A customized multi-dimensional meteorological digital twin method for major event security scenarios

Abstract

This invention provides a customized multi-dimensional meteorological digital twin method for major event support scenarios, belonging to the field of 3D meteorological information visualization technology. The invention first acquires raw meteorological data within the support area, then uses adaptive format automatic parsing to identify radar volume scan files and 3D real-time grid field files, converting them into intermediate data volumes containing timestamps, feature names, and original feature values. Based on these intermediate data volumes, a mapping transformation is performed, and a texture array is generated using GPU floating-point parallel computing. Real-time modeling and rendering are performed using WebGL 3D texture drawing, combined with real-time pruning, dynamic loading, and progressive rendering to optimize the refresh rate. Multi-scale exploration and quantitative analysis are achieved through volume rendering, layered slicing, profile analysis, and height field settings. This method enables efficient access to multi-source meteorological data and true 3D real-time presentation.

Description

Technical Field

[0001] This invention relates to the field of three-dimensional visualization technology of meteorological information, and in particular to a customized multi-dimensional meteorological digital twin method for supporting major events. Background Technology

[0002] Customized multidimensional meteorological digital twin technology for major event support is rooted in the paradigm shift in meteorological services from traditional forecasting to scenario-based, visualized, and intelligent decision support. The successful hosting of major events relies not only on high-precision spatiotemporal meteorological data but also urgently requires deep integration of data with specific scenarios to form intuitive and interactive three-dimensional visualizations that serve a wide range of users, including non-meteorological decision-makers. However, current operational support platforms are mostly built on two-dimensional visualization technology. While they can convey information through isolines and color maps, their understanding heavily relies on the spatial imagination and experience of professionals, resulting in limited information transmission efficiency and a bottleneck in intuitively presenting complex three-dimensional weather structures and their impact on specific sites.

[0003] To overcome the limitations of 2D visualization, the industry has attempted to introduce various 3D technologies into the meteorological field. One approach is lightweight Web3D technology based on WebGL (such as Three.js). This browser-based technology boasts high adoption and low initial costs, but developing advanced meteorological body rendering, particle systems, and other special effects requires strong graphics programming skills, resulting in long development cycles and difficulty meeting the demands of "short, frequent, and fast" tasks. Furthermore, its architecture is inefficient for directly processing and rendering massive amounts of meteorological grid data, often requiring complex pre-processing of data. Another approach is high-fidelity simulation technology using game engines (such as Unity and Unreal Engine). This approach can achieve photorealistic volumetric clouds, lighting, and physically realistic weather phenomena, offering outstanding visual appeal. However, its core bottleneck lies in the difficulty of integrating with professional meteorological data formats, requiring the development of dedicated middleware for parsing and conversion. Moreover, the resulting systems are typically large-scale deployments with high hardware requirements, hindering rapid business deployment and cross-platform access via the web. In addition, emerging 3D reconstruction technologies such as 3D Gaussian rasterization, while highly realistic in static scene rendering, are essentially models of static visual appearances and cannot dynamically drive and map real-time evolving numerical forecast data streams. Therefore, they cannot independently serve as the core dynamic carrier of meteorological digital twins.

[0004] In summary, general-purpose 3D visualization technology has significant limitations in addressing core requirements such as the fusion of massive geospatial data, direct parsing of professional formats, efficient development, and lightweight deployment needed for major event support. Summary of the Invention

[0005] To overcome the shortcomings of existing technologies, the purpose of this invention is to provide a customized multi-dimensional meteorological digital twin method for major event support scenarios. Through format adaptive parsing, mapping transformation, and GPU parallel texturing rendering link, it achieves efficient unified access and true 3D real-time presentation of multi-source meteorological data.

[0006] To achieve the above objectives, the present invention provides the following solution: A customized multi-dimensional meteorological digital twin method for major event support scenarios includes: Obtain raw meteorological data within the coverage area; The raw meteorological data is subjected to format adaptive automatic parsing technology. Based on the file naming rules and header file format, the corresponding parsing program is matched to identify radar volume scan files and three-dimensional real grid field files, and the parsed data is converted into a unified intermediate data volume. The intermediate data volume contains at least timestamps, feature names and corresponding original feature values. Based on the intermediate data volume, the multi-dimensional features of meteorological information are correlated with key elements of mapping and rendering, completing the mapping and transformation from meteorological data to graphics rendering elements, and using GPU floating-point parallel computing technology to generate texture arrays for WebGL 3D texture drawing. Based on WebGL 3D texture rendering technology, texture arrays are modeled, rendered, and displayed in real time. During the rendering process, real-time pruning, dynamic loading, and progressive rendering techniques with multi-level detail models are employed. The refresh rate of scene rendering is optimized according to the matching graphics rendering quality to control the continuous updating of scene content and achieve progressive rendering. Among them, dynamic loading includes selecting appropriate level of detail data for loading based on data layering, chunking, and dynamic updates of data pages. By integrating interactive functions such as volume rendering, layer slicing, profile analysis, and height field setting, the system performs multi-scale exploration and quantitative analysis on the 3D visualization scene obtained by real-time modeling and rendering of texture arrays, and outputs 3D visualization results.

[0007] Preferably, acquiring raw meteorological data within the protected area includes: The API interface is used to read ground observation data, vertical observation data, radar and satellite data, model forecast products, local meter-level three-dimensional real-time wind products, and service support text products within the coverage area. The collected data is compiled to form raw meteorological data.

[0008] Preferably, the raw meteorological data is subjected to format adaptive automatic parsing technology, including: The corresponding parser is matched based on file naming rules and header file format; The parsing program identifies radar volume scan files and 3D real-world grid field files. The radar volume scan file and the 3D real-world grid field file are parsed using a built-in binary format WebAssembly parser. The parsing results are standardized using a JSON data standardization processor; The standardized data is quality-assured by a multidimensional data quality verification system, resulting in parsed data.

[0009] Preferably, the parser matches the corresponding file naming rules and header file format, including: Perform file header feature recognition to obtain the first target data format identifier; When the format cannot be determined by header feature recognition, metadata analysis and recognition are performed to obtain the second target data format identifier. When the metadata analysis cannot determine the format, naming rule identification is performed to obtain the third target data format identifier; When the format cannot be determined by naming rule identification, a deep analysis of content features is performed to obtain the fourth target data format identifier. The parsing procedure that matches the original meteorological data is determined based on the first target data format identifier, the second target data format identifier, the third target data format identifier, or the fourth target data format identifier.

[0010] Preferably, the parsed data is converted into a unified intermediate data body, including: For site products, spatiotemporal consistency processing, basic quality control filtering, feature extraction, and spatial range interpolation are performed on the parsed data to generate an intermediate data body containing timestamps, feature names, site metadata, feature values, and quality feature codes. For grid products, spatial clipping, feature value filtering, latitude and longitude-based point interpolation, feature extraction and statistics are performed on the parsed data to generate an intermediate data volume containing timestamps, feature names, feature 3D spatial range description information, original feature values ​​and analysis result values. For service assurance text products, keyword retrieval, time matching, and location service matching are performed on the parsed data to generate an intermediate data body containing timestamps, feature names, and corresponding original feature values.

[0011] Preferably, the mapping and transformation from meteorological data to graphic rendering elements includes: Meteorological element datasets are used as input for mapping transformation, where the meteorological element datasets include a set of data elements, a set of element relationships, and a set of element spatial and attribute dimensions. The set of graphic rendering elements is used as the output of the mapping transformation, which includes graphic structural attributes, spatiotemporal relationships of graphic display, and a set of cartographic components. Establish the functional correspondence between inputs and outputs, and complete the mapping transformation based on the functional correspondence.

[0012] Preferably, the texture array for WebGL 3D texture rendering is generated using GPU floating-point parallel computing technology, including: Based on the mapping transformation results, the intermediate data volume is textured and encoded. For example, for 3D cloud data, a gray-white semi-transparent dot texture is applied, and the density of the dots represents the cloud state. For 3D wind data, a streamline particle texture blend is applied to render the texture, representing the wind direction and wind speed, ultimately resulting in a texture array. Set the texture array to an unsigned byte format to enable fast generation of texture arrays.

[0013] Preferably, the real-time modeling, rendering, and display of the texture array based on WebGL 3D texture rendering technology includes: GPU-accelerated algorithms are used to generate and render 2D mesh data into 3D terrain surfaces, isosurface meshes, and particle system geometry in real time. Balance visual effects and rendering performance through adaptive subdivision technology; The Cesium rendering pipeline is used to finalize the real-time modeling and rendering results.

[0014] Preferably, real-time cropping and dynamic loading include: A real-time occlusion clipping algorithm is used to remove invisible polygons early in the graphics pipeline; Dynamic loading of 3D meteorological data is based on data layering, block partitioning, and dynamic updating of data pages; During dynamic loading, select and load detail-level data that matches the view distance.

[0015] Preferably, the interactive functions of integrating volume rendering, layered slicing, profile analysis, and height field setting include: In a 3D scene, the profile of the 3D meteorological data rendering model is redrawn along a specified plane to generate a profile map and render it in real time. Perform profile analysis based on the profile map, and provide profile meteorological data values, timestamps, comparative change analysis, and control over the color transparency of the profile map.

[0016] The present invention discloses the following technical effects: This invention uses format-adaptive automatic parsing technology to identify raw meteorological data within the protected area as radar volume scan files and three-dimensional real-time grid field files, and parses them into a unified intermediate data volume. The intermediate data volume organizes the data with element key-value pairs at fixed time intervals, solving the problems of unstable access and poor reusability caused by the diverse formats of professional meteorological data, scattered parsing chains, and difficulty in unifying data structures in existing solutions from the source, thereby improving the consistency and scalability of data access.

[0017] This invention establishes a correspondence between multi-dimensional features of meteorological information and key elements of mapping and rendering based on intermediate data, completes the mapping and conversion of meteorological data to graphic rendering elements, and uses GPU floating-point parallel computing to generate texture arrays for WebGL 3D texture drawing. This allows meteorological numerical data to directly enter the visualization pipeline with a rendering-oriented data structure, reducing the latency and engineering coupling caused by a large amount of offline preprocessing and format conversion in the traditional process, thereby improving end-to-end processing efficiency.

[0018] This invention uses WebGL 3D texture rendering to model and render texture arrays in real time, forming an interactive 3D visualization scene. This allows the 3D weather structure and its continuous spatial changes to be presented intuitively, avoiding the problems of high spatial understanding costs and insufficient information expression caused by relying solely on 2D products or static layers. This enhances the intuitiveness and interpretability of situational awareness in the scenario of ensuring the success of major events.

[0019] This invention employs real-time pruning, dynamic loading, and progressive rendering techniques with multi-level detail models during the rendering process, and optimizes the scene rendering refresh rate based on the graphics rendering quality. Among them, dynamic loading is based on data layering, block division, and dynamic updating of data pages, selecting detail level data that matches the viewing distance for loading. This allows the system to maintain a controllable rendering load when the data scale increases and the viewing angle changes, reducing the stuttering and resource waste caused by full loading and full-detail rendering, thereby improving the real-time performance and stability on the browser side.

[0020] This invention utilizes interactive functions such as volume rendering, layered slicing, cross-sectional analysis, and height field setting to conduct multi-scale exploration and quantitative analysis of the three-dimensional visualization scene and output three-dimensional visualization results. This allows users to not only "see" the three-dimensional structure but also obtain quantitative information in the cross-sectional, layered, and height dimensions during the interactive process. This overcomes the shortcomings of simply providing a display without analytical tools, thereby improving the pertinence and efficiency of safeguard decision-making. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 A flowchart of the method provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the end-to-end 3D meteorological digital twin system architecture from data source to front-end rendering provided in an embodiment of the present invention; Figure 3 A flowchart illustrating the high-performance 3D visualization and interaction technology provided in this embodiment of the invention; Figure 4 This is a schematic diagram of the layered architecture of the meteorological activity support platform provided in an embodiment of the present invention; Figure 5 This is a schematic diagram illustrating the process of real-time access, parsing, preprocessing, and three-dimensional visualization interactive analysis of multi-source meteorological data for major event support scenarios provided in this embodiment of the invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] The purpose of this invention is to provide a customized multi-dimensional meteorological digital twin method for major event support scenarios. Through dynamic loading, real-time cropping and progressive rendering, it can achieve high-performance visualization of massive spatiotemporal data and support multi-scale interactive quantitative analysis output.

[0025] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0026] Figure 1 The method flowchart provided in the embodiments of the present invention is as follows: Figure 1 As shown, this invention provides a customized multi-dimensional meteorological digital twin method for major event support scenarios, including: Step 100: Obtain raw meteorological data within the coverage area; Step 200: Apply format adaptive automatic parsing technology to the raw meteorological data, match the corresponding parsing program based on file naming rules and header file format to identify radar volume scan files and three-dimensional real grid field files, and convert the parsed data into a unified intermediate data volume; the intermediate data volume shall at least contain timestamps, feature names and corresponding original feature values; Step 300: Based on the intermediate data volume, establish a correspondence between the multi-dimensional features of meteorological information and the key elements of mapping and rendering, complete the mapping and transformation from meteorological data to graphics rendering elements, and use GPU floating-point parallel computing technology to generate a texture array for WebGL 3D texture drawing. Step 400: Based on WebGL 3D texture rendering technology, the texture array is modeled, rendered, and displayed in real time. During the rendering process, real-time pruning, dynamic loading, and progressive rendering techniques with multiple levels of detail are adopted. The refresh rate of scene rendering is optimized according to the matching graphics rendering quality to control the continuous updating of scene content and achieve progressive rendering. Among them, dynamic loading includes selecting appropriate level of detail data for loading based on data layering, chunking, and dynamic updating of data pages. Step 500: Through the interactive functions of integrated volume rendering, layer slicing, profile analysis and height field setting, the 3D visualization scene obtained by real-time modeling and rendering of the texture array is explored and quantitatively analyzed at multiple scales and the 3D visualization results are output.

[0027] In one feasible embodiment, this customized multidimensional meteorological digital twin technology is a true 3D visualization integrated technology solution based on WebGIS. The system is based on the Cesium 3D WebGIS framework and adopts a B / S architecture to achieve front-end and back-end separation. Clients do not need to install plugins; only a browser supporting WebGL2 is required to access the 3D visualization platform. The system is geared towards supporting major events, unifying professional meteorological data products with geospatial data elements to form a collection of visualized objects that can be overlaid and displayed in the same 3D scene. It supports interactive functions such as overlay analysis, filtering, charting, playback, downloading, and searching to quickly capture key information on severe weather and support on-site command and coordinated decision-making.

[0028] like Figure 2 As shown, the key link from data to rendering in this invention can be abstracted into the input of visualized meteorological data products and the output of visualized interactive output data, with three stages set between the two: structural mapping, spatiotemporal dimension mapping, and element mapping. Figure 2 The left side represents the storage and supply end of visualized meteorological data products, which include radar volume scan files and three-dimensional real-time grid field files, etc. Figure 2 The right side shows the visualized interactive output data, which is used to drive the rendering, display, and interactive analysis of the 3D scene. Figure 2The three-stage mapping in the middle is used to complete the data structure unification, spatiotemporal dimension alignment and rendering element construction, thereby forming a closed-loop link from data source to visual interactive results.

[0029] At the data request layer, the system employs format-adaptive automatic parsing technology to analyze and preprocess raw meteorological data. Specifically, the system uses file naming rules and header file format matching to match corresponding parsing programs to identify radar volume scan files and 3D real-time gridded field files. On the client side, a built-in binary format WebAssembly parser, a JSON data standardization processor, and a multi-dimensional data quality verification system are used to complete efficient parsing and quality assurance. The parsing results are organized into feature key-value pairs at fixed time intervals, where the key is the meteorological feature name and the value is the spatial gridded field corresponding to the meteorological feature name. These feature key-value pairs are stored and scheduled as intermediate data at the front end. The above process corresponds to... Figure 2 The structure mapping stage is used to convert multi-source heterogeneous meteorological data into a unified data organization form and provide a stable data entry point for the subsequent generation of rendering elements.

[0030] In the spatiotemporal mapping stage, the system maps intermediate front-end data onto a unified spatial reference and time axis, completing spatiotemporal alignment and index organization for rendering. Taking a major event support scenario as an example, the system overlays two-dimensional and three-dimensional geographic elements based on CGCS2000 geographic coordinates, overlaying digital models of support venues and various meteorological elements such as radar echoes, temperature, and wind in the same scene, and supports data playback and process review through temporal sequence organization. The goal of this stage is to explicitly make the temporal and spatial continuity of meteorological data explicit, so that subsequent rendering only needs to extract the corresponding rendering input according to the current viewpoint and current time, avoiding repeated calculations across formats and coordinates, thereby improving real-time performance.

[0031] In the element mapping stage, the system utilizes GPU floating-point parallel computing technology to rapidly generate texture arrays for WebGL 3D texture rendering, and further completes the rendering representation of multi-dimensional meteorological elements. The mapping transformation from meteorological data to graphics rendering elements can be represented as a functional correspondence with the meteorological element data set as the domain and the graphics rendering element set as the value domain. The rendering model is represented as a functional relationship M, satisfying that M maps the element data in the domain to the set of renderable elements in the value domain. The domain includes the set of data elements, the set of element relationships, and the set of element space and attribute dimensions; the value domain includes graphic structure attributes, spatiotemporal relationships of graphic display, and the set of cartographic components. Scalar field elements such as temperature and air pressure can be multi-dimensionally expressed using color mapping and altitude mapping; vector field elements such as wind fields and ocean currents can express motion characteristics using dynamic particles and streamlines. The color band system, as part of the cartographic component set, can provide a variety of professional color bands that conform to meteorological industry practices and supports dynamic color band generation and user configuration to ensure consistent readability of different elements within the same scene.

[0032] After generating the texture array, the system utilizes WebGL 3D texture rendering technology to achieve rapid processing of multi-layer data and products, as well as real-time modeling, rendering, and display. The rendering engine employs GPU-accelerated algorithms to generate 3D terrain surfaces, isosurface meshes, and particle system geometry from 2D mesh data in real time, achieving a balance between visual effects and rendering performance through adaptive subdivision. The generated visualization data is ultimately presented through the Cesium rendering pipeline, which includes efficient management of GPU texture memory, dynamic compilation of shader programs, and real-time rendering of 3D scenes, thus achieving end-to-end GPU acceleration from data to pixels.

[0033] Furthermore, to adapt to the real-time presentation requirements of massive meteorological data in web services, the system employs real-time cropping, dynamic loading, and progressive rendering techniques with multi-level detail models during the rendering process, while optimizing the refresh rate of scene rendering to ensure visualization effects. Dynamic loading is based on layered and segmented data and data page updates, selecting and loading data of matching detail levels according to the viewing distance; real-time cropping reduces the rendering load of invisible data; progressive rendering maintains screen stability and smooth interaction under continuous data updates. Through the above process, this invention can achieve high-performance real-time rendering of 3D meteorological elements on the browser side, forming visualized interactive output data that can be used for interactive analysis and command decision-making. The overall process and module division are as follows: Figure 2 As shown.

[0034] In one feasible embodiment, the present invention provides an interactive analysis and visualization technology integrating two-dimensional and three-dimensional meteorological data. The system provides high-precision terrain and imagery services at multiple scales based on Cesium's functional API. These multiple scales include at least large-scale, medium-scale, small-scale, and micro-scale, thus supporting continuous zooming and detailed browsing from global to venue-level perspectives. It also supports dynamic data display based on a time axis, enabling the loading, playback, and replay of meteorological element data according to temporal context. Simultaneously, the system provides temporal visualization capabilities across multiple scenarios, including at least three-dimensional, 2.5-dimensional, and two-dimensional scenes. This ensures that the same meteorological element maintains temporal and spatial consistency across different scenarios, facilitating rapid switching between macro-level situational assessment and detailed local analysis by operational personnel, thereby improving the efficiency of comprehensive assessment in scenarios supporting major events.

[0035] Building upon the aforementioned multi-scale and multi-scene capabilities, this invention utilizes the HXGIS 3D Engine to construct a custom rendering engine and integrates a sectioning algorithm, forming an interactive analysis toolset for 3D meteorological products. By integrating core interactive functions such as volume rendering, layered slicing, profile analysis, and height field settings, the system enables users to perform multi-scale exploration and quantitative analysis of 3D meteorological element data, from overall structure to local details. Taking the profile analysis of radar 3D reflectivity data as an example, the system can sample radar 3D reflectivity data using custom shaders and texture mapping. It receives a specified position line input by the user as a sectioning constraint, extracts sampled data at different heights along the specified position line, and constructs a vertical sectioning plane using a nearest-neighbor sampling interpolation algorithm. Furthermore, it calculates the intersection of the sectioning plane and the radar 3D reflectivity data to form profile data, renders the profile map in real time, and provides a profile display function with legends. It also supports arbitrary direction sectioning and transparency control, thereby achieving refined analysis and intuitive presentation of the vertical structure of radar data without departing from the 3D scene spatial reference.

[0036] As an example, the method for parameterizing the vertical cutting plane in this embodiment is as follows: For any three-dimensional meteorological element field, its space is a three-dimensional vertical coordinate system. If two arbitrary points (x1, y1) and (x2, y2) are selected on a given horizontal plane (xy-axis plane), and a straight line l is drawn, then all points on the cutting plane bounded by l satisfy the following condition: their x and y coordinates lie on the straight line l, and their z coordinates can take any real number. Therefore, the set of points on this plane is: Where A = y1 - y2, B = x2 - x1, and C = x1y2 - x2y1.

[0037] Right now , .

[0038] Furthermore, the method for parameterizing the horizontal cutting plane is as follows: 1) On a given horizontal plane (xy-axis plane), arbitrarily select three or four points P1(x1, y1), P2(x2, y2), P3(x3, y3), and P4(x4, y4). Connect the vertices in order to obtain the sides. Each side corresponds to a linear inequality such that the points inside the polygon satisfy all inequalities (pointing inward). When the user selects three points that are not on a straight line to form a triangle, the parameterized form is: Among them u, v≥0, and u+v≤1; Let the coefficient matrix be... If the triangle is non-degenerate, The matrix is ​​invertible; this can be achieved by solving a system of linear equations. get, Then import the specified height layer You can obtain a cross-section of a specified area at a specified height: If a user selects four non-collinear points to form a convex quadrilateral, and connects the vertices in sequence to obtain the sides, each side corresponds to an inequality such that all points within the polygon satisfy these inequalities. Let the equation of the line on the i-th side be... And the interior points of the polygon make The point sets of each layer are then: The selected quadrilateral is represented by a bilinear parameter as follows: z=z0; If we remember: , , , , but: Treating the equation as a linear equation of t, and simplifying it, we get: (x-x1-s ()( +s )=(y-y1-s ()( +s ) Rearranging this into a quadratic equation of length s: as² + bs + c = 0, where: Solve the equation and choose s The root of [0, 1] can be used to obtain s, and: If the user selects the selection box function, the two points selected by the user are used as the diagonals of the selection rectangle, and in the above formula, Wx=Wy=0: Furthermore, the nearest neighbor sampling interpolation algorithm in this embodiment is a high-precision interpolation algorithm based on Gaussian process regression, used to reconstruct a complete, continuous function from a finite set of noisy data points. A Gaussian process consists of a mean function m(x) and a kernel function. Definition, writing: The kernel function is defined as the function at any two points x and xy. The covariance between these parameters controls the smoothness, periodicity, and other properties of the function. In practical 3D rendering, the data set D within the nearest boundary is represented as... Given an interpolation input point x (the location to be rendered), the joint distribution of the original data y and the interpolation result f can be written as: Where K(X, X) is an n×n matrix with elements (i, j) being k(xi, xj); K(X, x) is a vector of length n with elements i being k(xi, x). It is a scalar.

[0039] The mathematical core of this algorithm lies in the properties of the multivariate Gaussian distribution and the application of the kernel function. It encodes prior knowledge about the behavior of the function through the covariance matrix and obtains the posterior predicted distribution through Bayesian inference after observing the data.

[0040] In one feasible embodiment, to ensure the professional consistency and readability of the 3D meteorological profile rendering, this invention adopts a standardized color gamut and legend mapping mechanism for 3D meteorological graphic products. The basic colors of the color gamut consist of red, orange, yellow, magenta, green, cyan, blue, and purple, with the RGB values ​​of the main hues referring to the industry standard "Color Gamut of Meteorological Service Graphic Products QX / T180-2013". When color mapping meteorological element values, a gradient algorithm is used to smoothly transition between color levels to eliminate artifacts caused by color level boundaries and improve profile continuity. The system has built-in mapping parameters for common meteorological elements and uses different mapping algorithms for different elements to meet the numerical range and display resolution requirements of different physical quantities.

[0041] In the above mapping mechanism, the physical range of radar reflectivity is -20dBZ to 80dBZ, using a linear mapping formula: stored value = (physical value + 20) × 2.55. This formula maps -20dBZ to 0 and 80dBZ to 255, with intermediate values ​​linearly distributed. Temperature data is mapped from -50℃ to 50℃, using a similar linear mapping formula: stored value = (physical value + 50) × 2.55, to ensure the continuity and accuracy of temperature data. Wind speed data is mapped from 0m / s to 60m / s, using the formula: stored value = physical value × 4.25; for extreme wind speeds exceeding 60m / s, a uniform mapping of 255 is applied. Precipitation data is mapped from 0mm to 200mm, using the formula: stored value = physical value × 1.275. Relative humidity is mapped from 0% to 100%, using the formula: stored value = physical value × 2.55.

[0042] To ensure the controllable presentation of missing data in profile rendering, all missing values ​​are uniformly mapped to 0 and treated as transparent pixels during front-end rendering. The system predefines a missing value identifier for each element, with a missing value of -99dBZ for radar reflectivity and -999.9℃ for temperature. Through the synergy of the aforementioned color gamut standard, mapping range, mapping formula, missing data processing, and transparent pixel strategy, this invention achieves consistency in legends, continuity of color bands, and controllable visualization of missing data in profile rendering.

[0043] In one feasible embodiment, this invention constructs a multi-element real-time forecast integrated service on a three-dimensional geographic information platform. Using a self-developed data engine, data from encrypted automatic weather stations spaced 100 meters apart in the core activity area are mapped onto a three-dimensional terrain model using an interpolation algorithm. This generates three-dimensional volumetric maps of temperature, relative humidity, specific humidity, wind speed, and radar echo intensity, along with corresponding horizontal and vertical cross-sectional maps, thus visually displaying microclimate characteristics. Simultaneously, this invention provides a time-series map query function integrating real-time data and forecasts for each station, enabling comparative analysis of real-time data and forecasts for the same station on the same time axis.

[0044] Furthermore, this invention supports access to forecast field data such as intelligent networks and numerical models, and displays key element information in a three-dimensional scene; and provides the ability to identify key forecast weather processes that affect activities in response to event support needs, thereby providing refined forecast product services and key element curve services based on event locations, realizing element focus, process identification and location services for support tasks.

[0045] In one feasible embodiment, the present invention sets up a special support module for major events, targeting specific support areas to provide customized support capabilities for specific major events. The customized support capabilities include the integration and docking of spatial information such as the actual 3D model of the venue, route planning, and key locations under specific event scenarios, and the linkage of the above spatial information with meteorological data to achieve integrated display and decision support for event scenarios.

[0046] In one feasible embodiment, the present invention provides a key meteorological element curve module. This module connects to multi-source 2D and 3D cloud meteorological products through a standardized interface, uses front-end rendering technology to achieve a high-fidelity 3D display of the 3D cloud structure, and provides interactive analysis tools. Furthermore, based on a self-developed multi-element radiosonde curve drawing tool and a real-time forecast integrated drawing tool, this module provides a pop-up key meteorological element curve service for any venue or geographical location. The key meteorological element curves include curves showing changes in multiple elements such as real-time and forecast temperature, precipitation probability, and wind speed; and synchronizes the background lighting of the curve service with the weather evolution in the 3D scene to achieve a combined display of macro-weather evolution and micro-viewpoint curves.

[0047] In one feasible embodiment, this invention develops a high-precision 3D cloud, 3D wind field, and 3D radar product parsing and data transmission module to realize the data processing and packaged transmission of 3D atmospheric real-time cloud, 3D wind field, and 3D radar products. This module captures the latest multi-source 3D real-time products through a standardized interface and employs technologies such as real-time cropping, dynamic loading, and multi-level detail models to achieve data mapping, cropping, and transmission of visual regions; by progressive rendering, it reduces the volume of data transmitted each time, thereby improving the efficiency of dynamic interaction.

[0048] In terms of radar data coordinate processing, the original coordinate system of radar data is a polar coordinate system, with the radar position as the origin, defined by three parameters: azimuth, elevation, and slant range. This invention converts the polar coordinate data into a Cartesian coordinate system. First, the coordinates of each data point in the radar's local coordinate system are calculated, and then the coordinates are transformed to the WGS84 geographic coordinate system to form a spatial reference consistent with the three-dimensional geographic information base.

[0049] In terms of grid organization, this invention employs a regular three-dimensional grid structure. The horizontal direction uses a grid with equal latitude and longitude, and the resolution is determined based on the data coverage area, typically ranging from 500 meters to 1 kilometer. The vertical direction uses a grid with equal spacing from the ground, typically at 250-meter or 500-meter intervals. Each grid point stores parameters such as reflectivity and velocity, and supports simultaneous storage of multiple elements to meet the integrated organization, transmission, and front-end rendering needs of multi-source products such as 3D cloud, 3D wind field, and 3D radar.

[0050] In one feasible embodiment, this invention targets the scenario of ensuring the smooth operation of major events. It constructs an end-to-end execution chain around multi-source data access, 3D visualization analysis, early warning product generation, and visualization services, enabling the efficient flow and precise application of meteorological information during these events. The system takes multi-source meteorological data and 3D geographic baseline data as input, generates intermediate data volumes for rendering through a unified real-time access and parsing preprocessing mechanism, and completes GPU-accelerated rendering, integrated visualization of real-time conditions and forecasts, and interactive analysis output on the 3D scene side.

[0051] Figure 5 The upper part illustrates the relationship between importing 3D geographic base data and carrying 3D scenes. For example... Figure 5 As shown above, the 3D geographic foundation data is organized based on a multi-dimensional real-time assessment model, including at least Tianditu tile data, event support model data, and other administrative vector data. This 3D geographic foundation data is imported into the 3D scene via a data import channel to construct the spatial reference and scene representation of the support area. Based on this spatial representation, the subsequent 3D overlay of meteorological elements can achieve consistent spatial matching and visual presentation with spatial objects such as building digital models, image base maps, and administrative divisions within the support area, while GPU-accelerated rendering on the scene side ensures smooth interaction.

[0052] Figure 5 The lower part illustrates the process chain for real-time access, parsing, processing, rendering, and interactive analysis of meteorological data. For example... Figure 5 As shown in the lower part, the meteorological data comes from the Tianqing system and local observation system, including at least local characteristic observation data, ground observation data, radar and satellite observation data, model forecast data and microscale wind field data. The above data enters the processing link through the interface-based real-time data reading module, and sequentially performs data parsing and element processing, generates binary streams and 3D texture rendering, and finally enters the 3D scene side to form output capabilities such as 3D visualization of real-time data, 3D visualization of forecast data and 3D interactive analysis.

[0053] Furthermore, in the real-time access and preprocessing steps of multi-source meteorological data, the system reads ground observation data, vertical observation data, radar and satellite data, model forecast products, local meter-level 3D real-time wind products, and service support text products within the coverage area through API interfaces, and forms different intermediate data volumes for different types of data. For station products, spatiotemporal consistency processing, basic quality control filtering, feature extraction, and spatial range interpolation are performed to generate an intermediate data volume containing timestamps, feature names, station metadata, feature values, and quality feature codes. The station metadata includes at least station number, station name, administrative division, latitude and longitude, and altitude. For grid products, spatial clipping, feature value filtering, latitude and longitude-based point interpolation, feature extraction, and statistics are performed to generate an intermediate data volume containing timestamps, feature names, feature 3D spatial range description information, original feature values, and analysis result values. For text service products, keyword retrieval, time matching, and location service matching are performed. The processed data is used for service support-related visualization service calls and displays.

[0054] To ensure the accuracy and efficiency of identifying and parsing multi-source heterogeneous files, this embodiment further employs an intelligent format recognition system, which is a four-level progressive strategy. The first level is file header feature recognition, reading the first 64 bytes of the file and comparing them with predefined feature patterns. SA radar base data begins with RADD, NetCDF files begin with CDF plus control characters, and GRIB files begin with GRIB. The first level offers the fastest recognition speed and highest accuracy. The second level is metadata analysis and recognition, reading metadata information for self-describing formats and analyzing global and variable attributes to determine the data format and content structure. The third level is naming rule recognition, maintaining a naming rule library and performing regular expression pattern matching on business filenames. SWAN radar product filenames contain the Z_RADR_I prefix and carry fields such as time information, product type, and radar station number, enabling rapid identification. The fourth level is in-depth content feature analysis, analyzing byte distribution and data structure features when the format cannot be determined by the first three levels. Statistical analysis and pattern recognition are used to infer the most likely format type for handling non-standard format data.

[0055] In terms of intermediate data structure design, the parsed data is uniformly converted into an intermediate data structure, which includes at least a basic information layer, a spatial reference layer, a data body layer, and a quality control layer. The basic information layer records the data source, processing time, and data quality; the spatial reference layer uniformly adopts the WGS84 coordinate system and records the latitude and longitude range and grid parameters; the data body layer selects the organization method according to the data type; and the quality control layer records completeness and accuracy information. For W3DM binary format, the file header includes the magic number w3dm, version number, and total file length; the Feature Table JSON part adopts the standard JSON format and includes spatial range, physical dimensions, and texture parameters; the data body part is organized according to a 3D texture format and supports pixel formats such as RGBA. For JSON format data, taking grid data as an example, the JSON object contains grid parameters such as xNum, yNum, startX, and startY, and stores data values ​​through a grid array to balance data integrity and front-end parsing and display efficiency.

[0056] In the spatial model data processing step, the system improves the spatial matching and lighting design of building digital model data, Tianditu tile data, administrative vector data, and manually modeled data within the protected area, providing a unified spatial basis and visual consistency for meteorological element overlay. In the two-dimensional product production step, the system generates two-dimensional time series maps, heat maps, and profile maps: for a single point in three-dimensional space, it extracts and decodes the real-time meteorological element files and the latest time forecast files corresponding to that latitude and longitude, creating a time series map integrating real-time and forecast data; it decodes meteorological elements at multiple points and creates heat maps; through scientific calculation, it obtains the variation values ​​of multiple elements at a single point with height, and draws vertical profile map products to form a multi-dimensional auxiliary analysis capability for points and profiles.

[0057] In the dynamic cropping and loading steps of 3D meteorological data, the system employs a real-time occlusion cropping algorithm to remove invisible polygons early in the graphics pipeline, avoiding unnecessary processing of invisible parts. Based on data layering, segmentation, and dynamic updates of data pages, it achieves multi-layered, large-scale dynamic loading of 3D scenes, selecting appropriate levels of detail for loading. Distant objects are loaded with lower levels of detail, while nearby objects are loaded with higher levels of detail, thus improving dynamic interaction efficiency while maintaining visualization effects. Figure 5 The three-dimensional texture rendering and three-dimensional scene GPU-accelerated rendering output shown in the lower part demonstrate that this embodiment realizes a closed-loop execution process from real-time access and analysis of multi-source meteorological data to three-dimensional visualization and interactive analysis.

[0058] As an example, the real-time occlusion cropping algorithm in this embodiment is as follows: Suppose there is an object O that blocks the view, and the light source point of the camera is E. An extended plane formed by E and the contour edge of object O creates the occlusion. There exists a plane π such that point P lies on the opposite side of this plane, and the line connecting E and P passes through O. Therefore, point P is occluded. Let the equation of plane π be... ,in It is the unit normal vector pointing inwards from the plane. Let P be any point in space, and C be a constant. If an object lies in the positive half of space (inside) relative to this plane, then the object may be visible. For point P, if... Points located inside the view frustum are considered visible. The view frustum consists of six planes (left, right, top, bottom, near, and far). Only points that pass the tests on all six planes and are not occluded are rendered. Before performing complex occlusion determination, a level-order traversal of the octree combined with view frustum testing is performed to quickly eliminate nodes that are not within the camera's field of view. Then, a node hierarchy is constructed using a depth pyramid, and the occlusion state of the parent node is used to infer the state of the child nodes, thus avoiding testing and rendering a large number of occluded details.

[0059] Furthermore, the data layering algorithm in this embodiment is as follows: First, the selected major occlusions (such as buildings and terrain) in the scene are rendered to the Z-buffer. Then, a chain of mipmaps, the depth pyramid H, is generated from this full-resolution Z-buffer. Each pixel value in each layer Lk of the pyramid corresponds to the maximum depth value (i.e., the depth furthest from the camera) of a 2×2 pixel block in the layer above it, Lk-1. Mathematically, for the k-th layer: Furthermore, the hierarchical loading and rendering algorithm in this embodiment is as follows: The data is traversed from front to back and divided into an octree. For the current node, view frustum culling is performed first. If the node is outside the view frustum, the subtree is skipped. If the node is inside the view frustum, occlusion testing is performed. The node's bounding box is compared with the depth pyramid using the above formula. If the node is determined to be occluded, all its child nodes are skipped. If the node is potentially visible, it is analyzed that the node is a leaf node, and its contained geometry is submitted for rendering. If the node is not a leaf node, its child nodes are traversed again. Finally, the data points that pass the test are rendered according to the visual depth, and the complete image is synthesized. This hierarchical processing method utilizes spatial coherence and inter-frame coherence to reduce the time complexity of visibility determination from O(n) to O(logn), greatly improving the rendering efficiency of large-scale 3D scenes.

[0060] This invention employs a high-performance progressive texture rendering mechanism. By combining Level of Detail (LOD) technology with multi-resolution tiles and a smooth transition strategy, it adaptively selects and displays matching level-of-detail data under different viewing distances and angles. Furthermore, it optimizes the scene rendering refresh rate based on the graphics rendering quality to control continuous updates of scene content and achieve progressive rendering. Simultaneously, by integrating core interactive functions such as volume rendering, layered slicing, profile analysis, and height field settings, it enables users to conduct multi-scale exploration and quantitative analysis of 3D meteorological element data, from overall structure to local details. This provides more targeted situational assessment and precise decision-making support for meteorological support during major events.

[0061] In one feasible embodiment, this invention addresses the challenges of providing support for major events, specifically targeting the diverse sources of professional meteorological data, complex spatial data types, and long and easily coupled 3D rendering chains. It unifies and collaboratively presents professional meteorological data, 2D and 3D GIS engines, BIM information models, and geospatial information. Leveraging a centralized geospatial database, topographic data, and 3D model resources, it constructs a high-performance 3D visualization and interactive system based on a Web technology stack. This system adopts a front-end / back-end separation design, using standardized interfaces to connect data acquisition, real-time processing, and front-end rendering. While ensuring decoupling between data and display, it optimizes the entire performance chain from data source to front-end rendering, thereby providing smooth, stable, and scalable visualization and interactive capabilities for 3D command and coordinated decision-making in meteorological event support.

[0062] like Figure 3 As shown, the system's top layer is the visualization and interactive page. On the page side, the Cesium 3D engine handles 3D scene organization and rendering scheduling, the VUE 3.0 framework provides interface construction capabilities, and Pinia manages state and maintains the consistency of the interactive state. It also includes a custom WebGL layer management module, a WebGL parallel computing module, a shader program manager, and a dataset time-series update module to support unified visualization of multi-source meteorological elements and various spatial objects. The page accesses data through a standard API service, which adopts a RESTful API format and provides two types of capabilities to the front end: a data time-series table to describe the time series and update rhythm of available data, and a data file interface to obtain data files for specific elements or spatial ranges on demand, thus providing a stable data supply channel for front-end rendering and interaction.

[0063] exist Figure 3In the illustrated chain, the standard API interface service connects to the meteorological data processing and 3D construction module. This module caters to different data formats and includes functional units such as real-time data reading, parsing processor, data compression, 3D tiling, and dynamic loader. Real-time data reading continuously acquires raw data from different data sources; the parsing processor parses data of different formats and forms a unified intermediate structure; data compression reduces memory pressure on the transmission and rendering sides; 3D tiling organizes data for 3D rendering into loadable 3D tiles; and the dynamic loader loads and updates tile data on demand based on viewport, viewing distance, and interaction status. Through this processing chain, the system transforms complex and heterogeneous data sources into resource objects that can be directly used for WebGL rendering on the front end, and collaborates with the dataset time-series update module on the page side to form a stable real-time update and progressive loading mechanism.

[0064] Figure 3 The system further outlines a hierarchical organization of data sources to support the integrated presentation of meteorological elements, topography, and 3D objects. Firstly, professional meteorological data sources include station data and grid data, accessible via data lakes and FTP. Secondly, topography data sources include DEM, DOM, and vector data, used to construct the basic geographic foundation and spatial reference. Thirdly, 3D model data sources include oblique photogrammetry, handcrafted models, and architectural models, supporting OSG, OBJ, GLTF, GLB, and BIM model formats for refined representation of key scenes, important areas, and facilities. By integrating the above data into a unified standard API interface service and dynamic loading chain, the system can collaboratively display meteorological element overlay, topography, and BIM-level object presentation within the same 3D scene.

[0065] Figure 4 This is a schematic diagram of the layered architecture of the meteorological activity support platform of the present invention. Figure 4 As shown, the platform is divided into a basic support layer, a data layer, a service layer, a business layer, and a presentation layer from bottom to top: The basic support layer includes hardware resources, network resources, and storage resources to ensure computing, transmission, and storage capabilities; the data layer includes capabilities such as data acquisition, real-time access, real-time processing, and real-time querying, and sets up a real-time data lake, Tianqing data interface, high-precision spatial data, and thematic libraries to form a spatiotemporal data foundation for meteorological support; the service layer uses a digital twin rendering engine as its core to provide meteorological visualization services, spatiotemporal analysis services, intelligent early warning services, and data analysis services, decoupling upper-layer business from specific implementations; the business layer consists of business modules such as data viewing and updating, dynamic display and interaction, and data visualization and analysis; the presentation layer uses the national-level meteorological integrated 3D command platform as a unified entry point to achieve multi-business collaboration and unified presentation. Figure 4Both sides also provide a horizontal protection system, in which a security management system runs through all layers to ensure access control, operational security and audit requirements; and a standards and specifications system runs through all layers to constrain interface specifications, data specifications and operation and maintenance specifications, thereby improving the platform's maintainability and scalability.

[0066] Building upon the aforementioned embodiments, the platform can adopt a cloud-native microservice design to achieve front-end and back-end separation, service decoupling, and data and presentation separation. On the one hand, standardized data and service interfaces enable rapid access to multiple data sources and service orchestration; on the other hand, a multi-layered service governance and monitoring system enhances high availability and maintainability. Simultaneously, GPU-accelerated rendering and distributed computing capabilities ensure high performance for 3D visualization and meteorological data processing. Thus, the platform not only meets the real-time and stability requirements for major event support but also provides a unified technical foundation for future expansion to more support scenarios, more element types, and more business services through standardization and layered decoupling.

[0067] As an optional implementation method, meteorological support for large-scale sporting events is a sophisticated service that spans the entire process from pre-event preparation, in-event execution, and post-event review. Taking the 2025 National Games as an example, the opening and closing ceremonies, outdoor competition organization, and the experience of athletes and spectators are all highly sensitive to weather conditions: low-altitude wind fields affect the safety of the main torch and the organization of high-altitude performances; precipitation and strong convection may interfere with the progress of the event and pose safety risks to personnel; and factors such as temperature, humidity, and ultraviolet radiation are directly related to the competitive state of athletes and the comfort of spectators. This platform addresses the needs of large-scale sporting event support by constructing a three-dimensional meteorological support system with the entire event area, core venues, and key locations as its hierarchical levels. Through high-precision three-dimensional digital modeling of the event venues and their surrounding environment, a spatial carrying capacity and risk expression framework consistent with the business scenario is formed.

[0068] In terms of data and service capabilities, the platform accesses multi-source meteorological data updated every minute, including minute-level and meter-level three-dimensional microscale wind field data, micro-station observation data of the opening ceremony venue, as well as three-dimensional radar echo combined reflectivity products, gridded numerical forecast products and high-precision multi-source fusion meteorological data products. Through three-dimensional visualization, it transforms abstract weather risks into intuitive, dynamic and quantifiable spatial entities, achieving full-view decision support from macro-level situation to micro-level impact. The platform provides immersive 3D real-scene monitoring capabilities, covering the venue and a 3-kilometer radius around it. It supports switching between multiple perspectives, including panoramic, overhead, and internal / external views. It uses LOD (Level of Detail) to achieve smooth scaling and 360-degree free rotation from a height of 5 kilometers to 100 meters above the ground. Simultaneously, through the data panel, it enables the fusion and analysis of multi-source data, such as venue micro-stations, micro-scale wind fields, and 3D radar. It provides hourly and minute-level time-series curve analysis and supports comparison and judgment of real-time and forecast data by selected time periods. The tool panel provides functions such as perspective reset, data time-series graphs, multi-view switching, and operation instructions. It also integrates national and provincial Tianqing data sources to ensure data authority and real-time performance, thereby supporting professional decision-making in event organization and emergency response.

[0069] The beneficial effects of this invention are as follows: Existing meteorological support platforms primarily offer two-dimensional meteorological information services based on planar maps. This invention provides a three-dimensional, immersive decision support tool for meteorological support work during major national events. On one hand, this invention breaks through the limitations of traditional two-dimensional meteorological charts, achieving a three-dimensional representation of meteorological elements and more intuitively showcasing the three-dimensional meteorological conditions of the supported objects in a real environment. On the other hand, this technology simulates and displays the generation, evolution, and dissipation of three-dimensional cloud clusters in real time, dynamically maps the flow trajectory of low-level wind fields, and realistically reproduces the three-dimensional structure and movement path of precipitation systems, providing non-meteorological users with insights into the spatiotemporal and physical changes and development of meteorological elements. Furthermore, for large-scale sporting events, this technology can perform minute-level dynamic analysis of meteorological conditions in key areas such as the main stadium and outdoor tracks; for international summits, it can conduct real-time assessments of meteorological risks around the venue, along transportation routes, and in important reception areas, significantly improving the accuracy of meteorological information services and the timeliness of emergency command.

[0070] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0071] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A customized multi-dimensional meteorological digital twin method for major event support scenarios, characterized in that, include: Obtain raw meteorological data within the coverage area; The raw meteorological data is subjected to format adaptive automatic parsing technology. Based on the file naming rules and header file format, the corresponding parsing program is matched to identify the radar volume scan file and the three-dimensional real grid field file, and the parsed data is converted into a unified intermediate data body. The intermediate data body contains at least a timestamp, feature name and corresponding original feature value. Based on the intermediate data body, the multi-dimensional features of meteorological information are correlated with key elements of mapping and rendering, completing the mapping and transformation from meteorological data to graphic rendering elements, and using GPU floating-point parallel computing technology to generate a texture array for WebGL 3D texture drawing. The texture array is modeled, rendered, and displayed in real time using WebGL 3D texture rendering technology. During the rendering process, real-time clipping, dynamic loading, and progressive rendering techniques with multi-level detail models are employed. The refresh rate of scene rendering is optimized according to the matching graphics rendering quality to control the continuous updating of scene content and achieve progressive rendering. The dynamic loading includes selecting appropriate level-of-detail data for loading based on data layering, chunking, and dynamic updates of data pages. By integrating interactive functions such as volume rendering, layer slicing, profile analysis, and height field setting, the three-dimensional visualization scene obtained by real-time modeling and rendering of the texture array is explored and quantitatively analyzed at multiple scales, and the three-dimensional visualization results are output.

2. The customized multidimensional meteorological digital twin method for major event support scenarios as described in claim 1, acquiring raw meteorological data within the support area, includes: The API interface is used to read ground observation data, vertical observation data, radar and satellite data, model forecast products, local meter-level three-dimensional real-time wind products, and service support text products within the coverage area. The read data is then compiled to form the raw meteorological data.

3. The customized multi-dimensional meteorological digital twin method for major event support scenarios as described in claim 1, characterized in that, The raw meteorological data is subjected to format adaptive automatic parsing technology, including: The corresponding parser is matched based on file naming rules and header file format; The parsing program identifies radar volume scan files and three-dimensional real-time grid field files. The radar volume scan file and the three-dimensional real-time grid field file are parsed using a built-in binary format WebAssembly parser. The parsing results are standardized using a JSON data standardization processor; The standardized data is quality-assured by a multidimensional data quality verification system, resulting in parsed data.

4. The customized multi-dimensional meteorological digital twin method for major event support scenarios as described in claim 3, characterized in that, The parser matches the corresponding file naming rules and header file format, including: Perform file header feature recognition to obtain the first target data format identifier; When the format cannot be determined by the file header feature recognition, metadata analysis and recognition are performed to obtain the second target data format identifier; When the metadata analysis and identification cannot determine the format, naming rule identification is performed to obtain a third target data format identifier; When the naming rule identification cannot determine the format, a deep analysis of content features is performed to obtain the fourth target data format identifier; The parsing procedure that matches the original meteorological data is determined based on the first target data format identifier, the second target data format identifier, the third target data format identifier, or the fourth target data format identifier.

5. The customized multi-dimensional meteorological digital twin method for major event support scenarios as described in claim 1, characterized in that, The parsed data is then converted into a unified intermediate data body, including: For site products, spatiotemporal consistency processing, basic quality control filtering, feature extraction, and spatial range interpolation are performed on the parsed data to generate an intermediate data body containing timestamps, feature names, site metadata, feature values, and quality feature codes. For grid products, spatial clipping, feature value filtering, latitude and longitude-based point interpolation, feature extraction and statistics are performed on the parsed data to generate an intermediate data volume containing timestamps, feature names, feature 3D spatial range description information, original feature values ​​and analysis result values. For service assurance text products, keyword retrieval, time matching, and location service matching are performed on the parsed data to generate an intermediate data body containing timestamps, feature names, and corresponding original feature values.

6. The customized multi-dimensional meteorological digital twin method for major event support scenarios as described in claim 1, characterized in that, Complete the mapping and transformation from meteorological data to graphic rendering elements, including: Meteorological element datasets are used as input for mapping transformation, wherein the meteorological element datasets include a set of data elements, a set of element relationships, and a set of element spatial and attribute dimensions; The set of graphic rendering elements is used as the output of the mapping transformation, wherein the set of graphic rendering elements includes graphic structural attributes, spatiotemporal correlation of graphic display, and a set of cartographic components. Establish a functional correspondence between the input and the output, and complete the mapping transformation based on the functional correspondence.

7. The customized multi-dimensional meteorological digital twin method for major event support scenarios as described in claim 1, characterized in that, The texture array for WebGL 3D texture rendering is generated using GPU floating-point parallel computing technology, including: Based on the result of the mapping transformation, the intermediate data volume is textured and encoded to obtain a texture array; The texture array is set to an unsigned byte format to enable rapid generation of the texture array.

8. The customized multi-dimensional meteorological digital twin method for major event support scenarios as described in claim 1, characterized in that, The texture array is modeled, rendered, and displayed in real time using WebGL 3D texture rendering technology, including: GPU-accelerated algorithms are used to generate and render 2D mesh data into 3D terrain surfaces, isosurface meshes, and particle system geometry in real time. Balance visual effects and rendering performance through adaptive subdivision technology; The Cesium rendering pipeline is used to finalize the real-time modeling and rendering results.

9. The customized multi-dimensional meteorological digital twin method for major event support scenarios according to claim 1, characterized in that, The real-time cropping and the dynamic loading include: A real-time occlusion clipping algorithm is used to remove invisible polygons early in the graphics pipeline; Dynamic loading of 3D meteorological data is based on data layering, block partitioning, and dynamic updating of data pages; During the dynamic loading process, detailed level data that matches the line-of-sight is selected for loading.

10. The customized multi-dimensional meteorological digital twin method for major event support scenarios according to claim 1, characterized in that, Interactive functions including volume rendering, layered slicing, profile analysis, and height field settings are integrated, including: In a 3D scene, the profile of the 3D meteorological data rendering model is redrawn along a specified plane to generate a profile map and render it in real time. Based on the profile, perform profile analysis and provide profile meteorological data values, timestamps, comparative change analysis, and profile color transparency control.

Images