A geological disaster database construction method and device, electronic equipment and storage medium

Abstract

Embodiments of the present disclosure provide a geological disaster database construction method and device, electronic equipment and storage medium, to realize multi-source data fusion, intelligent quality control, dynamic update tracing in database construction. Including: acquiring multi-source geological disaster investigation data, constructing a trinity storage structure of a three-layer double-drive architecture; based on the trinity storage structure, generating a dynamic intelligent code for each geological disaster point to obtain a unified code, based on the unified code, performing spatial proximity matching and attribute similarity comparison through a multi-source data fusion engine to establish a correlation between the multi-source geological disaster investigation data; based on the correlation and the multi-source geological disaster investigation data, performing spatial verification, logical verification and correlation verification to generate a quality detection result; based on the quality detection result, dynamically updating the database and generating a version control record to obtain a traceable database containing base data and incremental data.

Description

Technical Field

[0001] This disclosure relates to the field of geological research technology, and in particular to a method, apparatus, electronic device and storage medium for constructing a geological disaster database. Background Technology

[0002] Geological hazard risk investigation is a core foundational task of the disaster prevention and mitigation system and a crucial prerequisite for identifying potential disaster hazards, assessing risk levels, and formulating prevention and control strategies.

[0003] Since the commencement of the geological hazard risk survey, the scale of data collection has grown exponentially, forming a massive, multi-source data system. However, existing technologies for constructing geological hazard risk assessment databases still face bottlenecks, severely hindering the transformation of survey results into dynamic risk assessment and disaster prevention decision support. For example, the integration of multi-source heterogeneous data is low, lacking a unified correlation mechanism; data quality control relies on manual review, resulting in delayed and inaccurate anomaly identification; and dynamic updating and data traceability mechanisms are lacking, making it difficult to adapt to real-time, full lifecycle management needs.

[0004] The aforementioned technical bottlenecks prevent existing geological hazard risk investigation and assessment databases from achieving deep integration of data and decision-making: on the one hand, fragmented storage of multi-source data fails to provide complete and accurate input parameters for risk assessment models (such as numerical simulation of landslide stability); on the other hand, the real-time data access and historical trend analysis capabilities required for dynamic risk assessment are insufficient, making it difficult for investigation results to support refined and timely disaster prevention decisions. Therefore, there is an urgent need for a database construction solution that can overcome technical bottlenecks such as multi-source data fusion, intelligent quality control, and dynamic update traceability. Summary of the Invention

[0005] In view of this, the present disclosure provides a method, apparatus, electronic device and storage medium for constructing a geological disaster database, so as to realize multi-source data fusion, intelligent quality control and dynamic update traceability in database construction.

[0006] Firstly, a method for constructing a geological disaster database is provided, comprising: acquiring multi-source geological disaster survey data, constructing a three-layer dual-drive database storage structure to obtain a three-in-one storage structure including a relational database, a spatiotemporal database, and a knowledge base; the multi-source geological disaster survey data includes field survey data, remote sensing data, and real-time monitoring data; based on the three-in-one storage structure, generating a dynamic intelligent code for each geological disaster point to obtain a unified code containing a basic code, a feature code, and a time code, wherein the basic code is an administrative division code, the feature code includes a disaster type code and a remote sensing interpretation code, and the time code is a time identifier; based on the unified code, performing spatial proximity matching and attribute similarity comparison through a multi-source data fusion engine to establish an association relationship between the field survey data, the remote sensing data, and the real-time monitoring data; based on the association relationship and the multi-source geological disaster survey data, performing spatial verification, logical verification, and association verification through an intelligent quality control system to generate quality inspection results; based on the quality inspection results, dynamically updating the database using an incremental update mode and generating version control records to obtain a traceable database containing base state data and incremental data.

[0007] In one implementation, a three-tiered, dual-drive database storage structure is constructed to obtain a unified storage structure comprising a relational database, a spatiotemporal database, and a knowledge base. This includes: constructing a relational database based on the field survey data using a hierarchical master-slave table structure for standardized storage; wherein the master table of the relational database stores the unified disaster point number, spatial coordinates, and disaster type, and the slave tables of the relational database store differentiated attribute parameters according to disaster type; constructing a spatiotemporal database based on the remote sensing data and the real-time monitoring data using a triplet storage model containing spatial coordinates, timestamps, and monitoring values; obtaining disaster type criteria and evaluation model parameters, constructing a knowledge graph using graph database storage technology, and establishing a knowledge base; wherein the knowledge base contains association nodes between disaster types, criteria, and evaluation models, thus obtaining the unified storage structure.

[0008] In one implementation, based on the remote sensing data and the real-time monitoring data, a spatiotemporal database is constructed using a triplet storage model that includes spatial coordinates, timestamps, and monitoring values. This includes: acquiring disaster point deformation monitoring data from the remote sensing data and the real-time monitoring data, storing it in a triplet format that includes spatial coordinates, timestamps, and monitoring values, and generating a spatiotemporal data model; and based on the spatiotemporal data model, establishing a spatiotemporal index and integrating the functions of a temporal geographic information system (GIS) to obtain a spatiotemporal database that supports spatiotemporal range queries and historical status backtracking.

[0009] In one implementation, disaster type criteria and evaluation model parameters are obtained, and a knowledge graph is constructed using graph database storage technology to establish a knowledge base. This includes: obtaining disaster type criteria and evaluation model parameters, constructing a knowledge graph between disaster types, criteria, and evaluation models, and establishing knowledge base nodes; and based on the knowledge base nodes, associating numerical simulation parameters to obtain a knowledge base that supports intelligent reasoning and automatic matching of model parameters.

[0010] In one implementation, a dynamic intelligent code is generated for each geological disaster site to obtain a unified code containing a base code, a feature code, and a time code. This includes: acquiring the administrative division information of the geological disaster site, obtaining a 6-digit county-level code from the National Bureau of Statistics interface, and generating the base code; based on the type attribute of the disaster site, matching the disaster type code table from a knowledge base to generate a 7-10 digit disaster type code; acquiring the interferometric synthetic aperture radar (InSAR) deformation level data of the disaster site, performing level encoding with a remote sensing interpretation system to generate an 11-14 digit remote sensing interpretation code; acquiring the current system date, extracting time information accurate to the day, and generating a 15-18 digit time code; and based on the base code, the disaster type code, the remote sensing interpretation code, and the time code, performing a uniqueness check and establishing a historical code mapping relationship to obtain an 18-digit unified code.

[0011] In one implementation, a multi-source data fusion engine is used to perform spatial proximity matching and attribute similarity comparison to establish the correlation between the field survey data, the remote sensing data, and the real-time monitoring data. This includes: acquiring radar point cloud data and high-resolution remote sensing images, performing radiometric calibration, atmospheric correction, and point cloud denoising to obtain preprocessed remote sensing data; based on the preprocessed remote sensing data, using the Scale Invariant Feature Transform (SIFT) algorithm to extract corresponding feature points and performing registration using the Random Sample Consensus (RANSAC) algorithm to obtain registered point cloud data; based on the registered point cloud data, using a region growing algorithm to extract the three-dimensional contour of the disaster body and generate a vector surface file; and based on the spatial contour of the vector surface file... The system uses attributes to establish a correlation between remote sensing data and field survey data by spatial overlay analysis and matching with the unified coding. It receives data from BeiDou monitoring terminals, which includes terminal ID, timestamp, spatial coordinates, displacement, and tilt angle data, and caches this data using a message queue to obtain a cached monitoring data stream. Outliers are removed from the cached monitoring data stream, and linear interpolation is used to complete missing data, resulting in cleaned monitoring data. Based on the timestamp and BeiDou monitoring terminal ID of the cleaned monitoring data, it is matched with the unified disaster point number in the database, and a continuous surface data is generated using the Kriging interpolation algorithm, establishing a correlation between the real-time monitoring data and the field survey data.

[0012] In one implementation, a quality inspection result is generated by performing spatial verification, logical verification, and correlation verification through an intelligent quality control system. This includes: acquiring disaster point location data collected in the field; comparing the spatial positional relationship between the acquisition point and the vector boundary of the survey area in real time using a BeiDou positioning module to generate a spatial verification result; wherein, when the distance between the acquisition point and the vector boundary of the survey area exceeds a first distance, it is marked as a spatial anomaly; acquiring disaster point attribute data collected in the field; performing a legality check based on a preset value range and correlation logic rules to generate a logical verification result; acquiring EXIF ​​information of exchangeable image files (EXIF format) of photos collected in the field; extracting the GPS coordinates and shooting time of the photos and comparing them with the disaster point coordinates and acquisition time to generate a correlation verification result; wherein, when the deviation between the GPS coordinates of the photos and the disaster point coordinates exceeds a second distance, the upload is rejected.

[0013] In one implementation, the spatial verification, logical verification, and correlation verification performed through the intelligent quality control system further include: acquiring remote sensing image interpretation results, performing batch detection using a trained lightweight convolutional neural network (CNN) model, identifying boundary delineation errors and land cover classification deviations, and generating interpretation anomaly reports; acquiring historical time series of monitoring data, calculating the mean and standard deviation of the monitoring data using a sliding window and performing trend analysis, identifying data mutation anomalies, and generating anomaly data notifications.

[0014] In one implementation, an incremental update mode is used to dynamically update the database and generate version control records to obtain a traceable database containing base state data and incremental data. This includes: generating a base state snapshot based on initial survey results, containing complete backups of attribute databases, spatial databases, and data databases, and establishing a base state database; acquiring data change information, including new disaster points, parameter updates, and data deletions, recording the operation type, updated fields, old values, new values, updater, and timestamp of the changed data, and generating incremental records; and setting a version number for the database based on the base state database and the incremental records, supporting version backtracking by version number, update time, or operator, to obtain the traceable database.

[0015] In one embodiment, the method further includes: based on the hazard body parameters in the traceable database, the hazard body parameters including landslide thickness and soil mechanical properties, calling a three-dimensional continuous medium fast Lagrangian analysis program through a dynamic evaluation engine. The numerical simulation interface generates stability evaluation results. Based on the stability evaluation results and preset evaluation factors, including the number of people threatened, economic losses, and terrain slope, the weights are calculated using the analytic hierarchy process (AHP) and a risk level evaluation is performed to obtain the risk level results. Based on the risk level results, when the risk level escalates from medium risk to high risk, an early warning threshold mechanism is automatically triggered to generate early warning information and push it to the management platform.

[0016] Secondly, embodiments of this disclosure provide a geological disaster database construction apparatus, comprising:

[0017] The database storage structure construction module is used to acquire multi-source geological disaster survey data, construct a three-layer dual-drive database storage structure, and obtain a three-in-one storage structure including a relational database, a spatiotemporal database, and a knowledge base; the multi-source geological disaster survey data includes field survey data, remote sensing data, and real-time monitoring data;

[0018] A unified coding generation module is used to generate a dynamic intelligent code for each geological disaster point based on the three-in-one storage structure, resulting in a unified code containing a basic code, a feature code, and a time code. The basic code is an administrative division code, the feature code includes a disaster type code and a remote sensing interpretation code, and the time code is a time identifier.

[0019] The association establishment module is used to establish the association between the field survey data, the remote sensing data and the real-time monitoring data based on the unified coding and through the multi-source data fusion engine to perform spatial proximity matching and attribute similarity comparison;

[0020] The quality inspection module is used to generate quality inspection results by performing spatial verification, logical verification, and correlation verification through an intelligent quality control system based on the correlation relationship and the multi-source geological disaster survey data.

[0021] The database generation module is used to dynamically update the database and generate version control records based on the quality inspection results using an incremental update mode, thereby obtaining a traceable database containing base state data and incremental data.

[0022] Thirdly, embodiments of this disclosure also provide an electronic device, including a processor and a memory, wherein the memory is used to store computer instructions, and the processor is used to execute the computer instructions to implement the geological disaster database construction method described in any of the above embodiments.

[0023] Fourthly, a computer-readable storage medium is provided, wherein computer instructions are stored therein, and when executed, the computer instructions implement the geological disaster database construction method described in any of the above embodiments.

[0024] The geological disaster database construction scheme described in this embodiment utilizes a three-in-one storage structure consisting of a relational database, a spatiotemporal database, and a knowledge base. This allows for differentiated storage based on the characteristics of different data types. For example, the relational database can standardize the management of disaster point attribute data (such as unified numbering and landslide thickness), while the spatiotemporal database can store extended timestamped spatial data (such as disaster point deformation trajectories). The knowledge base typically stores evaluation model parameters, eliminating data storage barriers at the underlying architecture level. Furthermore, unified encoding using a base code, feature code, and time code achieves semantic consistency in the data. Spatial proximity matching and attribute similarity comparison further enable automatic association between field survey data, remote sensing data, and real-time monitoring data.

[0025] In addition, during the field data collection phase, spatial verification, logical verification, and correlation verification are used to prevent erroneous data from entering the database at the source.

[0026] Furthermore, this embodiment employs a dynamic update mode based on a ground-state + incremental approach. The ground-state database stores initial survey results, while the incremental database records only changed data, avoiding the efficiency loss caused by full coverage and significantly improving data update efficiency. Moreover, version control can be combined to support arbitrary version backtracking, ensuring data traceability and immutability throughout its entire lifecycle, resolving the problem of poor historical data traceability, and guaranteeing data authenticity and credibility.

[0027] In summary, the embodiments disclosed herein lay the foundation for risk assessment model integration through the synergy of structured storage, precise correlation, and high-quality data. A unified encoding and fusion engine ensures that the model can quickly retrieve complete data including disaster point attributes, remote sensing deformation, and real-time monitoring. Intelligent quality control guarantees the accuracy of input parameters, and dynamic updates ensure that model input is synchronized with the actual state of the disaster body. Ultimately, this achieves automation and precision in stability assessment, risk level classification, and early warning range generation, providing scientific and accurate data support for disaster prevention decisions such as emergency evacuation and disaster prevention engineering layout. Attached Figure Description

[0028] Figure 1 A flowchart illustrating a method for constructing a geological disaster database, as provided in this embodiment of the disclosure;

[0029] Figure 2 A schematic diagram of a geological disaster database construction device is also provided as an embodiment of this disclosure;

[0030] Figure 3 This is a schematic diagram of an electronic device provided as an embodiment of the present disclosure. Detailed Implementation

[0031] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0032] The terminology used in this disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The singular forms “a,” “the,” and “the” as used in this disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0033] It should be understood that although the terms first, second, third, etc., may be used in this disclosure to describe various information, such information should not be limited to these terms. These terms are used only to distinguish information of the same type from one another. For example, without departing from the scope of this disclosure, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0034] Furthermore, the symbol “ / ” in this disclosure indicates that there is an “or” relationship between the related objects before and after the symbol, or that the related objects have an exemplary relationship that can coexist.

[0035] Research has revealed that the existing geological disaster risk investigation and assessment database faces technical bottlenecks, including difficulties in integrating multi-source heterogeneous data, low efficiency of manual quality control, lack of dynamic updating and traceability mechanisms, and difficulty in supporting the in-depth application of risk assessment models. Systematic innovation is urgently needed in core aspects such as database architecture design, data association methods, quality control systems, and update and traceability mechanisms to overcome the current technical difficulties.

[0036] To address this, this disclosure proposes a method for constructing a geological disaster risk investigation and assessment database. It employs a three-layer, dual-drive architecture integrating a relational database, a spatiotemporal database, and a knowledge base to generate dynamic intelligent coding containing basic codes, feature codes, and time codes. A multi-source data fusion engine enables precise correlation between field survey data, remote sensing data, and real-time monitoring data. An intelligent quality control system ensures data quality, and an incremental update mode and version control records are used to construct a traceable database. Ultimately, this method achieves efficient integration, precise quality control, dynamic updates, and in-depth application of geological disaster risk investigation data, providing strong support for dynamic risk assessment and disaster prevention and mitigation decision-making.

[0037] The solutions of the present disclosure will be further described in detail below through specific embodiments.

[0038] like Figure 1 As shown, a method for constructing a geological disaster database is provided in an embodiment of this disclosure, including:

[0039] S101: Acquire multi-source geological disaster survey data, construct a three-layer dual-drive database storage structure, and obtain a three-in-one storage structure including a relational database, a spatiotemporal database, and a knowledge base; the multi-source geological disaster survey data includes field survey data, remote sensing data, and real-time monitoring data.

[0040] Here, field survey data consists of core information collected on-site at disaster sites, such as the site's unique identification number, coordinates, type (landslide / collapse / debris flow), landslide thickness, main sliding direction, and number of people at risk. This data forms the foundational attribute data for the database. Remote sensing data, acquired through technologies such as satellites, LiDAR, and InSAR, includes high-resolution optical imagery, 3D point cloud data, and surface deformation data, providing information on the spatial morphology and dynamic changes of the disaster body. Real-time monitoring data, collected in real-time via terminals, includes data such as the displacement, tilt angle, and displacement rate of disaster sites, reflecting the current activity status of the disaster body.

[0041] Based on the different characteristics (attributes, spatiality, rules) of the aforementioned multi-source data, a differentiated storage structure is designed to classify and store data in corresponding databases, achieving dedicated database storage and facilitating association. Specifically, the relational database adopts a master-detail table hierarchical structure, storing standardized attribute data (such as disaster point numbers, coordinates, and threatened populations from field surveys). The master table stores core common attributes, while the detail tables store differentiated attributes according to disaster type (e.g., landslide detail tables store landslide thickness, debris flow detail tables store source supply type). Through unified numbering, multi-table joint queries are supported. The spatiotemporal database is built based on PostGIS extensions, storing time-stamped spatial data (such as deformation trajectories of disaster points from remote sensing and displacement time-series data from real-time monitoring), supporting spatiotemporal indexing and historical status backtracking (e.g., querying the range of landslide displacement exceeding 5mm during a certain period). The knowledge base uses a graph database to store rule-based data (such as disaster type criteria, risk assessment model parameters, etc.). (Numerical simulation default values), constructing a disaster type-criteria-evaluation model knowledge graph to support subsequent intelligent reasoning and model parameter matching. The core storage layer in this three-layer dual-drive architecture (core storage layer + intelligent engine layer + interface adaptation layer) is a three-in-one structure composed of these three types of libraries. It is a storage scheme customized according to the characteristics of data types, avoiding inefficient queries and difficulties in association caused by mixed storage of multi-source data, and laying the underlying architectural foundation for subsequent data fusion and intelligent applications.

[0042] In one implementation, a three-tiered, dual-drive database storage structure is constructed to obtain a unified storage structure comprising a relational database, a spatiotemporal database, and a knowledge base. This includes: First, based on the field survey data, a standardized storage structure using a master-slave table hierarchical structure is employed to construct a relational database. The master table of the relational database stores the unified disaster point number, spatial coordinates, and disaster type, while the slave tables store differentiated attribute parameters according to disaster type. Second, based on the remote sensing data and the real-time monitoring data, a spatiotemporal database is constructed using a triplet storage model containing spatial coordinates, timestamps, and monitoring values. This spatiotemporal database is indexed using PostGIS extensions. Third, disaster type criteria and evaluation model parameters are obtained, and a knowledge graph is constructed using graph database storage technology to establish a knowledge base. This knowledge base contains the association nodes between disaster types, criteria, and evaluation models, thus obtaining the unified storage structure.

[0043] Here, by customizing three types of dedicated databases based on the differences in the types and uses of multi-source data and forming a collaborative storage system, the problems of difficult data association and inefficient querying under the traditional mixed storage model are solved. Specifically, for the characteristics of field survey data (such as disaster point number, coordinates, landslide thickness, etc.) with many attributes and the need for classification and association, a hierarchical structure of master table + slave table is adopted to build the relational database. The master table uniformly stores the key information common to all disaster points (unified number, spatial coordinates, disaster type) as the basis for association (such as all landslide and collapse points need to record coordinates). The slave table stores differentiated attributes, specifically split by disaster type (such as the landslide slave table recording landslide thickness and main sliding direction, and the debris flow slave table recording material supply type), avoiding redundancy of fields in a single table. The master table and slave table are linked by a unified number (master table primary key), supporting multi-table joint queries (such as querying "coordinates + landslide thickness of landslide point in County A"), realizing the structured and standardized storage of attribute data.

[0044] To address the characteristics of remote sensing data (such as InSAR deformation trajectories) and real-time monitoring data (such as BeiDou displacement rates) which contain spatial location, time dimension, and monitoring values, a spatiotemporal database is constructed using a triplet model of spatial coordinates, timestamps, and monitoring values. For example, when storing data for a landslide point, it is necessary to simultaneously record the longitude AA° (spatial), 2024-05-20 14:30 (timestamp), and displacement 2.3mm (monitoring value). This design can directly support spatiotemporal range queries (such as querying landslide points with displacements exceeding 5mm from January to March 2024) and historical status retrospection (such as comparing deformation trends in different months), adapting to dynamic monitoring needs.

[0045] For rule-based data required for disaster assessment (such as landslide stability criteria, Simulation parameters are used to construct a knowledge graph-style knowledge base using a graph database. The stored content includes three types of related nodes: disaster type (e.g., landslide), criteria (e.g., landslide thickness > 10m), and evaluation model (e.g., stability evaluation model). Reasoning logic is formed through node association (e.g., landslide thickness > 10m + water content > 25% → calling model parameters to determine poor stability). Subsequently, evaluation model parameters can be automatically matched to support intelligent risk assessment.

[0046] In summary, this implementation method is based on a data characteristic-adaptive storage scheme. It stores attribute data from field surveys, spatiotemporal data from remote sensing / measurement, and rule data for evaluation in relational databases, spatiotemporal databases, and knowledge bases, respectively. This ensures efficient storage of each type of data and lays the foundation for subsequent data fusion and intelligent applications through unified numbering, spatiotemporal indexing, and node association, ultimately forming a three-in-one core storage system.

[0047] In one implementation, based on the remote sensing data and the real-time monitoring data, a spatiotemporal database is constructed using a triplet storage model that includes spatial coordinates, timestamps, and monitoring values. This can include: acquiring disaster point deformation monitoring data from the remote sensing data and the real-time monitoring data, storing it in the form of triplets containing spatial coordinates, timestamps, and monitoring values, and generating a spatiotemporal data model; based on the spatiotemporal data model, establishing a spatiotemporal index and integrating the functions of a temporal geographic information system (GIS) to obtain a spatiotemporal database that supports spatiotemporal range queries and historical status backtracking.

[0048] In this implementation, a triplet storage model is used to achieve structured data storage. For the core information of disaster point deformation monitoring in remote sensing data (such as InSAR deformation data) and real-time monitoring data (such as BeiDou displacement data), three key dimensions—space, time, and numerical value—are extracted. This data is stored as a triplet of spatial coordinates (such as latitude and longitude) + timestamp (such as 2024-05-20 14:30:00) + monitored value (such as displacement 2.3mm, tilt angle 0.02°), generating a standardized spatiotemporal data model. In this way, the originally fragmented deformation monitoring data (such as InSAR XML files and BeiDou CSV data) can be transformed into storage units with unified dimensions and clear structure, ensuring that each piece of monitoring data clearly corresponds to "where (space), when (time), and what state (monitored value)," laying the foundation for subsequent association and querying.

[0049] Based on the triplet data model, the practicality of the spatiotemporal database is enhanced through two key designs: establishing a spatiotemporal index and integrating temporal GIS functionality. Firstly, the index, built based on the spatial coordinates and timestamps of the data (such as ST_TemporalIndex mentioned in the documentation), allows for rapid location of data within a specific spatiotemporal range. For example, it can efficiently query "monitoring data of landslide points with displacement exceeding 5mm within County A from January to March 2024," avoiding the inefficiency of traditional full-volume scanning. Secondly, the integration of temporal GIS functionality supports recording the historical changes in the spatial morphology of disaster points (such as the annual positional shift of landslide trailing edges) and allows for retrospective analysis of spatial data status at any given time point (such as viewing the deformation range of a landslide in 2023), meeting the needs of dynamically analyzing disaster development trends. In summary, this implementation method addresses the problem of fragmented spatiotemporal information in remote sensing / monitoring data through a triplet model. Furthermore, by leveraging spatiotemporal indexing and temporal GIS functions, the spatiotemporal database not only stores data but also enables efficient data retrieval (spatiotemporal range query) and historical data retrospective (status retrospective). Ultimately, it provides precise spatiotemporal dimension data support for subsequent multi-source data association and dynamic risk assessment.

[0050] In one implementation, acquiring disaster type criteria and evaluation model parameters, constructing a knowledge graph using graph database storage technology, and establishing a knowledge base may include: acquiring disaster type criteria and evaluation model parameters, constructing a knowledge graph between disaster types, criteria, and evaluation models using Neo4j graph database technology, and establishing knowledge base nodes; based on the knowledge base nodes, associating them with the Fast Lagrangian Analysis of Continua in 3 Dimensions (FLA) program. Numerical simulation parameters are used to obtain a knowledge base that supports intelligent reasoning and automatic matching of model parameters.

[0051] This implementation transforms the rule data required for disaster assessment into a structured knowledge graph and associates it with numerical simulation parameters, upgrading the knowledge base from static storage to dynamic support for intelligent applications. First, a knowledge graph is constructed using graph database technology, targeting disaster type criteria (e.g., landslide determination requires a slip surface + displacement deformation) and evaluation model parameters (e.g., the Fs threshold for landslide stability evaluation (Fs < 1.0 indicates instability)). Specifically, three types of core knowledge base nodes are defined: corresponding to disaster types (e.g., landslides, collapses), criteria (e.g., slip surface existence, displacement rate > 10 mm / month), and evaluation models (e.g., stability evaluation models, risk level classification models). Then, relationships (graph edges) are established between nodes: for example, landslide type → associated with "slip surface existence" and "displacement deformation" criteria → then associated with the landslide stability evaluation model, forming a clear logical chain of type-criteria-model, transforming scattered rules into structured knowledge and avoiding the difficulty of rule lookup in traditional storage. Second, based on the basic knowledge graph (type-criteria-model), specific numerical simulation parameters (e.g., ...) are further associated. The simulation requires default values ​​for the internal friction angle and cohesion. For example, in the landslide stability evaluation model node, when the associated landslide thickness is >10m, The simulation uses parameters such as a cohesion of 18 kPa and an internal friction angle of 28°. This implementation method enables the knowledge base to intelligently reason and automatically match parameters. For example, when the system needs to evaluate the stability of a landslide, it can automatically retrieve the corresponding parameters based on the landslide type, the stability evaluation model, and the landslide thickness of 12m. Simulated parameters, eliminating the need for manual database lookups and input, support the automation of subsequent risk assessments. In short, this implementation method uses a knowledge graph to chain together rules such as criteria and models, and then associates them with numerical simulation parameters. This allows the knowledge base to not only store rules but also proactively provide reasoning logic and parameter support for intelligent assessment, addressing the pain point of traditional databases where rules and parameters are scattered and unable to support automated applications. It provides crucial knowledge support for the subsequent dynamic risk assessment engine.

[0052] S102: Based on the aforementioned three-in-one storage structure, a dynamic intelligent code is generated for each geological disaster point, resulting in a unified code containing a basic code, a feature code, and a time code. The basic code is an administrative division code, the feature code includes a disaster type code and a remote sensing interpretation code, and the time code is a time identifier.

[0053] Here, key information is extracted from a three-in-one storage structure to generate dynamic intelligent codes containing spatial, feature, and temporal dimensions, addressing the issues of confusing identifiers and inefficient matching when associating multi-source data. Specifically, a unique and unified code is generated for each disaster point. This code condenses core information from relational databases (such as administrative divisions and disaster types), spatiotemporal databases (such as time information), and knowledge bases (such as remote sensing interpretation results) into a single code string. This allows data from different sources (field surveys, remote sensing, and monitoring) to be quickly associated through the code, avoiding the problems of multiple identifiers and difficulty in matching data for the same disaster point. The encoding consists of three parts: a basic code, a feature code, and a time code. Each part corresponds to a specific data dimension, ensuring information integrity and ease of interpretation. The basic code (administrative division code) is obtained from the spatial coordinates of disaster points in the relational database. For example, a 6-digit standard administrative division code clarifies the spatial affiliation of disaster points, facilitating data filtering and management by region. The feature code (disaster type code + remote sensing interpretation code) integrates information from the relational database and the knowledge base. The disaster type code (e.g., 01 represents landslide, 02 represents collapse) is extracted from the disaster type field in the relational database, while the remote sensing interpretation code (e.g., "0002" represents InSAR level 2 deformation) is extracted from the remote sensing interpretation results in the knowledge base, reflecting the core attributes and remote sensing characteristics of the disaster point. The time code (time identifier) ​​is extracted from the monitoring / investigation timestamps in the spatiotemporal database. For example, an 8-digit year-month-day format (e.g., "20240520") records the first investigation or latest update time of the disaster point, reflecting the dynamic nature of the encoding and facilitating the tracking of data timeliness. The generated unified code can be used throughout the entire data lifecycle; for example, by encoding "530621-01-0002-20240520", all data of "XX County landslide point, InSAR secondary deformation, updated on May 20, 2024" can be quickly linked (the thickness of the landslide body in the field survey, the deformation trajectory of the remote sensing, and the displacement data monitored in real time). At the same time, the time code can be adjusted synchronously with the data update, realizing dynamic code update and real-time data linkage, providing an efficient identification foundation for subsequent intelligent quality control and risk assessment.

[0054] In one implementation, a dynamic intelligent code is generated for each geological disaster site to obtain a unified code containing a base code, a feature code, and a time code. This includes: acquiring the administrative division information of the geological disaster site, obtaining a 6-digit county-level code from the National Bureau of Statistics interface, and generating the base code; based on the type attribute of the disaster site, matching the disaster type code table from a knowledge base to generate a 7-10 digit disaster type code; acquiring the interferometric synthetic aperture radar (InSAR) deformation level data of the disaster site, performing level encoding with a remote sensing interpretation system to generate an 11-14 digit remote sensing interpretation code; acquiring the current system date, extracting time information accurate to the day, and generating a 15-18 digit time code; and based on the base code, the disaster type code, the remote sensing interpretation code, and the time code, performing a uniqueness check and establishing a historical code mapping relationship to obtain an 18-digit unified code.

[0055] Here, the codes are first generated by field. The basic code (1-6 digits) is a 6-digit county-level code (e.g., "530621" for County A) obtained from the administrative division information of the disaster site via the National Bureau of Statistics interface, used to clarify the spatial affiliation of the disaster site. The disaster type code (7-10 digits) is a 4-digit code (e.g., "0001" for landslide) generated by matching the disaster site type attribute (e.g., landslide) from the knowledge base code table, used to identify the core disaster type. The remote sensing interpretation code (11-14 digits) is a 4-digit code (e.g., "0002" for level 2 deformation) obtained from the InSAR deformation level data of the disaster site and converted by the remote sensing interpretation system, used to associate remote sensing feature information. The time code (15-18 digits) is a 4-digit code generated by extracting the current date (accurate to the day) from the system (e.g., "0520" for May 20, 2024), used to reflect the dynamic timeliness of the code. After combining the four fields, a uniqueness check is first performed (to avoid duplication with existing codes), and then a historical code mapping relationship is established (such as retaining the association between the old code and the new code after the code is updated). Finally, an 18-bit unified code (such as "530621000100020520") is generated, which not only ensures that the code of each disaster point is unique, but also allows for the tracing of historical code changes, supporting the full life cycle management of data.

[0056] S103: Based on the unified coding, spatial proximity matching and attribute similarity comparison are performed through a multi-source data fusion engine to establish the correlation between the field survey data, the remote sensing data and the real-time monitoring data.

[0057] Here, using unified coding as the core link, the multi-source data fusion engine employs dual comparison (spatial + attribute) to accurately link previously scattered field survey, remote sensing, and real-time monitoring data to the same geological hazard point. Unified coding integrates key information such as the administrative division (spatial), type / remote sensing characteristics (attribute), and time of the hazard point, becoming a common identifier for all source data. The fusion engine first uses the coding as a basis to filter out multi-source data for the same hazard point (such as field survey records, remote sensing deformation maps, and real-time displacement monitoring data for a landslide point). Dual comparison ensures the accuracy of data association. Specifically, spatial proximity matching further compares the spatial coordinates of the selected multi-source data (such as the hazard point coordinates in field survey records, the monitoring area coordinates in remote sensing data, and the installation coordinates of real-time monitoring equipment). By determining whether the coordinates are within a proximity range (such as within an error threshold), data with spatial mismatch (such as data mistakenly associated with adjacent hazard points) is excluded. Attribute similarity comparison simultaneously compares the core attributes of the data (such as whether the disaster type in the field survey is consistent with the type interpreted by remote sensing, and whether the deformation trend in real-time monitoring matches the trend in remote sensing analysis). By judging attribute similarity, it ensures that the associated data are logically consistent in content (such as avoiding incorrect association between landslide data and collapse data). In this way, through coding anchoring and dual comparison, the correspondence between field survey, remote sensing, and real-time monitoring data of the same disaster point is finally clarified (such as "2024 field survey landslide thickness data of a certain landslide point → concurrent remote sensing InSAR deformation data → real-time displacement monitoring data of the same time period"), laying the foundation for subsequent data integration and analysis (such as comprehensive assessment of disaster risk).

[0058] In one implementation, the correlation between the field survey data, the remote sensing data, and the real-time monitoring data can be established according to the following process:

[0059] Establishing the correlation between remote sensing data and field survey data: Radar point cloud data and high-resolution remote sensing images are acquired, and radiometric calibration, atmospheric correction, and point cloud denoising are performed to obtain preprocessed remote sensing data. Based on the preprocessed remote sensing data, the Scale Invariant Feature Transform (SIFT) algorithm is used to extract corresponding feature points, which are then registered using the Random Sample Consensus (RANSAC) algorithm to obtain registered point cloud data. Based on the registered point cloud data, a region growing algorithm is used to extract the three-dimensional contour of the disaster body, generating a vector surface file. Based on the spatial attributes of the vector surface file, spatial overlay analysis is used to perform correlation matching with the unified coding to establish the correlation between the remote sensing data and the field survey data.

[0060] Establish the correlation between the real-time monitoring data and the field survey data: Receive data from a BeiDou monitoring terminal, which includes terminal ID, timestamp, spatial coordinates, displacement, and tilt angle data. Cache this data using a message queue (e.g., Kafka message queue) to obtain a cached monitoring data stream. Remove outliers from the cached monitoring data stream and perform linear interpolation to complete missing data, obtaining cleaned monitoring data (specifically, based on the cached monitoring data stream, use the 3σ criterion to remove outliers and perform linear interpolation to complete missing data to obtain cleaned monitoring data). Based on the timestamp and BeiDou monitoring terminal ID of the cleaned monitoring data, match it with the unified disaster point number in the database and use the Kriging interpolation algorithm to generate continuous surface data, thus establishing the correlation between the real-time monitoring data and the field survey data.

[0061] This implementation method establishes associations between remote sensing, real-time monitoring, and field survey data through a source-specific processing and precise matching strategy. The core of this approach uses a unified coding / terminal ID as the link, combined with data preprocessing and spatial / attribute matching to ensure association accuracy. Specifically, radar point clouds and remote sensing images undergo radiometric calibration and denoising. Feature points are extracted using the SIFT algorithm, registered using the RANSAC algorithm, and then the 3D contours of the disaster bodies are extracted using a region growing algorithm to generate vector surface files (containing spatial attributes). Based on the spatial coordinates of the vector surface files, spatial overlay analysis is used to match the corresponding disaster points with the unified coding (association is established if the contour and field survey coordinates meet the required overlay degree), thus achieving the binding of remote sensing data and field survey data. BeiDou terminal data (including terminal ID, coordinates, displacement, etc.) is cached in a Kafka message queue. Outliers are removed using the 3σ criterion, and missing values ​​are filled in using linear interpolation, resulting in cleaned data. The cleaned data is then matched with the unified code of the disaster point in the database (terminal ID and code are pre-bound), and the spatiotemporal dimensions are aligned using timestamps. Kriging interpolation is then used to generate continuous surface data, completing the association between real-time monitoring data and field survey data. Both associations are based on field survey data (unified code identifier). Remote sensing data is matched with the code through spatial contour overlay, and real-time monitoring data is matched with the code through terminal ID mapping, ultimately achieving accurate association of the three types of data for the same disaster point.

[0062] S104: Based on the aforementioned correlation and the aforementioned multi-source geological disaster survey data, spatial verification, logical verification, and correlation verification are performed through an intelligent quality control system to generate quality inspection results.

[0063] Here, the quality of multi-source geological disaster data (field survey, remote sensing, and real-time monitoring data) that has been linked is assessed. An intelligent system verifies the data from three dimensions to determine its quality. Specifically, based on multi-source data linked to the same disaster point, the intelligent quality control system performs three types of verifications: first, spatial verification, to check the accuracy of the data location, such as whether the field coordinates are within the remote sensing disaster area; second, logical verification, to check the rationality of the data content, such as whether the disaster type and deformation values ​​are consistent; and third, association verification, to check the effectiveness of the data association, such as whether there is isolated data and whether the spatiotemporal matching is correct. Finally, a detection result indicating whether the data quality meets the standards is generated.

[0064] In one implementation, a quality inspection result is generated by performing spatial verification, logical verification, and correlation verification through an intelligent quality control system. This may include: acquiring disaster point location data collected in the field, comparing the spatial positional relationship between the collection point and the vector boundary of the survey area in real time using a BeiDou positioning module, and generating a spatial verification result, wherein when the distance of the collection point beyond the vector boundary of the survey area is greater than a first distance (e.g., 5 meters), it is marked as a spatial anomaly; acquiring disaster point attribute data collected in the field, performing a legality check based on a preset value range and correlation logic rules, and generating a logical verification result; acquiring EXIF ​​information of exchangeable image files (EXIF format) of photos collected in the field, extracting the GPS coordinates and shooting time of the photos, and comparing them with the disaster point coordinates and collection time, and generating a correlation verification result, wherein when the deviation between the GPS coordinates of the photos and the disaster point coordinates is greater than a second distance (e.g., 10 meters), the upload is rejected.

[0065] Here, during spatial verification, BeiDou positioning is used to compare the vector boundary between the field collection point and the survey area in real time. If the distance exceeds 5 meters (first distance), spatial anomalies are marked. During logical verification, the legality of the disaster point attribute data collected in the field is checked according to the preset value range and association rules. During association verification, the GPS coordinates and shooting time of the collected photos are extracted from the EXIF ​​data and compared with the disaster point coordinates and collection time. If the GPS deviation exceeds 10 meters (second distance), the upload is rejected, and a verification result is generated.

[0066] The aforementioned value range is a valid interval set for each attribute data. Values ​​exceeding this range are considered abnormal. This is common for numeric and enumeration-type (drop-down selection) attributes. For example, the attribute "Number of people threatened by disaster" (numerical) has a preset value range of "positive integer (1 person or more)". If "0 people" or "-5 people" is entered during field data collection (recording a disaster point when no one is threatened is unreasonable) will trigger an abnormality and be marked as "value range error". Association rules are used for multiple attributes with dependencies to ensure they conform to real-world logic or business requirements. For example, the attributes "Disaster Type" and "Material Source Type" are associated with the preset rule: "If the disaster type is 'mudslide,' then the material source type must include loose soil and rocks." If "Disaster Type = Mudslide" is entered during field data collection, but "Bedrock Exposed" is selected for "Material Source Type" (mudslides require loose soil and rocks as a material source, and bedrock cannot form), a "logical contradiction" will be triggered, and an abnormality will be marked.

[0067] In another implementation, remote sensing image interpretation results can be obtained, and batch detection can be performed using a trained lightweight convolutional neural network (CNN) model to identify boundary delineation errors and land cover classification biases and generate interpretation anomaly reports; historical time series of monitoring data can be obtained, and the mean and standard deviation of the monitoring data can be calculated using a sliding window and trend analysis can be performed to identify data mutation anomalies and generate anomaly data notifications; here, when the monitoring data exceeds the range of mean plus or minus three times the standard deviation, it is marked as anomaly.

[0068] In this implementation, a pre-trained lightweight CNN model is used to perform batch detection of remote sensing image interpretation results, automatically identify boundary delineation errors (such as unclosed disaster body boundaries) and land cover classification biases (such as farmland being misclassified as debris flow areas), and generate interpretation anomaly reports; for the historical time series of monitoring data, the mean and standard deviation are calculated using a sliding window and the trend is analyzed. When the data exceeds the range of "mean ± 3σ", it is marked as a sudden anomaly and an anomaly data notification is generated.

[0069] Here, a lightweight convolutional neural network (CNN) model is pre-trained with a large number of correctly interpreted samples (such as accurately labeled hazard boundaries and land cover type labels), enabling it to quickly learn correctly interpreted features (such as the arc-shaped contour of landslide boundaries and the messy texture of debris flow areas). The lightweight design ensures efficient batch processing, avoiding speed bottlenecks caused by model complexity. When detecting boundary delineation errors, the model compares the hazard boundaries (such as landslide extent lines) in the interpretation results with the actual features of the image. If the boundary deviates significantly from the actual hazard area (such as including non-hazardous farmland within the landslide area, or boundary fractures / overlaps), it is marked as an anomaly. When detecting land cover classification bias, the model judges whether the interpreted land cover type (such as debris flow) matches the actual land cover features in the image. For example, if classification errors occur, such as "misclassifying forest land as a landslide area" or "misclassifying roads as debris flow accumulation areas," these are also marked as anomalies. Then, an interpretation anomaly report is generated, clearly listing the anomaly location (such as image coordinates) and anomaly type (boundary error / ...). Errors in classification and description (such as the landslide boundary being extra 10 meters into farmland) can be corrected by interpreters to facilitate targeted corrections.

[0070] Monitoring data (such as the displacement and tilt angle of the disaster body collected by the Beidou terminal) is the core data for dynamically tracking disaster changes. If there is a sudden change in data (such as the displacement suddenly jumping from 5mm to 50mm), it may be invalid data caused by equipment failure or signal interference, which needs to be investigated in time. This solution achieves anomaly detection through statistical analysis. First, the historical time series of monitoring data is obtained (such as the daily displacement data of a landslide point within one month). Then, a sliding window (which can be understood as a fixed time window, such as taking data for 7 consecutive days each time) is used to calculate the mean (average displacement) and standard deviation (data fluctuation degree) of the data within the window segment by segment. At the same time, the overall trend of data change is analyzed by sliding the window (such as whether the displacement is steadily increasing or fluctuating downward). If the monitoring data at a certain moment exceeds the range of the mean ± 3 times the standard deviation of the window, it means that the data deviates greatly from the normal fluctuation level of the same period, and is likely an outlier (such as erroneous data caused by temporary equipment failure), and is directly marked as a time series abrupt change anomaly. Finally, an anomaly data notification form is generated, clearly indicating the collection time, anomaly value (e.g., "Displacement on May 10th: 52mm, exceeding the mean ±3σ range"), and the monitoring point (e.g., "BeiDou Terminal ID: BD-2024-01"), facilitating equipment inspection or data verification by maintenance personnel. This solves the problems of human error in remote sensing interpretation and dynamic anomalies in monitoring data, and achieves automated, batch processing, significantly improving the efficiency and accuracy of geological disaster data quality control.

[0071] S105: Based on the quality inspection results, the database is dynamically updated using an incremental update mode, and version control records are generated to obtain a traceable database containing base state data and incremental data.

[0072] Here, the database is updated by only changing the changed data, and a log of each update is kept, ultimately forming a database that can be searched historically and is traceable.

[0073] In practical implementation, based on the initial survey results, a base-state snapshot containing complete backups of attribute database, spatial database, and data database can be generated to establish a base-state database; data change information can be obtained, including new disaster points, parameter updates, and data deletions, and the operation type, updated fields, old values, new values, updaters, and timestamps of the changed data can be recorded to generate incremental records; based on the base-state database and the incremental records, a version number can be set for the database and version backtracking can be supported by version number, update time, or operator to obtain the traceable database.

[0074] Here, based on the initial survey results, a full data snapshot mode is used to structurally integrate and completely back up core data from multiple dimensions, including attributes, space, and information, forming an immutable data baseline. Its core value is anchoring the data source, providing a comparable original reference system for all subsequent changes, and solving the underlying problem of ambiguous initial states in data tracing. Next, incremental control rules are established that fully cover all change types and record all elements: on the one hand, the capture scope for three core change scenarios—addition, update, and deletion—is clearly defined; on the other hand, a five-dimensional record model of operation-field-value-responsibility-time is used to transform abstract data changes into auditable structured information. This mechanism records only differences and avoids redundancy, reducing storage costs while ensuring precise location of each change. Finally, using version numbers as a connecting link, base-state data and incremental records are linked to construct a multi-dimensional backtracking index (version number / time / operator). Its core is to achieve complete reproduction of the database state at any historical node through the dynamic combination logic of "base state + corresponding increment". In essence, it is to establish a full-link traceability system for data from the initial state to the current state, and solve the management pain points of data change without trace and error correction without any way back.

[0075] In one embodiment, the method further includes: based on the hazard body parameters in the traceable database, the hazard body parameters including landslide thickness and soil mechanical properties, calling a three-dimensional continuous medium fast Lagrangian analysis program through a dynamic evaluation engine. The numerical simulation interface generates stability evaluation results. Based on the stability evaluation results and preset evaluation factors, including the number of people threatened, economic losses, and terrain slope, the weights are calculated using the analytic hierarchy process (AHP) and a risk level evaluation is performed to obtain the risk level results. Based on the risk level results, when the risk level escalates from medium risk to high risk, an early warning threshold mechanism is automatically triggered to generate early warning information and push it to the management platform.

[0076] Here, a data-simulation linkage mechanism is constructed using the core parameters of the disaster body (basic physical parameters such as landslide thickness and soil mechanical properties) from a traceable database as input. This is achieved by using a dynamic evaluation engine to call upon professional numerical simulation tools (…). This interface transforms structured data into stability assessment results, essentially establishing an automated transformation chain from "basic data → professional simulation → evaluation indicators," addressing the pain points of traditional assessments such as the disconnect between data and simulation and excessive manual intervention. By establishing a comprehensive evaluation system of core indicators and auxiliary factors, using stability assessment results as the core benchmark, and integrating multiple evaluation factors such as the number of people threatened, economic losses, and terrain slope, the system combines subjective weights with objective data through the analytic hierarchy process (AHP) to quantify risk levels. Based on the risk level results, dynamic early warning rules (such as medium to high risk escalation thresholds) are set, constructing an automated response chain of "level monitoring - threshold triggering - information push." ​​When the risk level reaches the early warning conditions, early warning information is automatically generated and pushed to the management platform. Thus, through engineering design, the entire process of accurate geological disaster risk assessment, dynamic monitoring, and timely early warning is automated, adapting to the high requirements of scientific rigor and timeliness in disaster prevention and control.

[0077] In one implementation, consensus nodes can be deployed on the Hyperledger Fabric consortium blockchain to generate blocks containing the operator, timestamp, and data hash value for data entry, quality audit, update, and deletion operations, thereby obtaining blockchain traceability records. Based on the data hash value, data integrity is verified through a hash value verification mechanism to achieve data tamper-proofing and full lifecycle traceability.

[0078] Here, consensus nodes are deployed on the Hyperledger Fabric consortium blockchain to perform operations such as data entry, review, update, and deletion, generating blocks containing the operator, timestamp, and data hash value and storing them on the chain to form an immutable blockchain traceability record. Through data hash value verification, the hash value generated by the current data is compared with the hash value recorded on the chain. If they match, the data has not been tampered with, ultimately achieving tamper-proof and traceable data throughout its entire lifecycle.

[0079] In one implementation, server status indicators, including CPU utilization, memory utilization, disk space utilization, and network bandwidth, can be monitored in real time using the Zabbix monitoring system. Multi-level alarm thresholds can be set to generate monitoring alarm information. Based on the monitoring alarm information, when the disk space utilization exceeds 85%, a data archiving strategy is triggered to compress and store historical data of more than 3 years to a remote node while retaining index information.

[0080] Here, Zabbix is ​​used to monitor the server's CPU, memory, disk, bandwidth, and other statuses, setting multiple thresholds and generating alarms. When the disk usage exceeds 85%, historical data of more than 3 years is automatically compressed and stored off-site, while the index is retained for easy querying, ensuring server operation and data security.

[0081] In one implementation, a database backup file to be submitted can be obtained, encrypted using the AES-256 algorithm, and a key digest can be generated to obtain encrypted submission data. The key digest includes the encryption algorithm, key ID, and file size. Based on the encrypted submission data, a dual backup mode of offline hard drive and blockchain notarization is used for submission to generate a submission record. The submission record includes the submitter, submission time, and data hash value and is written to the blockchain node.

[0082] Here, the database backup file to be submitted can be obtained, encrypted using the AES-256 algorithm, and a key digest containing the encryption algorithm, key ID, and file size can be generated to obtain encrypted submission data. The encrypted data is submitted through a dual method of offline hard drive + blockchain notarization, generating a submission record containing the submitter, time, and data hash value and writing it to the blockchain node to ensure that the data submission is secure and traceable.

[0083] like Figure 2 As shown in the embodiments of this disclosure, a geological disaster database construction apparatus 200 is also provided, comprising:

[0084] The database storage structure construction module 21 is used to acquire multi-source geological disaster survey data, construct a three-layer dual-drive database storage structure, and obtain a three-in-one storage structure including a relational database, a spatiotemporal database, and a knowledge base; the multi-source geological disaster survey data includes field survey data, remote sensing data, and real-time monitoring data;

[0085] The unified coding generation module 22 is used to generate a dynamic intelligent code for each geological disaster point based on the three-in-one storage structure, and obtain a unified code containing a basic code, a feature code and a time code. The basic code is an administrative division code, the feature code includes a disaster type code and a remote sensing interpretation code, and the time code is a time identifier.

[0086] The association establishment module 23 is used to establish the association between the field survey data, the remote sensing data and the real-time monitoring data based on the unified coding and through the multi-source data fusion engine to perform spatial proximity matching and attribute similarity comparison;

[0087] The quality inspection module 24 is used to generate quality inspection results by performing spatial verification, logical verification and correlation verification through an intelligent quality control system based on the correlation relationship and the multi-source geological disaster survey data.

[0088] The database generation module 25 is used to dynamically update the database and generate version control records based on the quality inspection results using an incremental update mode, thereby obtaining a traceable database containing base state data and incremental data.

[0089] The specific implementation process of the functions and roles of each unit in the above device can be found in the implementation process of the corresponding steps in the above method, and will not be repeated here.

[0090] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and 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 modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0091] Based on the same technical concept, this disclosure also provides an electronic device 300, which can be a server or a mobile device, specifically as described in the mobile terminal description, see reference 300. Figure 3 The diagram shown is a schematic representation of the structure of an electronic device according to an exemplary embodiment of this disclosure, comprising:

[0092] The processor 310, memory 320, and bus 330 are included. The memory 320 is used to store execution instructions and includes main memory 321 and external memory 322. The main memory 321, also known as internal memory, is used to temporarily store the operation data in the processor 310 and the data exchanged with external memory 322 such as hard disk. The processor 310 exchanges data with external memory 322 through main memory 321.

[0093] In this embodiment, the memory 320 is specifically used to store application code that executes the scheme of this disclosure, and its execution is controlled by the processor 310. That is, when the electronic device 300 is running, the processor 310 communicates with the memory 320 through the bus 330, or the processor 310 communicates with the memory 320 through other means, so that the processor 310 executes the application code stored in the memory 320, and then executes the steps of the geological disaster database construction method described in any of the foregoing embodiments.

[0094] The memory 320 may be, but is not limited to, random access memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), etc.

[0095] Processor 310 may be an integrated circuit chip with signal processing capabilities. The aforementioned processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor.

[0096] It is understood that the structures illustrated in the embodiments of this disclosure do not constitute a specific limitation on the electronic device 300. In other embodiments of this disclosure, the electronic device 300 may include more or fewer components than illustrated, or combine some components, or split some components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0097] This disclosure also provides a computer-readable storage medium including instructions stored thereon, wherein, when executed by a processor, the instructions perform the geological hazard database construction method described in any of the preceding embodiments. The storage medium may be a volatile or non-volatile computer-readable storage medium.

[0098] This disclosure also provides a computer program product, which stores a computer program. When the computer program is run by a processor, it executes the steps of the geological disaster database construction method provided in any of the above embodiments of this disclosure. For details, please refer to the above method embodiments, which will not be repeated here.

[0099] The aforementioned computer program product can be implemented through hardware, software, or a combination thereof. In one optional embodiment, the computer program product is specifically embodied in a computer storage medium, which can be a volatile or non-volatile computer-readable storage medium. In another optional embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.

[0100] Furthermore, embodiments of the subject matter and functional operation described in this specification can be implemented in the following ways: digital electronic circuits, tangibly embodied computer software or firmware, computer hardware including the structures disclosed in this specification and their structural equivalents, or combinations thereof. Embodiments of the subject matter described in this specification can be implemented as one or more computer programs, i.e., one or more modules of computer program instructions encoded on a tangible, non-transitory program carrier for execution by a data processing apparatus or for controlling the operation of a data processing apparatus. Alternatively or additionally, program instructions may be encoded on artificially generated propagation signals, such as machine-generated electrical, optical, or electromagnetic signals, which are generated to encode information and transmit it to a suitable receiving device for execution by the data processing apparatus. The computer storage medium may be a machine-readable storage device, a machine-readable storage substrate, a random or serial access memory device, or combinations thereof.

[0101] The processing and logic flows described in this specification can be executed by one or more programmable computers that execute one or more computer programs to perform corresponding functions by operating on input data and generating output. The processing and logic flows can also be executed by dedicated logic circuits—such as field-programmable gate arrays (FPGAs) or application-specific integrated circuits (ASICs), and the device can also be implemented as dedicated logic circuits.

[0102] Suitable computers for executing computer programs include, for example, general-purpose and / or special-purpose microprocessors, or any other type of central processing unit. Typically, the central processing unit receives instructions and data from read-only memory and / or random access memory. The basic components of a computer include a central processing unit for implementing or executing instructions and one or more memory devices for storing instructions and data. Typically, a computer will also include one or more mass storage devices for storing data, such as disks, magneto-optical disks, or optical disks, or the computer will be operatively coupled to such mass storage devices to receive data from or transfer data to them, or both. However, a computer is not required to have such devices. Furthermore, a computer can be embedded in another device, such as a mobile phone, a personal digital assistant (PDA), a mobile audio or video player, a game console, a Global Positioning System (GPS) receiver, or a portable storage device such as a Universal Serial Bus (USB) flash drive, to name a few.

[0103] Computer-readable media suitable for storing computer program instructions and data include all forms of non-volatile memory, media, and memory devices, such as semiconductor memory devices (e.g., EPROM, EEPROM, and flash memory devices), magnetic disks (e.g., internal hard disks or removable disks), magneto-optical disks, and CD-ROM and DVD-ROM disks. Processors and memory may be supplemented by or incorporated into dedicated logic circuitry.

[0104] While this specification contains numerous specific implementation details, these should not be construed as limiting the scope of any invention or the scope of the claims, but rather are primarily intended to describe features of specific embodiments of a particular invention. Certain features described in the various embodiments herein may also be implemented in combination in a single embodiment. Conversely, various features described in a single embodiment may also be implemented separately in various embodiments or in any suitable sub-combination. Furthermore, while features may function in certain combinations as described above and even initially claimed in this way, one or more features from a claimed combination may be removed from that combination in some cases, and a claimed combination may refer to a sub-combination or a variation thereof.

[0105] Similarly, although the operations are depicted in a specific order in the accompanying drawings, this should not be construed as requiring these operations to be performed in the specific order shown or sequentially, or requiring all illustrated operations to be performed to achieve the desired result. In some cases, multitasking and parallel processing may be advantageous. Furthermore, the separation of various system modules and components in the above embodiments should not be construed as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product or packaged into multiple software products.

[0106] Thus, specific embodiments of the subject matter have been described. Other embodiments are within the scope of the appended claims. In some cases, the actions recited in the claims may be performed in a different order and still achieve the desired result. Furthermore, the processes depicted in the drawings are not necessarily shown in a specific order or sequence to achieve the desired result. In some implementations, multitasking and parallel processing may be advantageous.

[0107] The above description is merely a preferred embodiment of this disclosure and is not intended to limit this disclosure. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for constructing a geological disaster database, characterized in that, include: By acquiring multi-source geological disaster survey data, a three-layer dual-drive database storage structure is constructed, resulting in a three-in-one storage structure including a relational database, a spatiotemporal database, and a knowledge base. The multi-source geological disaster survey data includes field survey data, remote sensing data, and real-time monitoring data; Based on the aforementioned three-in-one storage structure, a dynamic intelligent code is generated for each geological disaster point, resulting in a unified code that includes a basic code, a feature code, and a time code. The basic code is an administrative division code, the feature code includes a disaster type code and a remote sensing interpretation code, and the time code is a time identifier. Based on the unified coding, spatial proximity matching and attribute similarity comparison are performed through a multi-source data fusion engine to establish the association between the field survey data, the remote sensing data and the real-time monitoring data; Based on the aforementioned correlation and the multi-source geological disaster survey data, spatial verification, logical verification, and correlation verification are performed through an intelligent quality control system to generate quality inspection results. Based on the quality inspection results, the database is dynamically updated using an incremental update mode and version control records are generated to obtain a traceable database containing base state data and incremental data. The aforementioned construction of a three-tier, dual-drive database storage structure results in a unified storage structure comprising a relational database, a spatiotemporal database, and a knowledge base, including: Based on the field survey data, a hierarchical structure of master table and slave table is used for standardized storage to build a relational database. The master table of the relational database stores the unified number, spatial coordinates and disaster type of the disaster point, and the slave tables of the relational database store differentiated attribute parameters according to the disaster type. Obtain disaster point deformation monitoring data from the remote sensing data and the real-time monitoring data, store them in the form of triples containing spatial coordinates, timestamps and monitoring values, and generate a spatiotemporal data model; based on the spatiotemporal data model, establish a spatiotemporal index and integrate the functions of a temporal geographic information system (GIS) to obtain a spatiotemporal database that supports spatiotemporal range query and historical status backtracking; Obtain disaster type criteria and evaluation model parameters, construct a knowledge graph between disaster types, criteria, and evaluation models, and establish knowledge base nodes; based on the knowledge base nodes, associate numerical simulation parameters to obtain a knowledge base that supports intelligent reasoning and automatic matching of model parameters.

2. The method according to claim 1, characterized in that, A dynamic intelligent code is generated for each geological hazard point, resulting in a unified code that includes a basic code, a feature code, and a time code, including: Obtain the administrative division information of geological disaster sites, retrieve the 6-digit county-level code from the National Bureau of Statistics interface, and generate the aforementioned basic code; Based on the type attribute of the disaster point, the disaster type code table is called from the knowledge base for matching, and a 7-10 digit disaster type code is generated; Obtain the Interferometric Synthetic Aperture Radar (InSAR) deformation level data of the disaster point, connect it to the remote sensing interpretation system for level encoding, and generate the 11-14 bit remote sensing interpretation code; Obtain the current system date, extract the time information accurate to the day, and generate the 15-18 digit time code; Based on the base code, the disaster type code, the remote sensing decoding code, and the time code, a uniqueness check is performed and a historical coding mapping relationship is established to obtain an 18-bit unified code.

3. The method according to claim 1, characterized in that, By using a multi-source data fusion engine to perform spatial proximity matching and attribute similarity comparison, the correlation between the field survey data, the remote sensing data, and the real-time monitoring data is established, including: Acquire radar point cloud data and high-resolution remote sensing images, perform radiometric calibration, atmospheric correction and point cloud denoising to obtain preprocessed remote sensing data; Based on the preprocessed remote sensing data, the SIFT algorithm is used to extract corresponding feature points and the RANSAC algorithm is used for registration to obtain the registered point cloud data. Based on the registered point cloud data, a region growing algorithm is used to extract the three-dimensional contour of the disaster body and generate a vector surface file. Based on the spatial attributes of the vector surface file, spatial overlay analysis and the unified coding are used to establish the association between the remote sensing data and the field survey data; The system receives data from a BeiDou monitoring terminal, which includes the terminal ID, timestamp, spatial coordinates, displacement, and tilt angle data. The data is then cached using a message queue to obtain a cached monitoring data stream. Outliers are removed from the cached monitoring data stream and missing data is filled in by linear interpolation to obtain cleaned monitoring data. Based on the timestamp and BeiDou monitoring terminal ID of the cleaned monitoring data, it is matched with the unified disaster point number in the database and continuous surface data is generated using the Kriging interpolation algorithm to establish the correlation between the real-time monitoring data and the field survey data.

4. The method according to claim 1, characterized in that, The intelligent quality control system performs spatial verification, logical verification, and correlation verification to generate quality inspection results, including: The system acquires disaster location data collected in the field, compares the spatial positional relationship between the collection point and the vector boundary of the survey area in real time using the BeiDou positioning module, and generates spatial verification results. When the distance of the collection point beyond the vector boundary of the survey area is greater than a first distance, it is marked as a spatial anomaly. Acquire disaster site attribute data collected in the field, perform legality checks based on preset value ranges and association logic rules, and generate logical verification results; The system acquires EXIF ​​information from photos collected in the field, extracts the GPS coordinates and shooting time of the photos, compares them with the coordinates of the disaster point and the collection time, and generates a correlation verification result. The upload is rejected when the deviation between the GPS coordinates of the photos and the coordinates of the disaster point is greater than a second distance.

5. The method according to claim 1, characterized in that, Performing spatial verification, logical verification, and correlation verification through an intelligent quality control system also includes: The system acquires remote sensing image interpretation results, performs batch detection using a trained lightweight convolutional neural network (CNN) model, identifies boundary delineation errors and land cover classification biases, and generates interpretation anomaly reports. The historical time series of monitoring data is obtained, the mean and standard deviation of the monitoring data are calculated using a sliding window and trend analysis is performed to identify data mutations and anomalies and generate abnormal data notifications.

6. The method according to claim 1, characterized in that, An incremental update model is used to dynamically update the database and generate version control records, resulting in a traceable database containing base-state data and incremental data, including: Based on the initial survey results, a base state snapshot containing complete backups of the attribute library, spatial library, and data library is generated to establish a base state database; Acquire data change information, including newly added disaster points, parameter updates, and data deletions; record the operation type, updated fields, old values, new values, updaters, and timestamps of the changed data; and generate incremental records. Based on the base state database and the incremental records, a version number is set for the database and version backtracking is supported by version number, update time, or operator to obtain the traceable database.

7. The method according to claim 1, characterized in that, The method further includes: Based on the hazard body parameters in the traceable database, including landslide thickness and soil and rock mechanical properties, a three-dimensional continuous medium fast Lagrangian analysis program is invoked through a dynamic evaluation engine. Numerical simulation interface to generate stability evaluation results; Based on the stability evaluation results and the preset evaluation factors, including the number of people threatened, economic losses and terrain slope, the weights are calculated using the analytic hierarchy process and the risk level is evaluated to obtain the risk level results. Based on the risk level results, when the risk level escalates from medium risk to high risk, an early warning threshold mechanism is automatically triggered to generate early warning information and push it to the management platform.

8. A geological disaster database construction device, characterized in that, include: A database storage structure construction module is used to acquire multi-source geological disaster survey data and construct a three-layer dual-drive database storage structure, resulting in a unified storage structure including a relational database, a spatiotemporal database, and a knowledge base. The multi-source geological disaster survey data includes field survey data, remote sensing data, and real-time monitoring data. When constructing the three-layer dual-drive database storage structure, a unified storage structure including a relational database, a spatiotemporal database, and a knowledge base is obtained, it is used to construct a relational database based on the field survey data, using a master-slave table hierarchical structure for standardized storage. The master table of the relational database stores the unified disaster point number, spatial coordinates, and disaster type. The slave tables of the relational database... The table stores differentiated attribute parameters according to disaster type; it acquires disaster point deformation monitoring data from the remote sensing data and the real-time monitoring data, and stores it in the form of triplets containing spatial coordinates, timestamps, and monitoring values ​​to generate a spatiotemporal data model; based on the spatiotemporal data model, it establishes a spatiotemporal index and integrates the functions of a temporal geographic information system (GIS) to obtain a spatiotemporal database that supports spatiotemporal range query and historical state backtracking; it acquires disaster type criteria and evaluation model parameters, constructs a knowledge graph between disaster types, criteria, and evaluation models, and establishes knowledge base nodes; based on the knowledge base nodes, it associates numerical simulation parameters to obtain a knowledge base that supports intelligent reasoning and automatic model parameter matching; A unified coding generation module is used to generate a dynamic intelligent code for each geological disaster point based on the three-in-one storage structure, resulting in a unified code containing a basic code, a feature code, and a time code. The basic code is an administrative division code, the feature code includes a disaster type code and a remote sensing interpretation code, and the time code is a time identifier. The association establishment module is used to establish the association between the field survey data, the remote sensing data and the real-time monitoring data based on the unified coding and through the multi-source data fusion engine to perform spatial proximity matching and attribute similarity comparison; The quality inspection module is used to generate quality inspection results by performing spatial verification, logical verification, and correlation verification through an intelligent quality control system based on the correlation relationship and the multi-source geological disaster survey data. The database generation module is used to dynamically update the database and generate version control records based on the quality inspection results using an incremental update mode, thereby obtaining a traceable database containing base state data and incremental data.

9. An electronic device, characterized in that, It includes a processor and a memory, the memory being used to store computer instructions, and the processor being used to execute the computer instructions to implement the geological disaster database construction method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions, which, when executed, implement the geological disaster database construction method according to any one of claims 1 to 7.

Images