Marine geological environment multi-dimensional data reconstruction and dynamic visualization mapping method and system

Abstract

The invention discloses a marine geological environment multi-dimensional data reconstruction and dynamic visualization mapping method and system, and belongs to the technical field of marine information technology and data visualization. In order to solve the problems that in the prior art, standard data asset structure is isolated, heterogeneous data deep fusion is difficult, and mass data visualization performance is bottleneck, the method mainly adopts the following technical scheme that standardized multi-source data is received, a data volume mapping rule is established, discrete observation data is reconstructed and fused into a unified and continuous three-dimensional data volume, and the three-dimensional data volume is obtained; and executing a three-dimensional space pattern recognition algorithm on the three-dimensional data volume to recognize a feature entity, and providing on-demand query and multi-resolution access services through a visual mapping interface. According to the method, efficient automatic modeling and high-performance interactive analysis of mass and multi-dimensional marine geological environment data can be realized.

Description

Technical Field

[0001] This invention belongs to the field of marine information technology and data visualization technology, and in particular relates to a method and system for multi-dimensional data reconstruction and dynamic visualization mapping of marine geological environment. Background Technology

[0002] In modern marine geological research and resource exploration, the deep integration and efficient interactive analysis of massive amounts of observation data are crucial for scientific discovery and precise decision-making. With the development of marine exploration technology, advanced data collaborative acquisition and real-time quality control systems can now produce standardized data products with unified spatiotemporal benchmarks and reliable quality, providing high-quality data sources for subsequent analysis.

[0003] However, with the increasing complexity of exploration tasks and the explosive growth of data volume, new and more severe challenges remain in the areas of in-depth data utilization and visualization analysis: The structural isolation of data assets: Although the data format is standardized, they are essentially still an assembly of discrete datasets with multiple physical dimensions and independent data structures (such as one-dimensional profiles, two-dimensional images, and three-dimensional point clouds). Existing technologies lack a systematic method for automatically internalizing and reconstructing these discrete data into a unified data volume that can be continuously queried and subjected to complex calculations in three-dimensional space.

[0004] The bottleneck of deep fusion of heterogeneous data: The current technological bottleneck lies in efficiently and automatically fusing observational data with vastly different dimensions and structures into a unified, continuous, and spatially interpolable 3D gridded data model. Existing processes heavily rely on domain experts for complex, time-consuming, and unreproducible manual modeling, lacking robust and efficient automated system-level solutions.

[0005] Performance challenges in massive data visualization: Traditional visualization analysis workflows follow a heavy client-side model of "overall loading - local rendering," which leads to a sharp drop in performance and severe lag when dealing with terabyte-scale datasets. Existing architectures lack a service-oriented mechanism to decouple the backend massive data model from the frontend visualization application, making it impossible to support real-time, dynamic, on-demand querying and smooth exploration of massive datasets. Summary of the Invention

[0006] This invention proposes a method and system for multidimensional data reconstruction and dynamic visualization mapping of marine geological environment. It aims to solve the problems of structural isolation of standard data assets, difficulty in deep integration of heterogeneous data, and performance bottleneck of visualization of massive data in the prior art by constructing a unified and continuous three-dimensional data volume, performing advanced three-dimensional feature entity recognition, and providing dynamic visualization mapping services. Ultimately, it achieves efficient and automated modeling and high-performance interactive analysis of massive and multidimensional marine geological environment data.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows.

[0008] A method for reconstructing and dynamically visualizing marine geological environment data includes the following steps: It receives one or more standardized multi-source data sources with a unified spatiotemporal reference and establishes data volume mapping rules according to preset configuration; Based on the data volume mapping rules, the discrete observation data in the multi-source data sources are reconstructed, and the reconstructed data is fused into a unified and continuous three-dimensional data volume covering the target area; A three-dimensional spatial pattern recognition algorithm is executed on the three-dimensional data volume to identify and mark three-dimensional feature entities that meet preset attribute conditions; A visualization mapping interface is established for the three-dimensional data volume and three-dimensional feature entities, supporting on-demand querying and multi-resolution data access.

[0009] Furthermore, the steps for establishing data volume mapping rules based on preset configurations include: Provides a mapping rule configuration interface for users to configure in a graphical or script-based manner; The mapping rule configuration interface defines the correspondence between physical quantity variables in different input data sources and voxel attribute fields in the three-dimensional data volume. The mapping relationship is then solidified into a machine-readable configuration file.

[0010] Furthermore, the discrete observation data in the multi-source data source includes one-dimensional profile points, two-dimensional image pixel centers, and three-dimensional point cloud points.

[0011] Furthermore, the step of reconstructing the discrete observation data from the multi-source data source includes: Unified spatial carrier generation: Based on the user-defined geographical range and target resolution, a three-dimensional regular mesh or voxel model is generated in three-dimensional geographic space as a unified spatial carrier for heterogeneous data; Multidimensional data spatial registration: Parse the metadata of each input data source, which includes spatial coordinates and lever values, and project and register the discrete observation data of different dimensions to the world coordinate system of the three-dimensional regular grid or voxel model; Multi-field joint spatial interpolation: Based on the registered discrete observation points, a spatial interpolation algorithm is invoked to calculate one or more physical attribute values ​​for each voxel node in the three-dimensional regular mesh or voxel model, thereby obtaining a three-dimensional data field covering the target area.

[0012] Furthermore, the steps for fusing the reconstructed data include: In the three-dimensional data field, the same voxel is used as a multi-attribute voxel unit. At least one physical attribute among temperature, salinity, formation acoustic impedance, and magnetic anomaly intensity, which are obtained by interpolation calculation from different source data, is used as the attribute set of the voxel unit and fused together. The set of all voxel units constitutes a unified three-dimensional data volume.

[0013] Furthermore, the steps for performing recognition using a 3D spatial pattern recognition algorithm include: In the three-dimensional data volume, a three-dimensional spatial pattern recognition algorithm is used to query and analyze the voxel data to identify continuous three-dimensional spatial regions that meet the preset physical attribute threshold conditions. The continuous three-dimensional spatial region is extracted as an independent three-dimensional entity, and the geometric boundary, volume, and center path parameters of the three-dimensional entity are calculated.

[0014] Furthermore, the preset attribute conditions refer to a set of physical attribute thresholds used to filter continuous three-dimensional spatial regions. The set of thresholds includes temperature thresholds, salinity thresholds, acoustic impedance thresholds, magnetic anomaly intensity thresholds, and combinations thereof.

[0015] Furthermore, the steps for establishing a visual mapping interface include: A standardized data service interface is constructed to provide the three-dimensional data volume and three-dimensional feature entities to the front-end application in the form of a visual mapping. It supports parameterized on-demand queries, extracting slices or data subsets from the 3D data volume or 3D feature entity according to the query request and returning them; The three-dimensional data volume is organized using a spatial index structure. Based on the viewpoint distance or target resolution parameter in the query request, data of the corresponding level is selected from the spatial index structure for aggregation or sampling and then returned.

[0016] A system for reconstructing and dynamically visualizing marine geological environment data includes: The data body construction input module is used to receive standardized multi-source data sources and provides a mapping rule configuration interface to establish data body mapping rules; The multidimensional data volume reconstruction and fusion engine is used to reconstruct discrete observation data according to data volume mapping rules, and fuse the reconstructed data into a unified and continuous three-dimensional data volume covering the target area. The 3D data volume feature entity recognition module is used to execute a 3D spatial pattern recognition algorithm on a 3D data volume to identify and mark 3D feature entities that meet preset attribute conditions. The Dynamic Visualization Mapping API service module is used to provide 3D data volumes and 3D feature entities in the form of visual mappings, and supports on-demand querying and multi-resolution data access.

[0017] Furthermore, it also includes a data volume management and indexing module, which is used to persistently store and manage the three-dimensional data volume generated by the multi-dimensional data volume reconstruction and fusion engine, and to build and maintain a spatial index structure to support on-demand querying and multi-resolution data access of the dynamic visualization mapping API service module. The spatial index structure selects the corresponding level of data for aggregation or sampling based on the viewpoint distance or target resolution parameter in the query request and then returns it.

[0018] Compared with the prior art, the present invention has at least the following beneficial effects.

[0019] (1) Automated reconstruction and efficiency improvement of discrete data to unified model: This invention constructs an end-to-end automated processing pipeline from standardized multi-source data input to a unified three-dimensional data volume service that can be dynamically queried. Its multi-dimensional data volume reconstruction and fusion engine shortens the complex script writing, manual interpolation and model building work that traditionally requires domain experts to spend several days to several weeks to the level of automated processing, thereby improving the efficiency of marine geological data from usability to ease of use and high efficiency, and ensuring the objectivity, consistency and reproducibility of the modeling process, solving the problems of low efficiency and strong subjectivity in traditional manual modeling.

[0020] (2) Providing a unified data foundation: This invention integrates discrete observation data with different physical dimensions and data structures, including one-dimensional profiles, two-dimensional images, and three-dimensional point clouds, into a unified and continuous three-dimensional digital twin model through a data volume reconstruction and interpolation fusion engine. In this unified data volume, different physical fields, such as temperature, salinity, and magnetic anomalies, are coupled as multi-attribute voxels in the same spatial grid and are no longer isolated variables. This provides a solid and reliable data foundation for realizing cross-physical field collaborative analysis and the automatic identification and quantitative assessment of three-dimensional geological bodies (such as hydrothermal plumes and ore bodies), fundamentally breaking through the limitations of traditional data silos.

[0021] (3) Dynamic Visualization of Massive Data: This invention utilizes a dynamic visualization mapping data service module, leveraging the decoupling of the backend data model and frontend application architecture, an on-demand query mechanism, and Level of Detail (LOD) technology to overcome the performance limitations of traditional overall loading and local rendering modes. Users can explore massive 3D data volumes in real-time, smoothly, and from multiple angles through lightweight clients such as web browsers without downloading the complete TB-level dataset, including arbitrary slicing, drill-down, and stepless scaling operations. This "Data as a Service (DaaS)" model significantly improves users' analytical efficiency and decision-making capabilities, supports cross-regional and cross-platform collaborative data analysis, and promotes the sharing and utilization of scientific data.

[0022] (4) Enhanced Data Analysis Capabilities: This invention elevates the analysis object from discrete data points to continuous three-dimensional data volumes through a three-dimensional data volume feature entity recognition module. Utilizing volumetric data mining algorithms, it automatically identifies and marks continuous three-dimensional spatial regions that meet preset physical attribute conditions, forming independent three-dimensional feature entities. This process extends from data quality assessment to three-dimensional geological body semantic recognition, achieving automated spatial entity labeling and parameter calculation, including geometric boundaries, volumes, and center paths. This provides users with high-level three-dimensional analysis results, reduces the burden of manual interpretation, and enhances the intelligence level of marine geological analysis. Attached Figure Description

[0023] Figure 1 This is a diagram illustrating the overall structure of a multidimensional data reconstruction and dynamic visualization mapping system for marine geological environment according to the present invention.

[0024] Figure 2 This is the overall flowchart of the marine geological environment multidimensional data reconstruction and dynamic visualization mapping method in Example 1.

[0025] Figure 3 This is a schematic diagram of the multidimensional data volume reconstruction and fusion process in Example 1.

[0026] Figure 4 This is a schematic diagram of the dynamic visualization mapping API service and interaction in Implementation Example 1. Detailed Implementation

[0027] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. The embodiments given are only for illustrating the present invention and are not intended to limit the scope of the present invention.

[0028] The core of this invention lies in constructing an integrated data fusion and application service framework that combines multi-dimensional data volume reconstruction, 3D feature entity recognition, and dynamic visualization mapping. This framework addresses the problems of structural isolation, complex fusion processes, and limited visualization performance inherent in existing technologies for deep fusion and interactive analysis of high-quality standardized data. This invention transforms the traditional file-centric discrete data analysis model into a model-centric data processing model, and provides intelligent data application capabilities as a data-as-a-service.

[0029] To achieve the above objectives, this invention provides a method and system for multidimensional data reconstruction and dynamic visualization mapping of marine geological environments. This system executes the processing steps of the method, and its structure is as follows: Figure 1 As shown, it includes the following modules.

[0030] The data body construction input module is used to receive standardized multi-source data sources and provides a mapping rule configuration interface to establish data body mapping rules.

[0031] The multidimensional data volume reconstruction and fusion engine is used to reconstruct discrete observation data according to data volume mapping rules, and then fuse the reconstructed data into a unified and continuous three-dimensional data volume covering the target area.

[0032] The 3D data volume feature entity recognition module is used to execute a 3D spatial pattern recognition algorithm on a 3D data volume to identify and mark 3D feature entities that meet preset attribute conditions.

[0033] The Dynamic Visualization Mapping API service module is used to provide 3D data volumes and 3D feature entities in the form of visual mappings, and supports on-demand querying and multi-resolution data access.

[0034] The data volume management and indexing module, as an optional module, is used to persistently store and manage the three-dimensional data volume generated by the multi-dimensional data volume reconstruction and fusion engine, and to build and maintain a spatial index structure to support on-demand querying and multi-resolution data access of the dynamic visualization mapping API service module. The spatial index structure selects data of the corresponding level for aggregation or sampling based on the viewpoint distance or target resolution parameter in the query request and then returns it.

[0035] This system is based on several key functional modules that can work together, including a data volume construction input module, a multi-dimensional data volume reconstruction and fusion engine, a three-dimensional data volume feature entity recognition module, a dynamic visualization mapping API service module, and a data volume management and indexing module. The specific details are as follows.

[0036] First, the multidimensional data volume reconstruction and fusion engine is the technological cornerstone of the entire system. This engine first receives one or more standardized multi-source data sources with a unified spatiotemporal reference. Through a mapping rule configuration interface (supporting graphical or scripted configuration), it establishes a mapping relationship between the physical quantity variables of the input data source (e.g., temperature, salinity, magnetic anomaly) and the attribute fields of multi-attribute voxels in the target 3D data volume. This mapping relationship is solidified in a machine-readable format such as XML or JSON. Based on this, the system performs unified spatial carrier generation: according to user-defined geographical range and target resolution parameters, it automatically creates a 3D regular mesh or voxel model in a virtual 3D geographic space, serving as a unified carrier for all heterogeneous data.

[0037] Secondly, the system performs multidimensional data spatial registration: it automatically parses the metadata of each input data source (including spatial coordinates and lever values), and accurately and automatically projects and registers discrete observation data from different dimensions (including one-dimensional profile points, two-dimensional image pixel centers, and three-dimensional point cloud points) into the world coordinate system of the voxel model, assigning a unique spatial position to each observation point.

[0038] Secondly, the system performs multi-field joint spatial interpolation: based on discrete observation points already registered to the voxel model, it calls preset or user-selected interpolation algorithms (such as Kriging interpolation, inverse distance weighted interpolation, or radial basis function interpolation) to calculate the physical attribute values ​​of each voxel node in the 3D mesh. Multiple physical attributes from different source data (such as temperature, salinity, formation acoustic impedance, magnetic anomaly intensity, etc.) are uniformly fused into multi-attribute voxel elements, ultimately forming a continuous, unified, multi-field coupled 3D voxelized data volume covering the target area. This achieves the reconstruction and fusion of discrete, sparse observation data into a structured digital twin model.

[0039] Secondly, the 3D data volume feature entity recognition module is the core component for achieving intelligent analysis. This module elevates the analysis object from raw discrete data points or 2D images to a constructed continuous 3D data volume. It employs 3D spatial pattern recognition or volume data mining algorithms to perform advanced queries and analyses within the data volume. For example, it can automatically identify all continuous 3D spatial regions that meet specific physical property combination thresholds (e.g., temperature above T1 and salinity below S1) through 3D connected component search or level set segmentation algorithms, and extract them as independent 3D entities. For each extracted entity, the system further calculates its geometric boundaries, volume, center path, and other parameters, and attaches semantic tags (e.g., "hydrothermal plume," "high-density ore body," or "shallow gas sac"), thereby achieving a leap from data quality assessment to 3D geological body semantic recognition and solidifying geological interpretation expert knowledge into an advanced analysis process that can be automatically executed by machines.

[0040] Finally, the Dynamic Visualization Mapping API service module acts as a bridge connecting the massive backend data model with the frontend visualization application, providing a standardized data-as-a-service interface. This module, through architectural decoupling, allows terabyte-scale complex data models to be managed independently of lightweight frontend tools (such as web browsers and GIS software). The system supports rich parametric on-demand queries; the frontend application does not need to download the complete dataset. Instead, it sends specific query commands, and the API service module extracts corresponding slices or subsets from the 3D data volume in real time and returns them in lightweight data formats (such as JSON, PNG, or binary streams).

[0041] To enable seamless interaction with massive amounts of data, this module organizes and manages the 3D data volume based on a spatial index structure (preferably an octree or data pyramid). When a query request is received, the system dynamically selects data of the appropriate level for aggregation or sampling based on the viewpoint distance or target resolution parameters in the request, providing the front-end application with the best resolution data, supporting stepless scaling and smooth navigation, thereby completely revolutionizing the traditional visualization and analysis mode of marine geological data.

[0042] This invention can be applied to various marine geological research and engineering applications, including 3D modeling of submarine hydrothermal activity zones, digital twin construction of submarine pipeline corridors, quantitative assessment of marine mineral resources, and analysis of marine environmental change trends. The following embodiments further illustrate the technical solutions and implementation methods of this invention in specific applications.

[0043] Example 1: 3D modeling and feature analysis of submarine hydrothermal plumes.

[0044] This embodiment will elaborate on the application of the method of the present invention in a typical marine geochemical and physical oceanography research scenario—three-dimensional modeling and feature analysis of hydrothermal plumes. In this scenario, the objective is to deeply fuse and intelligently analyze multiple observation datasets from different sensors, describing the same geographical area but with varying data structures, to construct a unified three-dimensional model that accurately characterizes the spatial morphology, physicochemical properties, and interactions with the surrounding environment of hydrothermal plumes. This embodiment will be based on... Figure 2 The overall process shown below provides a detailed explanation of the technical steps of this invention.

[0045] In this scenario, after a marine scientific expedition, researchers acquired a series of high-quality, standardized data products from a suspected area of ​​hydrothermal vent activity. These data products were generated using advanced acquisition systems, possess a unified spatiotemporal reference, and are stored in the industry-standard NetCDF format. However, these files are structurally independent and cannot be directly analyzed together.

[0046] Step 1: System initialization and data body mapping rule configuration.

[0047] Before performing the data volume reconstruction task, the first step is to use the data volume construction input module of the system described in this invention to initialize and configure the system. The core of this step is not simply data reading, but rather to pre-define a detailed, accurate, and machine-readable data volume mapping rule for the subsequent fully automated data fusion and modeling process.

[0048] In this embodiment, the system receives the following three independent NetCDF format data files as input sources: Multibeam topography data (topography_data.nc): This file contains high-resolution 3D point cloud data of seabed topography acquired by a multibeam echo sounder. It stores the longitude, latitude, and depth values ​​of each sounding point, forming a discrete description of the seabed morphology.

[0049] Temperature, salinity, and depth (CTD) profile data (ctd_profiles.nc): This file records vertical water profile data acquired at multiple locations by shipborne or towed CTDs. The data structure consists of a series of one-dimensional arrays, each corresponding to a profile and containing temperature and salinity observations varying with depth.

[0050] Water column acoustic backscatter data (watercolumn_backscatter.nc): This file stores data recorded by the water imaging function (WCI) of a multibeam echo sounder. Its data structure is a series of two-dimensional fan-shaped images, showing the intensity of backscattering generated by suspended particles or density interfaces encountered by sound waves as they propagate in the water. It is a powerful tool for directly detecting water anomalies such as hydrothermal plumes.

[0051] After determining the input data source, the operator begins configuration through a graphical or scripted mapping rule configuration interface provided by the system. This interface allows users to precisely define how to unify multiple discrete and heterogeneous data sources into a single target 3D data volume. The entire configuration process aims to answer three core questions: 1. What is the spatial extent and level of detail of the final 3D data volume to be constructed? 2. Which physical quantities from each input file should be used as the basis for modeling? 3. Which attribute fields in the final data volume should these physical quantities be assigned to?

[0052] Specifically, the process of establishing data volume mapping rules includes: Define the spatial parameters of the target data volume: The operator first needs to define the basic framework of the final generated 3D data volume. This includes: Geographic Bounding Box: Define a three-dimensional cuboid spatial bounding box that completely encompasses all features of interest, based on the area covered by all input data. For example, define the latitude and longitude coordinates of its southwest and northeast corners, as well as the depth values ​​of its top and bottom faces.

[0053] Voxel resolution: This sets the dimensions of the voxels—the basic units constituting the 3D data volume—in the X, Y, and Z directions. The choice of resolution directly affects the model's detail and computational complexity. In this embodiment, considering the spatial scale of the hydrothermal plume, an anisotropic resolution can be set, for example, 5 meters in the horizontal direction (X, Y) and 2 meters in the vertical direction (Z).

[0054] Establish source-target variable mapping: Next, the operator needs to establish a mapping path from the original input file to the target data volume properties for each physical quantity that is to be incorporated into the model, including the following:

[0055] Specify the source file and the names of variables within it, for example, the temperature variable in the ctd_profiles.nc file.

[0056] Specify a uniform and meaningful target attribute name for this physical quantity in the final data body, such as body_temperature.

[0057] Repeat this process to establish mappings for all physical quantities that need to be fused, such as salinity (salinity -> body_salinity) and backscatter intensity (backscatter -> body_acoustic_intensity).

[0058] Ultimately, all this configuration information is solidified into a standardized, machine-readable configuration blueprint. This blueprint is preferably stored in JSON format, as its clear structure facilitates system parsing and user access. The following shows an example of a JSON-formatted data body mapping rule file generated for this embodiment: { "jobID": "Hydrothermal_Plume_Modeling_01", "targetVolumeDefinition": { "boundingBox": { "southwest_corner": {"lon": 121.50, "lat": 24.80, "depth": 1400.0}, "northeast_corner": {"lon": 121.52, "lat": 24.82, "depth": 1650.0} }, "voxelResolution": { "x_resolution": 5.0, "y_resolution": 5.0, "z_resolution": 2.0, "unit": "meter" }, "coordinateSystem": "EPSG:4326" }, "sourceDataMapping": [ { "sourceFile": "ctd_profiles.nc", "sourceVariable": "temperature", "interpolationMethod": "kriging", "targetAttribute": "body_temperature" }, { "sourceFile": "ctd_profiles.nc", "sourceVariable": "salinity", "interpolationMethod": "kriging", "targetAttribute": "body_salinity" }, { "sourceFile": "watercolumn_backscatter.nc", "sourceVariable": "backscatter_intensity", "interpolationMethod": "inverse_distance_weighting", "targetAttribute": "body_acoustic_intensity" }, { "sourceFile": "topography_data.nc", "sourceVariable": "depth", "targetAttribute": "seabed_depth" } ] } It is worth noting that, in addition to basic mapping relationships, this configuration blueprint can also contain more advanced instructions, such as specifying different spatial interpolation algorithms for different physical quantities. This provides more granular control for the subsequent data volume reconstruction process.

[0059] By completing the above steps, the system of this invention lays a solid foundation for subsequent fully automated data processing. This configuration blueprint transforms multiple logically and structurally independent discrete data files into a clear and unambiguous structured construction guide for building a unified three-dimensional data volume. This step is a key prerequisite for realizing the paradigm shift from tedious manual modeling to efficient automated reconstruction, and is the technical starting point for implementing the entire inventive method.

[0060] Step 2: Multidimensional data volume reconstruction and fusion.

[0061] After system initialization is completed and the data volume mapping rules generated in step 1 are loaded, the core module of the system described in this invention—the multidimensional data volume reconstruction and fusion engine—begins execution. This step is the core technology of this invention, and its goal is to creatively reconstruct and fuse the aforementioned multiple discrete and heterogeneous input data sources into a unified, continuous, multi-attribute three-dimensional data volume through a series of automated calculation processes. The principle of this process can be combined with... Figure 3 This understanding will fundamentally change the organization of data, providing a structured data model capable of continuous spatial operations for subsequent 3D feature recognition and interactive analysis.

[0062] Specifically, the process of reconstructing and fusing discrete observation data into a unified, continuous three-dimensional data volume includes the following sequentially executed sub-steps: (1) Generation of unified spatial carriers.

[0063] This is the first step in the reconstruction process, aiming to create a standardized 3D spatial base to support all subsequent data to be merged. The system first parses the targetVolumeDefinition section defined in the data volume mapping rule file. Based on the geographic bounding box and voxel resolution parameters set therein, the system automatically constructs a 3D regular mesh, also known as a voxel model, in the computer's virtual 3D geographic space.

[0064] In this embodiment, the system generates a voxel grid covering a spatial range from 121.50° to 121.52° longitude, 24.80° to 24.82° latitude, and 1400.0 meters to 1650.0 meters depth, based on configuration parameters. This grid consists of a large number of closely packed cubic cells measuring 5 meters × 5 meters × 2 meters. This voxel grid provides a unified target coordinate system and structured container for all observation data from different sources and with varying structures, serving as a prerequisite for achieving spatial alignment and fusion of multi-source data.

[0065] (2) Spatial registration of multidimensional data.

[0066] After the unified spatial carrier is generated, the system begins processing the various input data sources, accurately projecting and registering the discrete observation data contained within them to the world coordinate system of the voxel model created above. The system automatically parses the metadata of each input NetCDF file, reading the geographic coordinate information (such as longitude, latitude, and depth) and necessary coordinate system transformation parameters contained therein. This process is detailed below.

[0067] For the 3D terrain point cloud in topography_data.nc, the system reads the 3D coordinates of each point and maps them to the corresponding or nearest voxel cell in the voxel grid.

[0068] For the one-dimensional vertical profile data in ctd_profiles.nc, the system first reads the geographical location (latitude and longitude) of each profile, and then, along the vertical direction, registers each temperature and salinity observation point (each with its corresponding depth value) on the profile to the corresponding voxel cell in the voxel grid.

[0069] For the two-dimensional acoustic image of water in watercolumn_backscatter.nc, the system calculates the precise position of each pixel in the two-dimensional image in the three-dimensional world coordinate system based on the sensor position, attitude, and opening angle and direction information of the acoustic wave sector recorded in the file, and registers it to the corresponding voxel unit.

[0070] After this step, the discrete observation points, originally stored in different files and following different data structures, are all assigned a unique spatial location in the unified three-dimensional voxel model. At this point, in the voxel mesh, only those voxels that exactly contain one or more original observation points have physical property values, while the vast majority of voxels are still null values, and the entire data field remains spatially discontinuous.

[0071] (3) Multi-field joint spatial interpolation.

[0072] To address the spatial discontinuity of the data, the system then performs multi-field joint spatial interpolation. The goal of this step is to use data from known observation points and mathematical methods to estimate a reasonable physical property value for all null voxels in the voxel grid that are not directly covered by observations.

[0073] The system calls the corresponding algorithm library to perform calculations based on the spatial interpolation method specified for each target attribute in the data volume mapping rule file.

[0074] For the target attributes body_temperature and body_salinity, the mapping rules specify the use of Kriging interpolation. This is a geostatistically optimal interpolation method based on spatial autocorrelation. The system takes all registered CTD temperature observations as input, analyzes their spatial variability, and then calculates an optimal unbiased estimate of temperature with minimum variance for each voxel node in the grid. The same process is performed on the salinity data.

[0075] For the target attribute body_acoustic_intensity, the mapping rule specifies the use of Inverse Distance Weighting (IDW). Based on the principle that "things that are close together are more relevant", this method assigns a weighted average of several nearest acoustic backscattering observations to each voxel to be interpolated, with the weights proportional to the inverse of the distance.

[0076] By performing this series of interpolation calculations, the system effectively "fills" the sparse, discrete observation data into all voxels of the entire three-dimensional data volume. Thus, the data undergoes a qualitative change from "discrete" to "continuous" in space, forming one or more continuous three-dimensional data fields covering the entire target area.

[0077] (4) Multi-attribute voxel fusion.

[0078] After calculating the temperature, salinity, acoustic intensity, and other attribute values ​​for each voxel in the mesh, the system performs the final fusion step. In this step, each voxel in the mesh is treated as a multi-attribute data voxel.

[0079] The system integrates the multiple independent continuous data fields generated by the interpolation in the previous step. For the same voxel with the same spatial location, the system treats it as a unified data structure and fuses all calculated physical properties (such as body_temperature: 1.85, body_salinity: 34.62, body_acoustic_intensity: -45.2) as the attribute set of that voxel. Simultaneously, topographic data from topography_data.nc can be used to define the effective lower boundary of the data volume; that is, all voxels located below the seabed can be marked as invalid.

[0080] Ultimately, the collection of all these multi-attribute voxels together constitutes a unified, continuous, multi-field coupled 3D data volume of the marine geological environment. This data volume is stored in the system's backend in an efficient, rapidly indexable format (such as a block-compressed binary array). It is no longer a simple collection of multiple independent files, but a high-fidelity 3D digital model of the marine geological environment with a unified internal structure, spatial continuity, and coexistence of multi-physics information. This innovative reconstruction process lays an indispensable data foundation for subsequent advanced 3D spatial analysis and intelligent feature recognition.

[0081] Step 3: 3D hydrothermal plume feature entity recognition.

[0082] After successfully constructing a unified and continuous 3D data volume in step 2, this invention enters a higher-level analysis stage—automatic identification and semantic labeling of 3D feature entities. This step is executed by the system's 3D data volume feature entity identification module. Its core objective is to automatically "discover" and "delineate" 3D objects with clear geological or physical oceanographic significance on the constructed data volume, which includes multiple continuous physical fields such as temperature and salinity, using advanced algorithms. This process achieves a fundamental leap from "data" to "information" and even "knowledge," transforming the complex interpretation work that previously relied on domain experts to complete through multi-view comparison and manual delineation into an objective, efficient, and repeatable automated process.

[0083] In this embodiment, the objective is to accurately identify and extract "hydrothermal plumes." Hydrothermal plumes exhibit significant physicochemical characteristics, typically displaying anomalies relative to the surrounding background seawater, including "high temperature, low salinity, and high concentration of suspended particles (leading to high acoustic scattering)." This invention utilizes these inherent physical laws, transforming them into deterministic identification rules that can be executed by a machine.

[0084] Specifically, the process of executing a three-dimensional spatial pattern recognition algorithm to automatically identify and label three-dimensional hydrothermal plume entities includes the following:

[0085] (1)Set semantic recognition rules.

[0086] The first step of the analysis is to provide the system with a quantification criterion for defining the target feature entity (i.e., hydrothermal plume). The operator can set one or more combined logic conditions based on physical properties through the system configuration interface. In this embodiment, a voxel is determined to be part of the "hydrothermal plume" if it meets all of the following conditions simultaneously.

[0087] Temperature anomaly condition: The value of the body_temperature attribute of this voxel must be higher than a preset threshold of the background temperature obtained from the data volume statistics. For example, body_temperature > T_background + 0.05 (unit: °C).

[0088] Salinity anomaly condition: The value of the body_salinity attribute of this voxel must be lower than a preset threshold of the background salinity. For example, body_salinity < S_background - 0.02 (unit: PSU).

[0089] Acoustic anomaly condition (optional): To improve the accuracy of recognition, the judgment of acoustic scattering intensity can be added. The value of the body_acoustic_intensity attribute of this voxel needs to be higher than a threshold of the surrounding background. For example, body_acoustic_intensity > I_background + 3 (unit: dB).

[0090] These rules transform complex oceanographic knowledge into an exact mathematical expression for performing Boolean logic judgment on each voxel in the data volume.

[0091] (2)Execute the three-dimensional spatial pattern recognition algorithm.

[0092] After setting the recognition rules, the system executes the preset three-dimensional spatial pattern recognition algorithm on the entire three-dimensional data volume. For recognizing entities with spatial continuity such as hydrothermal plumes, the present invention preferably adopts the three-dimensional connected component analysis (3D Connected-component Analysis) algorithm. The execution process of this algorithm is as follows.

[0093] Voxel binarization: The system first traverses each voxel in the three-dimensional data volume and judges it according to the semantic recognition rules set in the previous step. The voxels that meet all conditions are marked as "foreground" (value 1), and those that do not meet any condition are marked as "background" (value 0). In this way, the continuous multi-attribute data volume is temporarily converted into a huge three-dimensional binary matrix.

[0094] Connected Component Search: Next, the algorithm scans the 3D binary matrix. It finds any "foreground" voxel with a value of 1 as a seed point, and then uses breadth-first search or depth-first search to find all "foreground" voxels that are spatially connected to the seed point (e.g., adjacent by faces or edges). All these interconnected voxels constitute an independent "connected component".

[0095] Entity labeling and extraction: The algorithm repeats this process until all voxels with a value of 1 are assigned to a connected component. Each discovered independent connected component physically corresponds to a potential hydrothermal plume. The system assigns a unique ID to each connected component.

[0096] (3) Feature entity parameterization and semantic tagging.

[0097] After separating the set of voxels representing hydrothermal plumes from the background data, the system further performs quantitative parameter calculations and semantic labeling on these extracted 3D entities. For each identified entity (connected component), the system automatically calculates a series of geometric and physical properties, such as the following parameters.

[0098] Geometric parameters: volume (cubic meters), surface area (square meters), geographical extent, centroid location, maximum height, direction of extension, etc.

[0099] Physical parameters: Statistical values ​​of average temperature, maximum temperature, average salinity, minimum salinity, etc. of all voxels inside the entity.

[0100] Finally, the system integrates these calculated parameters, along with a clear semantic label, into a structured metadata record and associates it with the entity's geometric model. For example, in this embodiment, if two independent hydrothermal plumes are identified, the system will generate the following semantic tagging information: { "recognized_entities": [ { "entity_id": "plume_01", "entity_type": "hydrothermal_plume", "volume_m3": 12500, "max_temperature_anomaly_C": 0.21, "centroid_location": {"lon": 121.51, "lat": 24.81, "depth": 1520} }, { "entity_id": "plume_02", "entity_type": "hydrothermal_plume", "volume_m3": 8750, "max_temperature_anomaly_C": 0.15, "centroid_location": {"lon": 121.515, "lat": 24.815, "depth": 1555} } ] } By completing the above steps, this invention successfully elevates the level of analysis from raw, undifferentiated massive data points to three-dimensional geological entities with clear scientific significance and quantitative parameters. This automated and intelligent identification process not only greatly improves the efficiency and objectivity of marine geological interpretation but also provides direct and quantitative decision-making basis for subsequent scientific research and resource assessment.

[0101] Step 4: Interactive visualization and services of model results.

[0102] After completing the reconstruction of the 3D data volume (step 2) and the intelligent recognition of hydrothermal plume feature entities (step 3), this invention enters its final stage. The core task of this stage is to provide the large and complex model results stored on the backend server to the frontend user application through an efficient and flexible service interface, supporting high-performance, smooth, interactive visualization analysis. This process is executed by the dynamic visualization mapping API service module of the system described in this invention, and its working principle can be understood by combining... Figure 4 This step revolutionizes traditional data analytics, eliminating the need for users to download and process massive amounts of raw data files. Instead, users interact directly with a powerful backend data model through a lightweight client.

[0103] Specifically, the process of providing services to the outside world through APIs and supporting interactive visualization by users, which involves the three-dimensional data volume and the identified feature entities, includes the following typical application scenarios in this embodiment.

[0104] (1) Smooth roaming and multi-resolution browsing of massive data volumes.

[0105] In this scenario, a user (such as a marine geologist) wants to get a macro-level overview of the data for the entire exploration area in a three-dimensional virtual Earth environment.

[0106] Backend Support: After the data volume is generated, the system's data volume management and indexing module has pre-indexed the entire 3D data volume using an octree structure. The octree recursively divides the entire data space into eight sub-blocks, forming a multi-level pyramid structure. The top-level node of the tree represents a low-resolution overview of the entire data volume, while the bottom-level leaf nodes correspond to the highest-resolution original voxels.

[0107] Interaction Process: Users zoom or pan within a front-end visual application (e.g., a web-based browser interface). When the front-end application's viewpoint is far from the data volume, it sends a request to the back-end API service module containing the current viewpoint position and desired screen resolution. Upon receiving the request, the API service module uses an octree index to quickly extract aggregated, low-resolution data blocks from higher levels of the tree (e.g., level 3) and returns them to the front-end. Due to the small data volume, the front-end can render quickly, resulting in a smooth user experience. When the user zooms in, bringing the viewpoint closer to the data volume, the front-end sends a new request, and the API extracts and returns higher-resolution data from deeper levels of the octree (e.g., level 7).

[0108] Technical effect: This dynamic data scheduling mechanism based on multi-resolution technology ensures that no matter how large the backend data volume is (TB level), the frontend user can always obtain the best resolution data that matches the current view, thereby realizing stepless scaling and smooth roaming of massive data volumes and solving the performance bottleneck caused by loading a large amount of data at once in traditional visualization technology.

[0109] (2) Real-time slice analysis of arbitrary spatial profiles.

[0110] After discovering areas of interest through macro-level browsing, users may wish to explore the internal structure of the data volume in greater detail.

[0111] Interaction Process: The user activates a virtual tile tool on the front-end interface and defines a plane of arbitrary direction and position (e.g., an XY plane parallel to the sea level, or a vertical XZ plane) by dragging the mouse. Whenever the position of the tile plane changes, the front-end application immediately sends a parameterized tile request to the API service module. For example, requesting a horizontal tile showing the temperature distribution at a water depth of 1550 meters.

[0112] Backend processing: After receiving the request, the API service module does not scan the entire data volume. Instead, it uses spatial indexing to quickly locate all voxel blocks intersecting the slice plane, and then performs real-time calculations only on these blocks to interpolate the temperature value matrix on the precise plane.

[0113] Technical Effects: The calculation results (a two-dimensional temperature data matrix or a directly rendered PNG image) are quickly returned to the front end as a lightweight data package for display. The entire process has a response time in the millisecond range. Users can drag slices in real time and dynamically to observe the morphological and intensity changes of hydrothermal plumes at different depths and directions, achieving a "CT scan"-like exploration of the internal structure of the data volume.

[0114] (3) Focusing analysis and attribute query of three-dimensional feature entities.

[0115] After confirming the location of the hydrothermal plume through slice analysis, the user wanted to perform focused analysis on this specific geological entity.

[0116] Interaction Process: The user interface of the front-end application typically displays a legend listing all the characteristic entities identified in step 3 (such as plume_01, plume_02). When the user clicks on "plume_01" in the legend, the front-end sends an entity query request to the API.

[0117] Backend processing: The API service module retrieves all information about plume_01 from the entity database it manages, based on the ID. This includes the entity's 3D geometric model (usually stored in efficient glTF or VTK format) and all its parameterized attributes (volume, average temperature, etc.).

[0118] Technical Effects: The API returns this information to the front-end. Upon receiving it, the front-end application can highlight the 3D morphology of the hydrothermal plume in a 3D scene, while simultaneously setting the surrounding background data (such as seawater) to semi-transparent or completely hidden, allowing users to clearly observe the plume's precise boundaries, internal structure, and propagation path. Furthermore, its quantitative physicochemical parameters can also be displayed in an information panel.

[0119] Through the aforementioned series of innovative service-oriented and interactive mechanisms, this invention not only solves the problem of visualization performance for massive amounts of data, but more importantly, it provides marine scientists with an unprecedented, powerful, and intuitive set of data analysis tools, enabling them to explore and understand complex marine geological environmental phenomena with unprecedented depth and efficiency.

[0120] Example 2: Construction and interactive analysis of a three-dimensional digital twin of a subsea pipeline corridor.

[0121] Step 1: System configuration and construction of multi-attribute data bodies.

[0122] In this embodiment, the system processes multi-source, standardized data acquired during safety inspections of a section of a subsea natural gas pipeline corridor. The core task of this step is to configure the system according to the specific needs of engineering monitoring and utilize the data reconstruction and fusion capabilities of this invention to construct a unified, multi-attribute, three-dimensional digital twin basic model from various inspection data that aim to reflect different dimensions of pipeline status.

[0123] (1) System configuration and input data.

[0124] This step begins with the system's data volume construction input module. Similar to Example 1, the operator first needs to configure the data volume mapping rules, but the configuration target and the input source data are significantly more engineering-specific. In this scenario, the system receives the following set of typical inspection data files.

[0125] High-precision multibeam bathymetry data: used to accurately obtain the three-dimensional micro-topography of the pipeline body (if exposed) and the surrounding seabed, and is the key basis for determining whether the pipeline is suspended, exposed or eroded and covered.

[0126] Towed magnetometer data: Due to the strong magnetic anomalies of steel pipelines, this data is used for accurate tracking and location of pipelines completely covered by sediment.

[0127] Side-scan sonar image data: used to acquire acoustic images of the seabed on both sides of the pipeline over a wide area, and can efficiently identify topographic textures (such as sand waves), scour trenches, and third-party intrusions that may threaten pipeline safety (such as abandoned fishing nets, shipwrecks, etc.).

[0128] Operators configure the mapping rules interface to map key physical quantities from the aforementioned data sources to attribute fields of the target 3D data volume. A notable feature of this configuration is that, in addition to mapping conventional physical quantities (such as depth and magnetic anomaly intensity), it also maps the 2D pixel values ​​of the side-scan sonar—grayscale values ​​representing the intensity of acoustic backscattering from the seabed—to a voxel attribute called `acoustic_texture`. While this attribute is not a precise physical unit, it is crucial for interpreting seabed sediment types and identifying surface anomalies.

[0129] (2) Reconstruction and fusion of multi-attribute data volumes.

[0130] After configuration, the multidimensional data volume reconstruction and fusion engine begins execution. This process maintains the same methodology as in Example 1, namely, through automated steps such as unified spatial carrier generation, multidimensional data spatialization registration, and multi-field joint spatial interpolation, the selected inspection data is fused into a unified voxel mesh. Since these basic processes have been detailed in Example 1, they will not be repeated here. Instead, the focus will be on illustrating the unique results achieved in this project's application scenario.

[0131] The final output of this step is a highly integrated, multi-attribute 3D data volume. The core value of this data volume lies in the fact that each voxel within it becomes a composite information unit carrying multiple key engineering attributes. Especially in the pipeline body and its adjacent areas, the attribute set of a voxel may simultaneously contain the following information.

[0132] seabed_depth: Elevation value obtained by interpolation of multibeam data.

[0133] magnetic_anomaly: The magnetic anomaly intensity value obtained by interpolating magnetometer data.

[0134] acoustic_texture: The acoustic texture value obtained by registration and interpolation of side-scan sonar image data.

[0135] This deep fusion of multi-attribute voxels is the key technology for constructing a 3D digital twin of pipelines. It fundamentally breaks the traditional inspection and analysis paradigm where various data (topographic maps, magnetic field curves, sonar images) are independent and require repeated manual comparisons. In the data volume constructed by this invention, all information related to pipeline condition assessment is intrinsically and structurally unified within a unified 3D spatial coordinate system. For example, analysts can directly query the complete attribute set of a voxel at a specific location to instantly understand the precise water depth, the presence of magnetic anomalies, and the acoustic characteristics of its surface, without switching between multiple different software programs or maps.

[0136] This unified, multi-attribute three-dimensional data volume forms a high-fidelity, multi-physics digital twin foundation for the subsea pipeline corridor. It not only geometrically replicates the pipeline and its surrounding environment, but more importantly, it contains information from multiple physical fields at every spatial location within it. This provides an unprecedented, highly integrated data model foundation capable of complex spatial calculations, enabling subsequent automated engineering health status diagnosis and immersive interactive analysis.

[0137] Step 2: Interactive engineering analysis of the digital twin model.

[0138] After successfully constructing the multi-attribute 3D data volume of the pipeline corridor in step 1, this invention enters the interactive analysis phase for engineering applications. The core task of this phase is to utilize this high-fidelity digital twin model to provide engineers with a series of powerful, intuitive, and efficient analysis tools to support accurate diagnosis and assessment of the health status of subsea pipelines. This process is primarily driven by the dynamic visualization mapping API service module, combined with simplified feature entity recognition capabilities.

[0139] (1) Accurate extraction of the three-dimensional entity of the pipeline.

[0140] Unlike the exploration of unknown natural entities in Example 1, the core analysis object in this scenario—the subsea pipeline—is a known man-made object with strong physical characteristics. Therefore, the application of the 3D data volume feature entity recognition module is more direct and efficient here. The system can accurately extract the 3D geometric entity of the pipeline itself from the background data volume using any of the following methods: Automatic extraction based on physical features: Since pipelines typically generate the strongest magnetic anomaly signals, the system can quickly identify continuous regions with high magnetic anomaly values ​​representing the pipeline's orientation within the magnetic anomaly data field through simple thresholding and 3D connected component analysis. Similarly, if the pipeline is exposed, its tubular structure can be identified in the terrain data using morphological algorithms.

[0141] Importing design data: In most cases, the as-built design coordinates of pipelines are known. Operators can directly import the pipeline's design centerline coordinate file into the system. Based on this, the system will extract all voxels within a certain radius around the centerline from the 3D data volume, thus constructing a precise 3D solid of the pipeline.

[0142] Regardless of the method used, the result of this step is to obtain an independent three-dimensional geometric model of the pipeline itself, which will be used to guide and constrain all subsequent targeted analyses.

[0143] (2) Interactive engineering analysis based on API services.

[0144] After the pipeline entity is extracted, the dynamic visualization mapping API service module can provide a series of highly customized, engineering application-oriented interactive query services based on the entity model. Engineering technicians do not need complex data processing skills to perform in-depth and quantitative analysis of the pipeline's health status through the front-end application. The following are two typical interactive analysis scenarios.

[0145] Scenario A: Automated diagnosis of pipeline suspension and burial depth status.

[0146] This scenario is designed to quickly assess the overall safety status along the pipeline.

[0147] Interaction process: The user clicks the "Pipeline Health Diagnosis" button on the front-end application interface. The front-end application sends an advanced analytics request to the back-end API service module.

[0148] Backend Processing: Upon receiving the request, the API service module triggers a pre-defined engineering analysis algorithm. This algorithm iterates along the 3D centerline of the pipeline extracted in step 1 with dense step sizes (e.g., every 0.5 meters). At each iteration point, the algorithm performs the following calculations: 1) Obtain the bottom elevation of the pipeline entity at that point; 2) Query the elevation of the seabed surface directly below that point (i.e., the seabed_depth attribute value). By comparing these two elevation values, the status of the pipeline at that point can be determined: if the difference between the two elevations is greater than a pre-defined suspension threshold (e.g., 0.3 meters), the point is marked as "suspended"; if the pipeline elevation is lower than the seabed elevation, its burial depth is calculated.

[0149] Technical Effect: After the algorithm completes execution, the API service module does not return massive amounts of raw data points, but instead directly returns a structured diagnostic report. This report, in JSON format, lists the start and end mileage markers, length, and key parameters such as maximum overhang height or minimum burial depth for all abnormal sections. Upon receiving this report, the front-end application can highlight these abnormal pipe sections on the 3D model using striking colors (e.g., red for overhang, blue for shallow burial), making the overall health status of the pipeline readily apparent and significantly improving the efficiency and intuitiveness of inspection report generation.

[0150] Scenario B: Generation of cross-sections and multi-attribute analysis for arbitrary mileage stations.

[0151] This scenario is used to conduct a detailed local condition assessment of key pipeline sections.

[0152] Interaction process: The engineer enters a pipeline mileage station number that he is interested in into the input box of the front-end application, such as K15+230.

[0153] Backend processing: The frontend converts the mileage marker into geographic coordinates and sends a request to the API service module. This request explicitly requests the generation of a cross-sectional data point with a specific width (e.g., 20 meters on each side) perpendicular to the pipeline's direction at that point. Upon receiving the request, the API service module performs calculations in real-time within the 3D data volume, extracting the multidimensional attributes of all voxels on the 2D cross-section.

[0154] Technical Impact: The API returns multi-attribute data of the cross-section to the front end. The front-end visualization engine can then integrate and render this information onto a single image: a curve represents the precise morphology of the seabed, a circle represents the outline and location of the pipeline, and background color (spectrum) can represent the distribution of magnetic anomalies or acoustic texture characteristics of the soil at that cross-section. With a single image, engineers can intuitively determine whether there is an erosion pit, whether the pipeline is stable, and whether there are changes in the surrounding seabed, among other things. This interactive approach provides unprecedented, highly integrated, and quantitative analytical support for engineering decisions.

[0155] In summary, this invention discloses a system and method for reconstructing and dynamically visualizing multi-dimensional data volumes of marine geological environments. By organically integrating key technology modules such as data volume mapping rule configuration, a multi-dimensional data volume reconstruction and fusion engine, a 3D feature entity recognition module, and a dynamic visualization mapping API service, this invention constructs an intelligent data application framework that transforms discrete observation data into a unified, continuous, and interactive 3D model. This framework fundamentally solves the core technical challenges of existing technologies, such as the structural isolation of standardized data products in deep fusion and interactive analysis, the lack of automated modeling capabilities, and the performance bottleneck of massive data visualization.

[0156] This invention utilizes an innovative multi-dimensional data volume reconstruction and fusion engine to automatically reconstruct discrete observation data with varying physical dimensions and data structures into a unified, continuous, and multi-attribute coupled three-dimensional digital twin model. This lays a solid foundation for collaborative analysis across physical fields and intelligent identification of three-dimensional geological bodies. Crucially, the invention's dynamic visualization mapping API service, through architectural decoupling, on-demand querying, and multi-resolution technology, enables real-time, seamless interaction with terabyte-level massive data volumes, revolutionizing traditional data analysis and visualization methods.

[0157] The flexible technical framework of this invention makes it widely applicable to various scenarios such as marine scientific discovery and marine engineering applications. By transforming dispersed and heterogeneous data assets into a unified model capable of complex spatial computation and providing high-performance data services, this invention not only greatly enhances the depth of marine data utilization and analysis efficiency, but also promotes the development of marine geological research and engineering monitoring towards a more intelligent and insightful direction, demonstrating significant technological advancement and broad application value.

[0158] The above embodiments are only for illustrating the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand and implement the present invention. They should not be construed as limiting the scope of protection of the present invention. All equivalent changes or modifications made in accordance with the spirit and essence of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for reconstructing and dynamically visualizing marine geological environment data, characterized in that, Includes the following steps: It receives one or more standardized multi-source data sources with a unified spatiotemporal reference and establishes data volume mapping rules according to preset configuration; Based on the data volume mapping rules, the discrete observation data in the multi-source data sources are reconstructed, and the reconstructed data is fused into a unified and continuous three-dimensional data volume covering the target area; A three-dimensional spatial pattern recognition algorithm is executed on the three-dimensional data volume to identify and mark three-dimensional feature entities that meet preset attribute conditions; A visualization mapping interface is established for the three-dimensional data volume and three-dimensional feature entities, supporting on-demand querying and multi-resolution data access.

2. The method as described in claim 1, characterized in that, The steps for establishing data volume mapping rules based on preset configuration include: Provides a mapping rule configuration interface for users to configure in a graphical or script-based manner; The mapping rule configuration interface defines the correspondence between physical quantity variables in different input data sources and voxel attribute fields in the three-dimensional data volume. The mapping relationship is then solidified into a machine-readable configuration file.

3. The method as described in claim 1, characterized in that, The discrete observation data in the multi-source data source includes one-dimensional profile points, two-dimensional image pixel centers, and three-dimensional point cloud points.

4. The method as described in claim 1 or 3, characterized in that, The steps for reconstructing discrete observation data from the multi-source data source include: Unified spatial carrier generation: Based on the user-defined geographical range and target resolution, a three-dimensional regular mesh or voxel model is generated in three-dimensional geographic space as a unified spatial carrier for heterogeneous data; Multidimensional data spatial registration: Parse the metadata of each input data source, which includes spatial coordinates and lever values, and project and register the discrete observation data of different dimensions to the world coordinate system of the three-dimensional regular grid or voxel model; Multi-field joint spatial interpolation: Based on the registered discrete observation points, a spatial interpolation algorithm is invoked to calculate one or more physical attribute values ​​for each voxel node in the three-dimensional regular mesh or voxel model, thereby obtaining a three-dimensional data field covering the target area.

5. The method as described in claim 4, characterized in that, The steps for fusing the reconstructed data include: In the three-dimensional data field, the same voxel is used as a multi-attribute voxel unit. At least one physical attribute among temperature, salinity, formation acoustic impedance, and magnetic anomaly intensity, which are obtained by interpolation calculation from different source data, is used as the attribute set of the voxel unit and fused together. The set of all voxel units constitutes a unified three-dimensional data volume.

6. The method as described in claim 1, characterized in that, The steps for performing a 3D spatial pattern recognition algorithm include: In the three-dimensional data volume, a three-dimensional spatial pattern recognition algorithm is used to query and analyze the voxel data to identify continuous three-dimensional spatial regions that meet the preset physical attribute threshold conditions. The continuous three-dimensional spatial region is extracted as an independent three-dimensional entity, and the geometric boundary, volume, and center path parameters of the three-dimensional entity are calculated.

7. The method as described in claim 1, characterized in that, The preset attribute conditions refer to a set of physical attribute thresholds used to filter continuous three-dimensional spatial regions. The set of physical attribute thresholds includes temperature threshold, salinity threshold, acoustic impedance threshold, magnetic anomaly intensity threshold, and combined thresholds.

8. The method as described in claim 1, characterized in that, The steps to establish a visual mapping interface include: A standardized data service interface is constructed to provide the three-dimensional data volume and three-dimensional feature entities to the front-end application in the form of a visual mapping. It supports parameterized on-demand queries, extracting slices or data subsets from the 3D data volume or 3D feature entity according to the query request and returning them; The three-dimensional data volume is organized using a spatial index structure. Based on the viewpoint distance or target resolution parameter in the query request, data of the corresponding level is selected from the spatial index structure for aggregation or sampling and then returned.

9. A multi-dimensional data reconstruction and dynamic visualization mapping system for marine geological environment, characterized in that, include: The data body construction input module is used to receive standardized multi-source data sources and provides a mapping rule configuration interface to establish data body mapping rules; The multidimensional data volume reconstruction and fusion engine is used to reconstruct discrete observation data according to data volume mapping rules, and fuse the reconstructed data into a unified and continuous three-dimensional data volume covering the target area. The 3D data volume feature entity recognition module is used to execute a 3D spatial pattern recognition algorithm on a 3D data volume to identify and mark 3D feature entities that meet preset attribute conditions. The Dynamic Visualization Mapping API service module is used to provide 3D data volumes and 3D feature entities in the form of visual mappings, and supports on-demand querying and multi-resolution data access.

10. The system as described in claim 9, characterized in that, It also includes a data volume management and indexing module, which is used to persistently store and manage the three-dimensional data volume generated by the multi-dimensional data volume reconstruction and fusion engine, and to build and maintain a spatial index structure to support on-demand querying and multi-resolution data access of the dynamic visualization mapping API service module. The spatial index structure selects the data of the corresponding level for aggregation or sampling based on the viewpoint distance or target resolution parameter in the query request and then returns it.

Images