1; when... When it approaches 0; This represents the spatial heterogeneity correction factor, with a value range of [0.5, 1.5].

[0075] An exponential decay time-related function is introduced to reflect the "rapid at first, then slow" decay of data value with age. A power function is used to handle the marginal effects of completeness and popularity, and a spatial heterogeneity correction factor is applied. The value score is adjusted upwards or downwards based on the geological structural complexity (LSE value).

[0076] Furthermore, based on the calculated comprehensive value score... In this embodiment, the grid is divided into high-value zones ( ≥0.7), medium value zone (0.3≤ <0.7) and low-value areas ( <0.3), and implement tiered storage migration for different value zones. For example, for grid cells in high-value zones, their data is automatically migrated or stored in a high-speed cache (such as SSD or memory) to ensure fast access; for grid cells in medium-value zones, their data is stored in a standard storage zone (such as HDD) to balance performance and cost; for grid cells in low-value zones and with low access frequency, feature extraction and archiving processes are triggered, and their data is compressed and migrated to a low-cost storage zone (such as tape library or cold storage cloud).

[0077] S602: Extract features from low-value, low-frequency grids and archive them to a low-cost storage area; In this step, for grid cells identified as low-value and low-frequency, the original data is not directly deleted. Instead, feature extraction and summary compression are performed before archiving to a low-cost storage area. Feature extraction includes three aspects: statistical features, calculating statistics such as the mean, variance, and variogram values ​​of each observed variable within the grid cell; spatial features, extracting the spatial location, boundary coordinates, and relationships between adjacent grid cells; and a resolution-downgraded version, generating a low-resolution data version using wavelet transform or sampling compression. All extracted features are combined into a feature summary, which has a storage space much smaller than the original data, enabling rapid response to statistical-level query needs. The original data is then migrated from the high-speed storage area to a low-cost storage area (such as a tape library, cold storage cloud, or hard disk archive), while retaining a pointer to the archive location and the feature summary in the original storage location. When subsequent queries require access to archived data, the feature summary is used to determine if the requirement is met: if only statistical-level information is needed, the summary is returned directly; if the original data is needed, it is read from the archive area. A value recovery mechanism is also set up: if the access popularity of a certain grid suddenly increases after archiving, its value score is recalculated, and when the score rises back to above 0.5, the data is automatically migrated back from the archive area to the standard storage area.

[0078] S603: Periodically perform consistency checks between adjacent grids and handle inconsistency issues in a tiered manner; In this step, within each evaluation cycle, all adjacent grid cell pairs (including horizontal and vertical adjacent pairs) in the adaptive grid system are traversed, and their data consistency near the boundary is compared. For each adjacent pair, the degree of difference (such as root mean square error or relative deviation) of the sampling points on both sides of the boundary is calculated. If the degree of difference exceeds a preset consistency threshold (such as a relative deviation greater than 10%), the adjacent pair is determined to have an inconsistency problem, and the ID, boundary location, and degree of difference of the inconsistent grid are recorded. Subsequently, physical constraint checks are automatically triggered for the inconsistent grids to re-verify whether the data conforms to basic physical laws such as numerical range constraints, logical correlation constraints, and spatial continuity constraints, and a consistency report is generated. Inconsistencies are categorized and handled according to the report results: For issues that can be automatically repaired (such as boundary discontinuities caused by improper interpolation parameters), anisotropic kriging or geologically guided interpolation methods are used to re-interpolate along the geological boundary direction, or the offset data is corrected using a coordinate registration algorithm. After repair, a repair log is recorded (including repair time, repair method, and numerical comparison before and after repair); For issues that cannot be automatically repaired (such as fundamental conflicts between data from different sources), the issue is marked as an alarm and automatically transferred to the work order system, generating a structural work order containing a problem description, grid location, and suggested handling methods, which is then submitted to geological experts or data administrators for manual review and adjudication.

[0079] S604: Successful repair cases are fed back to the data collection strategy optimizer to form a closed-loop adaptive management.

[0080] In this step, the successfully repaired cases from the previous steps (including feature descriptions of inconsistencies, repair methods used, and repair effect evaluations) are stored in a structured manner as experiential knowledge to build a repair case knowledge base. Successful cases from this knowledge base are periodically fed back to the acquisition strategy optimizer in step S100 to achieve closed-loop evolution of the acquisition strategy. The acquisition strategy optimizer automatically adjusts the following acquisition parameters based on the feedback information: Based on the regional characteristics of areas where inconsistencies frequently occur, it optimizes the initial sampling density allocation for that region, increasing the density level corresponding to the LSE threshold; based on the data missing patterns identified during the repair process, it optimizes the optimal sampling point recommendation algorithm, increasing the sampling weight for boundary regions; based on the data quality problem sources recorded in the repair log (such as frequent problems with data from a certain type of sensor or source institution), it adjusts the priority weight of version / source codes, reducing the weight of low-quality source data during subsequent multi-source fusion. Through the above feedback mechanism, a closed-loop adaptive management system is formed, from data acquisition (S100) to preprocessing (S200), grid management (S300), encoding and storage (S400, S500) and then to full life cycle assessment and repair (S600), so that the quality of subsequent batch acquisition and management can be continuously improved with the accumulation of experience.

[0081] In another exemplary embodiment, this application also provides a multi-source geoscience data lifecycle management system, such as... Figure 2As shown, the system includes: a data acquisition module 100, used to adaptively collect multi-source geoscientific data based on the geological structural complexity of the target area, obtaining an original multi-source geoscientific dataset that matches the geological structural complexity of the target area; a preprocessing module 200, used to preprocess the multi-source geoscientific data in the original multi-source geoscientific dataset to generate a standard multi-source geoscientific dataset; a grid construction module 300, used to construct an adaptive grid system for the target area based on the standard multi-source geoscientific dataset, and to achieve spatial adaptive management of the preprocessed multi-source geoscientific data through grid splitting and merging; and a coding system construction module 400, used to construct a coding system based on the adaptive grid system for the target area. A three-dimensional hybrid coding system is established to assign a unique identifier to each piece of multi-source geoscience data in the standard multi-source geoscience dataset; a hybrid index construction module 500 is used to implement dual-scale hierarchical storage and construct a hybrid index for each piece of multi-source geoscience data in the standard multi-source geoscience dataset based on the three-dimensional hybrid coding system; a closed-loop adaptive management module 600 is used to perform multi-dimensional value assessment on each piece of multi-source geoscience data in the standard multi-source geoscience dataset based on hierarchical storage and hybrid index, and implement hierarchical storage and archiving migration based on value scores; at the same time, adjacent grid consistency verification is performed, inconsistency problems are repaired in a hierarchical manner, and the repair experience is fed back to the acquisition stage to form a closed-loop adaptive management from acquisition to archiving.

[0082] Based on the above embodiments, refer to Figure 3 The computer-readable storage medium of exemplary embodiments of this application will be described below. Please refer to [link / reference]. Figure 3 The computer-readable storage medium shown is an optical disc 40, on which a computer program (i.e., a program product) is stored. When the computer program is run by a processor, it implements the steps described in the above-described method implementation, such as: adaptively and differentially collecting multi-source geoscientific data based on the geological structural complexity of the target area to obtain an original multi-source geoscientific dataset that matches the geological structural complexity of the target area; preprocessing the multi-source geoscientific data in the original multi-source geoscientific dataset to generate a standard multi-source geoscientific dataset; and constructing an adaptive grid system for the target area based on the standard multi-source geoscientific dataset, using grid splitting and merging to process the preprocessed multi-source geoscientific data. Spatial adaptive management is implemented; a three-dimensional hybrid coding system is constructed based on the target area adaptive grid system to assign a unique identifier to each multi-source geoscientific data in the standard multi-source geoscientific dataset; based on the three-dimensional hybrid coding system, dual-scale hierarchical storage is implemented for each multi-source geoscientific data in the standard multi-source geoscientific dataset and a hybrid index is constructed; based on the hierarchical storage and hybrid index, multi-dimensional value assessment is performed on each multi-source geoscientific data in the standard multi-source geoscientific dataset, and hierarchical storage and archiving migration are implemented based on the value score; at the same time, consistency verification of adjacent grids is performed, inconsistency problems are repaired in a hierarchical manner, and the repair experience is fed back to the acquisition stage to form a closed-loop adaptive management from acquisition to archiving.

[0083] It should be noted that the computer-readable storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other optical and magnetic storage media, which will not be described in detail here.

[0084] Based on the above embodiments, this application also provides an electronic device, which is described below with reference to... Figure 4 An electronic device for file downloading according to an exemplary embodiment of this application will be described.

[0085] Figure 4 A block diagram is shown of an exemplary electronic device 50 suitable for implementing embodiments of the present application. The electronic device 50 may be a computer system or a cloud server. Figure 4 The electronic device 50 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0086] like Figure 4 As shown, the electronic device 50 includes, but is not limited to: one or more processors or processing units 501, system memory 502, and bus 503 connecting different system components (including system memory 502 and processing unit 501).

[0087] Electronic device 50 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by electronic device 50, including volatile and non-volatile media, removable and non-removable media.

[0088] System memory 502 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 5021 and / or cache memory 5022. Electronic device 50 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, ROM 5023 may be used to read and write non-removable, non-volatile magnetic media (…). Figure 4 (Not shown in the image, usually referred to as "hard drive"). Although not shown in... Figure 4The diagram illustrates that a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk") and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) can be provided. In these cases, each drive can be connected to bus 503 via one or more data media interfaces. System memory 502 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments of this application.

[0089] A program / utility 5025 having a set (at least one) of program modules 5024 may be stored, for example, in system memory 502, and such program modules 5024 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment. Program modules 5024 typically perform the functions and / or methods described in the embodiments of this application.

[0090] Electronic device 50 can also communicate with one or more external devices 504 (such as a keyboard, pointing device, display, etc.). This communication can be performed through input / output (I / O) interface 505. Furthermore, electronic device 50 can also communicate with one or more networks (such as local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via network adapter 506. Figure 4 As shown, network adapter 506 communicates with other modules of electronic device 50 (such as processing unit 501) via bus 503. It should be understood that, although... Figure 4 As not shown, it can be used in conjunction with electronic device 50 with other hardware and / or software modules.

[0091] The processing unit 501 executes various functional applications and data processing by running programs stored in the system memory 502. For example, based on the geological structural complexity of the target area, it adaptively performs differentiated acquisition of multi-source geoscientific data to obtain an original multi-source geoscientific dataset that matches the geological structural complexity of the target area; it preprocesses the multi-source geoscientific data in the original dataset to generate a standard multi-source geoscientific dataset; based on the standard multi-source geoscientific dataset, it constructs an adaptive grid system for the target area, and achieves spatial adaptive management of the preprocessed multi-source geoscientific data through grid splitting and merging; based on the target area... The domain-adaptive grid system constructs a three-dimensional hybrid coding system to assign a unique identifier to each piece of multi-source geoscience data in the standard multi-source geoscience dataset. Based on the three-dimensional hybrid coding system, it implements dual-scale hierarchical storage for each piece of multi-source geoscience data in the standard multi-source geoscience dataset and constructs a hybrid index. On the basis of hierarchical storage and hybrid index, it performs multi-dimensional value assessment on each piece of multi-source geoscience data in the standard multi-source geoscience dataset, and implements hierarchical storage and archiving migration based on the value score. At the same time, it performs consistency verification between adjacent grids, repairs inconsistencies in a hierarchical manner, and feeds back the repair experience to the acquisition stage to form a closed-loop adaptive management from acquisition to archiving.

[0092] The specific implementation methods of each step will not be repeated here. It should be noted that although several units / modules or sub-units / sub-modules of the file concurrent download device have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to the embodiments of this application, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided and embodied by multiple units / modules.

[0093] In the description of this application, it should be noted that the terms "first", "second", and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0094] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0095] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the shown or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0096] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0097] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0098] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium, includes several instructions, and executes the methods described in the various embodiments of this application through a computer device (which may be a personal computer, a cloud server, or a network device, etc.). The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0099] The above embodiments are only for illustrating the technical concept and features of this application, and are intended to enable those skilled in the art to understand the content of this application and implement it accordingly. They should not be construed as limiting the scope of protection of this application. All equivalent changes or modifications made in accordance with the spirit and essence of this application should be included within the scope of protection of this application.

Claims

1. A method for full lifecycle management of multi-source geoscientific data, characterized in that, The method includes: Based on the geological structural complexity of the target area, multi-source geoscientific data are adaptively and differentially collected to obtain an original multi-source geoscientific dataset that matches the geological structural complexity of the target area. Preprocess the multi-source geoscience data in the original multi-source geoscience dataset to generate a standard multi-source geoscience dataset; Based on standard multi-source geoscience datasets, an adaptive grid system for the target region is constructed, and spatial adaptive management of preprocessed multi-source geoscience data is achieved through grid splitting and merging. A three-dimensional hybrid coding system is constructed based on the target region adaptive grid system to assign a unique identifier to each multi-source geoscience data in the standard multi-source geoscience dataset. Based on the three-dimensional hybrid coding system, each multi-source geoscience data in the standard multi-source geoscience dataset is stored in a dual-scale hierarchical manner and a hybrid index is constructed. Based on hierarchical storage and hybrid indexing, a multidimensional value assessment is performed on each piece of multi-source geoscience data in the standard multi-source geoscience dataset. Hierarchical storage and archiving migration are implemented based on the value score. At the same time, consistency verification between adjacent grids is performed, inconsistency issues are repaired in a hierarchical manner, and the repair experience is fed back to the collection stage to form a closed-loop adaptive management from collection to archiving.

2. The method for full lifecycle management of multi-source geoscientific data according to claim 1, characterized in that, The method of adaptively and differentially collecting multi-source geoscientific data based on the geological structural complexity of the target area to obtain an original multi-source geoscientific dataset that matches the geological structural complexity of the target area includes: Load prior geological maps and satellite gravity and magnetic data of the target area, and calculate the local geological structure entropy distribution map; Based on the local geological structure entropy distribution map, the target area is divided into regions of different complexity, and a differentiated initial sampling density is assigned to each region of complexity. Multi-source geoscientific data are collected in real time according to the allocated differentiated initial sampling density, and the local coefficient of variation of each sampling point is calculated. Determine whether the local coefficient of variation exceeds the set threshold. If it does, recommend the next optimal sampling point location and send its coordinates to the sampling stage to perform data supplementation until the local coefficient of variation falls back below the set threshold. All collected data are aggregated, and the raw multi-source geoscience dataset with LSE labels and sampling density levels is output.

3. The method for full lifecycle management of multi-source geoscientific data according to claim 2, characterized in that, The set threshold is an adaptive threshold, and the calculation formula is as follows: in, This indicates that a threshold value has been set. This represents the average local geological tectonic entropy of the target area; , This represents the empirical coefficient.

4. The method for full lifecycle management of multi-source geoscientific data according to claim 1, characterized in that, The preprocessing of multi-source geoscience data in the original multi-source geoscience dataset includes the following steps: Perform format standardization, field mapping, and missing value interpolation on multi-source geoscience data to output a standard multi-source geoscience dataset. Perform physical consistency checks on standard multi-source geoscience datasets; Statistical outlier points are identified by screening multi-source geoscience data that have passed the physical consistency check in the standard multi-source geoscience dataset. Data in the standard multi-source geoscience dataset that fails the physical consistency check are marked as suspicious physical anomalies and subjected to in-depth inspection.

5. The method for full lifecycle management of multi-source geoscientific data according to claim 1, characterized in that, The aforementioned method constructs an adaptive grid system for the target region based on a standard multi-source geoscience dataset. This system achieves spatial adaptive management of the preprocessed multi-source geoscience data through grid splitting and merging, including: Edge detection is performed on the preprocessed multi-source geoscientific data to identify geoscientific boundaries, including lithological abrupt change zones, fault lines, and unconformities. The identified geoscientific boundaries are embedded as hard constraints into the initial mesh generation framework, and the initial constraint mesh is generated. Calculate the multi-source information entropy of each grid cell within the initial constrained grid to assess the degree of joint variation of multiple variables such as lithology, geochemistry, and geophysics. Grid cells with multi-source information entropy exceeding the splitting threshold are split using a quadtree or octree, and grid cells with multi-source information entropy below the merging threshold and consistent with the geological semantics of adjacent grids are merged. Output a non-uniform adaptive grid system and establish a grid metadata table to record the level, boundary coordinates, geological label and entropy value of each grid.

6. The method for full lifecycle management of multi-source geoscientific data according to claim 5, characterized in that, The calculation of the multi-source information entropy of each grid cell within the initial constraint grid includes: Extract multivariate data, including discrete and continuous variables, from all sampling points within a grid cell; Calculating discrete information entropy based on discrete variables, and calculating continuous information entropy based on continuous variables; Multi-source information entropy is obtained based on discrete information entropy and continuous information entropy.

7. The method for full lifecycle management of multi-source geoscientific data according to claim 1, characterized in that, The aforementioned three-dimensional hybrid coding system, constructed based on the target region adaptive grid system, assigns a unique identifier to each piece of multi-source geoscience data in the standard multi-source geoscience dataset, including: Construct a geological time-based coding system and assign a fixed bit-width code to each level; A standard lithology coding table is established based on the geological age hierarchical coding system, and fixed-length codes are assigned to igneous rocks, sedimentary rocks, metamorphic rocks and their subclasses. Based on the target region adaptive grid system, a variable-length prefix code is generated for each grid cell using a space-filling curve, which serves as the spatial grid encoding. Each piece of multi-source geoscience data is assigned a version number and source code to distinguish the data source, collection batch, and processing history; The above four types of codes are combined to generate a three-dimensional hybrid code, which assigns a unique identifier to each data item and stores it in a unified index table.

8. A multi-source geoscientific data lifecycle management system, characterized in that, The system includes: The data acquisition module is used to adaptively collect multi-source geoscientific data based on the geological structural complexity of the target area, so as to obtain an original multi-source geoscientific dataset that matches the geological structural complexity of the target area. The preprocessing module is used to preprocess the multi-source geoscience data in the original multi-source geoscience dataset to generate a standard multi-source geoscience dataset. The grid construction module is used to build an adaptive grid system for the target area based on standard multi-source geoscience datasets. It enables spatial adaptive management of preprocessed multi-source geoscience data through grid splitting and merging. The coding system construction module is used to construct a three-dimensional hybrid coding system based on the target area adaptive grid system, and to assign a unique identifier to each piece of multi-source geoscience data in the standard multi-source geoscience dataset. The hybrid index building module is used to implement dual-scale hierarchical storage for each piece of multi-source geoscience data in the standard multi-source geoscience dataset based on the three-dimensional hybrid coding system and to build a hybrid index. The closed-loop adaptive management module is used to perform multi-dimensional value assessment on each piece of multi-source geoscience data in the standard multi-source geoscience dataset based on hierarchical storage and hybrid indexing, and implement hierarchical storage and archiving migration according to the value score; at the same time, it performs consistency verification between adjacent grids, repairs inconsistencies in a hierarchical manner, and feeds back the repair experience to the collection stage to form a closed-loop adaptive management from collection to archiving.

9. A storage medium, characterized in that, It includes instructions that, when executed on a computer, cause the computer to perform the method as described in any one of claims 1 to 7.

10. An electronic device, characterized in that, The electronic device includes: Memory, processor, and computer programs stored in memory and executable on the processor, wherein, When the processor executes the program, it implements the method as described in any one of claims 1 to 7.

Images