Scene generation method and device based on multi-source GIS data fusion

Abstract

The invention discloses a scene generation method and device based on multi-source GIS data fusion, and relates to the technical field of digital twinning and programmed generation. The method comprises the following steps: acquiring GIS data, preprocessing the GIS data, and storing the preprocessed GIS data in a geographic information resource library; a structured resource library is constructed, semantic parameters are added to the three-dimensional model in the structured resource library through the configuration file, and three-dimensional model resources with structured semantics are constructed; on the basis of the three-dimensional model resources with structured semantics and GIS data in a geographic information resource library, building and road generation and terrain processing are carried out in a programmed modeling engine through a configuration file, and scene data are generated; and importing the generated scene data into a real-time rendering engine, carrying out dynamic environment interaction and biocenosis simulation, and generating a city scene. The problems that in the prior art, an urban three-dimensional modeling method is low in efficiency and insufficient in environment interaction reality sense are solved.

Description

Technical Field

[0001] This application relates to the fields of digital twin and procedural generation technology, and in particular to a scene generation method and apparatus based on multi-source GIS data fusion. Background Technology

[0002] With the acceleration of urbanization and the rise of emerging concepts such as digital twins and smart cities, the demand for 3D urban models is becoming increasingly urgent and the requirements are becoming more stringent. As an important carrier of urban spatial information, 3D urban models not only provide intuitive decision-making basis for urban planning, management, and operation, but also play a key role in many fields such as disaster simulation, autonomous driving testing, and virtual tourism.

[0003] Currently, urban 3D modeling mainly relies on technologies such as oblique photogrammetry, laser scanning, BIM modeling, and manual modeling. Oblique photogrammetry requires the use of large-scale drones for field data collection, and subsequent data processing is time-consuming and labor-intensive. Traditional modeling workflows are rigid and difficult to respond to real-time changes in the city (such as new buildings and road reconstruction). Updates require re-collecting data and reconstructing the entire process, resulting in poor timeliness. In addition, traditional modeling techniques mainly rely on static data, which is seriously insufficient in simulating the realism of environmental interactions. It is difficult to achieve realistic simulations of dynamic water flow, swaying plants in the wind, and other effects closely related to environmental physical characteristics. These realistic environmental interactions are crucial for immersive simulation applications; the lack of these effects will seriously affect the reliability and practicality of simulation results. Summary of the Invention

[0004] In this embodiment of the application, a scene generation method based on multi-source GIS data fusion is provided, which solves the problems of low efficiency and insufficient realism of environmental interaction in existing urban 3D modeling methods.

[0005] In a first aspect, embodiments of this application provide a scene generation method based on multi-source GIS data fusion. The method includes: acquiring GIS data, preprocessing the GIS data, and storing the preprocessed GIS data in a geographic information resource database; wherein the GIS data includes OSM data, DEM data, satellite remote sensing image data, and water ripple data; constructing a structured resource database, adding semantic parameters to the 3D models in the structured resource database using configuration files to construct 3D model resources with structured semantics; wherein the resources in the structured resource database include parametric building components, road components, vegetation, materials, and special effects; based on the 3D model resources with structured semantics and the GIS data in the geographic information resource database, generating buildings, roads, and processing terrain in a procedural modeling engine using configuration files to generate scene data; and importing the generated scene data into a real-time rendering engine to perform dynamic environmental interaction and biological community simulation to generate an urban scene.

[0006] In one possible implementation, the acquisition of GIS data, preprocessing of the GIS data, and storage of the preprocessed GIS data in a geographic information resource database include: downloading OSM data of the target area from the OpenStreetMap platform, including geometric information and attribute labels of key vector elements such as road networks and building outlines; acquiring DEM data of the target area from a geospatial data cloud platform, including topographic elevation information; acquiring satellite remote sensing image data of the target area through remote sensing data processing software, including true color or infrared multispectral information of the land surface; acquiring urban electronic map data of the target area through a map source download tool, including thematic maps of water system geographic features; and using GIS software to perform coordinate system unification, geometric clipping, and semantic information extraction on the GIS data to obtain preprocessed GIS data.

[0007] In one possible implementation, the construction of the structured resource library involves adding semantic parameters to the 3D models in the structured resource library through configuration files to construct 3D model resources with structured semantics. This includes: using a hybrid architecture of object storage and spatial database to store resources when constructing the structured resource library; wherein unstructured digital asset files are stored in object storage, and metadata and their spatial relationship indexes are stored in the spatial database; setting the parameters to be generated in the structured resource library through a standardized JSON format configuration file; parsing and loading the contents of the configuration file through the Python script interface embedded in the procedural modeling engine; injecting the parameter set obtained by the Python script parsing into the global attribute system of the procedural modeling engine in the form of key-value pairs; the procedural modeling engine initiating a dual-condition query request to the spatial database based on attributes and spatial location according to the parameters in the configuration file; the spatial database returning a unique identifier of the matching asset, the path of the object storage, and the associated metadata attribute set according to the query conditions; and loading the required unstructured digital asset files from the object storage according to the returned object storage path to construct 3D model resources with structured semantics.

[0008] In one possible implementation, the generation of scene data, based on 3D model resources with structured semantics and GIS data from a geographic information resource library, involves building, road, and terrain processing in a procedural modeling engine using configuration files. This includes: extracting building outline polygons from OSM data and building heights from configuration files to generate building outline geometry; using the building outline geometry and building generation rules from configuration files, generating floor models, facade details, floor slabs, bases, and roofs using the procedural modeling engine to generate building clusters; generating road plan data, medians, and road equipment based on road centerline vectors from OSM data and road generation rules from configuration files to construct a complete road network; generating basic terrain and water body grids in the Gaea terrain engine based on DEM and water ripple data, sculpting negative terrain structures of rivers and lakes using Boolean subtraction operations, and generating terrain grids; and distinguishing green areas, soil areas, and building areas based on the colors of remote sensing image data and semantically attaching these areas to the terrain grid for local optimization and material allocation in the procedural modeling engine.

[0009] In one possible implementation, importing the generated scene data into a real-time rendering engine for dynamic environmental interaction and biome simulation to generate an urban scene includes: in a procedural modeling engine, using terrain data as a reference, integrating roads and terrain through geometric projection and terrain shaping techniques, integrating buildings and terrain through spatial positioning and surface adaptation techniques, and integrating the procedurally generated terrain, road network, and building clusters to generate a unified scene; using the particle system and material system of the real-time rendering engine, creating a water simulation based on a particle fluid system to simulate physically driven water effects and vegetation wind effects; and using the procedural content generation framework of the real-time rendering engine, procedurally distributing vegetation and surface rock elements according to the layered data of terrain attributes to generate an ecologically reasonable biome, and performing biome simulation to generate an urban scene.

[0010] Secondly, embodiments of this application provide a scene generation device based on multi-source GIS data fusion. The device includes: an acquisition module for acquiring GIS data, preprocessing the GIS data, and storing the preprocessed GIS data in a geographic information resource database; wherein the GIS data includes OSM data, DEM data, satellite remote sensing image data, and water ripple data; a construction module for constructing a structured resource database, adding semantic parameters to the 3D models in the structured resource database using configuration files to construct 3D model resources with structured semantics; wherein the resources in the structured resource database include parametric building components, road components, vegetation, materials, and special effects; a generation module for generating scene data based on the 3D model resources with structured semantics and the GIS data in the geographic information resource database, performing building and road generation and terrain processing in a procedural modeling engine using configuration files; and a rendering module for importing the generated scene data into a real-time rendering engine for dynamic environmental interaction and biological community simulation to generate an urban scene.

[0011] Thirdly, embodiments of this application provide a scene generation server based on multi-source GIS data fusion, including a memory and a processor; the memory is used to store computer-executable instructions; the processor is used to execute the computer-executable instructions to implement the method described in the first aspect or any possible implementation of the first aspect.

[0012] Fourthly, embodiments of this application provide a computer-readable storage medium storing executable instructions, which, when executed by a computer, enable the method described in the first aspect or any possible implementation thereof.

[0013] One or more technical solutions provided in this application embodiment have at least the following technical effects: This application embodiment provides a scene generation method based on multi-source GIS data fusion, which involves acquiring GIS data, preprocessing the GIS data, and storing the preprocessed GIS data in a geographic information resource database; wherein, the GIS data includes OSM data, DEM data, satellite remote sensing image data, and water ripple data; constructing a structured resource database, adding semantic parameters to the 3D models in the structured resource database through configuration files, and constructing 3D model resources with structured semantics; wherein, the resources in the structured resource database include parametric building components, road components, vegetation, materials, and special effects; based on the 3D model resources with structured semantics and the GIS data in the geographic information resource database, building and road generation and terrain processing are performed in a procedural modeling engine through configuration files to generate scene data; the generated scene data is imported into a real-time rendering engine for dynamic environmental interaction and biological community simulation to generate an urban scene. This solves the problems of low efficiency and insufficient realism of environmental interaction in existing urban 3D modeling methods. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments of this application or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 A flowchart illustrating a scene generation method based on multi-source GIS data fusion, provided for embodiments of this application; Figure 2 Satellite remote sensing imagery provided for embodiments of this application; Figure 3 A schematic diagram of water ripple data after semantic information extraction from a city's electronic map; Figure 4 A schematic diagram of the geometric outline of a building with language provided in the embodiments of this application; Figure 5 A schematic diagram of the generated building facade provided in an embodiment of this application; Figure 6 A schematic diagram illustrating the generation of building floor slabs and bases provided in an embodiment of this application; Figure 7 A schematic diagram of the generated building roof provided for an embodiment of this application; Figure 8 A schematic diagram illustrating the generation of building clusters provided in an embodiment of this application; Figure 9 This is a schematic diagram illustrating the generation of road plan data provided in an embodiment of this application; Figure 10 A schematic diagram of the traffic light arrangement at an intersection provided in an embodiment of this application; Figure 11 A schematic diagram of a road divider provided in an embodiment of this application; Figure 12 A schematic diagram of the generated building outline geometry provided in the embodiments of this application; Figure 13 This is a schematic diagram of terrain mesh generation provided in an embodiment of this application; Figure 14 This is a schematic diagram of scene terrain rendering provided in an embodiment of this application; Figure 15 A schematic diagram of the generated non-urban area provided in the embodiments of this application; Figure 16 A schematic diagram of the generated urban area provided in the embodiments of this application; Figure 17A schematic diagram of a scene generation device based on multi-source GIS data fusion provided in an embodiment of this application; Figure 18 This is a schematic diagram of a scene generation server based on multi-source GIS data fusion, provided as an embodiment of this application. Detailed Implementation

[0016] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0017] The following description of some technologies involved in the embodiments of this application is provided to aid understanding and should be considered merely exemplary. Therefore, those skilled in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this application. Similarly, for clarity and brevity, some descriptions of well-known functions and structures are omitted in the following description.

[0018] This application provides a method for scene generation based on multi-source GIS data fusion, such as... Figure 1 As shown, the method includes steps S101 to S104. Wherein, Figure 1 This is merely one execution order shown in the embodiments of this application, and does not represent the only execution order for a scene generation method based on multi-source GIS data fusion. Where the final result can be achieved, Figure 1 The steps shown can be performed in parallel or in reverse order.

[0019] S101: Acquire GIS data, preprocess the GIS data, and store the preprocessed GIS data in a geographic information resource database. The GIS data includes OSM data, DEM data, satellite remote sensing imagery data, and water ripple data.

[0020] Acquire GIS (Geographic Information System) data, preprocess the GIS data, and store the preprocessed GIS data in a geographic information resource database, including the following:

[0021] Download OSM data for the target area from the OpenStreetMap platform, including geometric information and attribute labels for key vector elements of the road network and building outlines.

[0022] Specifically, download OSM (vector road network) data (OSM) of the target area from the OpenStreetMap platform. Using the data export function provided by its official platform, select the download area of ​​the target region (city) based on geographic latitude and longitude coordinates or preset administrative boundaries. The exported data uses the open .osm format (a standard geographic data format based on XML), which fully contains the geometric information and attribute labels of key vector elements such as road networks and building outlines within the selected area.

[0023] Obtain DEM data of the target area from the geospatial data cloud platform, including terrain and elevation information.

[0024] Specifically, DEM data is topographic elevation data. In the geospatial data cloud platform, the dataset type (e.g., ASTER GDEM, SRTM), spatial reference system (e.g., WGS84, CGCS2000), and spatial extent of the target area (defined by latitude and longitude or map sheet number) need to be precisely configured. The downloaded data is a GeoTIFF (.tif) raster file, which embeds geographic coordinate information, and the grayscale value of each pixel represents the elevation of that point.

[0025] Satellite remote sensing image data of the target area is obtained through remote sensing data processing software, including the true color of the ground surface or infrared multispectral information.

[0026] Specifically, satellite remote sensing imagery data of the target area is acquired using remote sensing data processing software (such as SAS Planet). This software supports multi-source imagery services (such as Google Earth and Bing Maps), and users can select and download imagery of the target area through its map interface. The downloaded imagery is also in GeoTIFF format, containing the true color of the Earth's surface or multispectral information such as infrared, primarily used for subsequent surface texture mapping and visual enhancement.

[0027] Obtain urban electronic map data for the target area using a map source download tool, including thematic maps of water system geographical features.

[0028] Specifically, urban electronic map data for the target area is obtained using a map source download tool (such as an all-in-one map downloader). A professional map service containing clear geographical features such as waterways is selected and provided as the source to download map tile data for the target area. The final output is a raster image in .tif format, containing a thematic map including elements such as waterways.

[0029] GIS software is used to perform coordinate system 1, geometric clipping, and semantic information extraction on GIS data to obtain preprocessed GIS data.

[0030] GIS (Geographic Information System) software can include QGIS, ArcGIS, and Global Mapper.

[0031] The specific details of Coordinate System 1 are as follows. Due to the diversity of coordinate systems, the downloaded raw data cannot be directly used in the generation process. Professional GIS software must be used to uniformly convert all data to the same projected coordinate system (such as UTM-WGS84). The specific details of geometric clipping are as follows. Raw data often covers too large an area or contains redundant information. Geometric clipping is performed based on precise project boundaries (such as administrative division polygons or custom ranges) to obtain precise elevation data, image data, and water system data. This process ensures spatial consistency of the data and lays the foundation for multi-source data fusion. The specific details of semantic information extraction are as follows. For specific geographic elements such as water systems, semantic information needs to be further extracted from the clipped base raster data. The thematic map raster data containing water systems is converted into bitmap data and imported into image processing software (such as Adobe Photoshop). Using its color selection tools or channel separation technology, based on the specific color or spectral characteristics of water bodies, thresholding is used to accurately identify and select water body pixel areas.

[0032] Figure 2 Satellite remote sensing imagery provided for embodiments of this application. Figure 2 Specifically, it is used to distinguish terrain textures. The range of terrain textures such as green space and soil is extracted from remote sensing imagery, which serves as the basis for terrain textures in Unreal Engine. Figure 3 This is a schematic diagram of water ripple data after semantic information extraction from a city's electronic map. (Example:) Figure 3 As shown, the selected area is output as an independent, transparent raster image (such as PNG format). This image is the semantically meaningful water ripple data used in the procedural generation, with the white portion representing the water area. After the above preprocessing, the preprocessed GIS data that meets the requirements is obtained and stored in the geographic information resource database.

[0033] S102: Construct a structured resource library. Add semantic parameters to the 3D models in the structured resource library through configuration files to construct 3D model resources with structured semantics. The resources in the structured resource library include parametric building components, road components, vegetation, materials, and special effects.

[0034] Build a structured resource library, add semantic parameters to the 3D models in the structured resource library through configuration files, and build 3D model resources with structured semantics, including the following content.

[0035] When constructing the structured resource repository, a hybrid architecture consisting of object storage and a spatialized database is adopted. Unstructured digital asset files are stored in object storage, while metadata and their spatial relationship indexes are stored in the spatialized database.

[0036] The spatialized database in this application can be a spatialized MySQL database.

[0037] Specifically, a hybrid architecture consisting of Object Storage Service (OSS) and a spatialized database is used to store resources. This architecture separates storage from indexing, fully leveraging the advantages of both storage methods. All unstructured digital asset files, such as 3D models (parametric building components, road components), texture maps, and effect sequences, are stored in the object storage in a highly reliable and low-cost manner. For example, 3D model files of windows, doors, roofs, and walls in parametric building components, along with their associated texture resources, are stored here, ensuring the secure storage and convenient access to a large amount of unstructured data. The metadata corresponding to the object storage (including asset name, type, geometric dimensions, physical material parameters, spatial bounding box, geographic coordinate labels, etc.) and their spatial relationship indexes are stored in a database that supports spatial queries. Taking parametric building components as an example, the parametric attributes of each component (such as style type, size range, and material mapping relationship) are stored as metadata in the database, facilitating efficient retrieval based on attributes and spatial location.

[0038] The parameters required for generating the structured repository are set through a standardized JSON-formatted configuration file.

[0039] Specifically, a standardized JSON-formatted configuration file (config.json) is used to define and control the parameters required for generating the structured resource library throughout the programmatic generation process. This configuration file covers detailed parameter information for various types of resources.

[0040] Core configuration parameters for components: Define the hierarchical affiliation of components in the structure (using a hierarchical coding format, such as: building -> exterior wall -> window), component model source path (specifying the complete access path of the component model file in object storage (OSS), component module index (used to specify the index number of the extracted sub-components when the source model is a composite model), etc.

[0041] Building core configuration parameters include: OSM data source path (the file path for storing OSM source data), building area filtering threshold (the area threshold used to filter out buildings that are too small, in square meters), building height mapping (a dictionary structure that maps building IDs in OSM to their specified height values, in meters), etc.

[0042] Core road configuration parameters include: road ID (a unique identifier for a road), road red line width (the total width of the planned road red line, in meters), and road grade (the technical grade of a road, such as: expressway, arterial road, secondary arterial road, local road).

[0043] Core terrain configuration parameters include: Gaea project file path (storage path of Gaea terrain project files), maximum elevation (highest elevation of the terrain area in meters), minimum elevation (lowest elevation of the terrain area in meters), etc.

[0044] The contents of the configuration file are parsed and loaded through the Python script interface embedded in the programmatic modeling engine.

[0045] The parameter set obtained by parsing the Python script is injected into the global attribute system of the procedural modeling engine in the form of key-value pairs.

[0046] Specifically, after the system starts, the procedural modeling engine (such as the Houdini engine) parses and loads all the contents of the configuration file (config.json) through its embedded Python script interface. The specific process is as follows: the Python script reads the configuration file and parses the parameter set within it.

[0047] The parsed parameter set is systematically injected into the global attributes system of the procedural modeling engine in the form of key-value pairs. For example, parameters such as architectural style and height are injected into the global attributes system, providing a unified and callable parameter access interface for all subsequent procedural generation nodes. In this way, during the procedural generation process, each generation node can perform corresponding operations based on these global parameters.

[0048] The procedural modeling engine initiates a dual-condition query request to the spatial database based on attributes and spatial location, according to the parameters in the configuration file.

[0049] Specifically, the spatial database performs high-performance joint retrieval based on multiple constraints contained in the instructions (such as attribute conditions: architectural style = modern, window type = floor-to-ceiling window; spatial conditions: target geographic bounding box). The retrieval results not only return the unique identifier and OSS storage path of the matching asset, but also its associated metadata attribute set (such as component size tolerance, spatial positioning rules, LOD switching threshold).

[0050] Procedural modeling engines (such as the Houdini engine) dynamically load the basic mesh model based on the returned path information and automatically perform geometric transformations and instantiation operations based on the parameters in the metadata. (1) Geometric deformation adaptation: Through nodes such as Attribute Wrangle, the size parameters in the metadata are applied to the scaling transformation of the model, or the deformation nodes are driven to modify the vertex positions to generate component instances that conform to specific specifications. (2) Spatial transformation and alignment: The required position and rotation matrices of the instance are automatically calculated, and the alignment rules (such as anchor points and offsets) defined in the metadata are used for precise positioning. (3) PBR material binding: Based on the material identifier recorded in the metadata, the corresponding material sphere and texture resources (background color, normal, metallicity, roughness) are retrieved from the preloaded PBR material library and dynamically assigned to the geometric instance. Material parameters (such as hue and wear) can be fine-tuned at runtime based on the metadata to increase variation and realism.

[0051] For example, to retrieve window models of a specific style within a specific region, the query criteria could include: "Retrieve all window models located within a geographic bounding box, with a style of 'modern' and a width between 1.2 and 1.5 meters." Upon receiving the query request, the spatial database uses pre-defined indexes and query algorithms to quickly match and retrieve data from the stored metadata.

[0052] Based on the query criteria, the spatialized database returns a unique identifier for the matching asset, the path to the object storage, and the associated set of metadata attributes.

[0053] Based on the returned object storage path, load the required unstructured digital asset files from the object storage to construct a 3D model resource with structured semantics.

[0054] For example, the 3D file and texture resources of the window model are loaded based on the returned path, thereby obtaining resources from the structured resource library and providing rich material support for subsequent urban scene generation.

[0055] S103: Based on 3D model resources with structured semantics and GIS data in the geographic information resource library, scene data is generated by generating buildings, roads and processing terrain in the procedural modeling engine through configuration files.

[0056] Based on 3D model resources with structured semantics and GIS data in the geographic information resource library, scene data is generated by performing building, road and terrain processing in the procedural modeling engine through configuration files, including the following:

[0057] Building outline geometry is generated by extracting building outline polygons from OSM data and building heights from configuration files. Based on the building outline geometry and building generation rules in the configuration files, a procedural modeling engine is used to generate floor models, facade details, floor slabs, bases, and roofs to generate building clusters.

[0058] Generating building outline geometry: The Labs toolchain of a procedural modeling engine (such as Houdini) reads OSM data obtained from a geographic information resource repository and extracts the building outline polygon information. Combined with the building height parameters defined in the configuration file, a procedural algorithm generates the basic 3D geometry representing the building volume, i.e., the building outline geometry. This step provides the basic framework for subsequent building detail generation.

[0059] Figure 4 This is a schematic diagram of a building outline geometry with language provided in an embodiment of this application. Based on attribute parameters in the structured semantics of the geometry, such as floor names, floor order configurations, floor height, and depth, the procedural modeling engine retrieves and calls corresponding parametric component models from the structured resource library. Through instantiated assembly technology, standard floor unit models are generated, providing basic modules for the vertical structure construction of the building.

[0060] Figure 5 This is a schematic diagram of the generated building facade provided in an embodiment of this application. Based on the architectural style attributes, floor order configuration, and total height information carried by the building outline geometry, the procedural modeling engine programmatically calls the pre-generated standard floor models. These models are vertically stacked and combined according to preset rules, ultimately generating a building shell with diverse facade details, making the building's appearance richer and more realistic.

[0061] Figure 6 This is a schematic diagram illustrating the generation of building floor slabs and bases according to an embodiment of this application. The procedural modeling engine automatically calculates the number of floor slabs based on the total building height and standard floor height. Simultaneously, based on associated building structural parameters, such as base shrinkage / expansion values, base thickness, floor slab thickness, and floor slab material index, the internal floor slab structure and bottom base of the building are generated through Boolean operations or polygon extrusion technology, thus perfecting the internal structure of the building.

[0062] Figure 7 This is a schematic diagram of generating a building roof according to an embodiment of this application. Based on the top polygon of the building outline geometry, the procedural modeling engine generates a matching 3D roof model according to its associated roof type parameters, such as flat roof, pitched roof, etc., as well as information such as slope, cross-sectional height, and position offset, through dedicated procedural rules, thus adding a complete top structure to the building.

[0063] Figure 8This diagram illustrates the generation of building clusters provided in an embodiment of this application. The PDG / TOPs module of the Houdini procedural modeling engine encapsulates the aforementioned building generation steps into a parallelizable task network. This network traverses all building outline polygons in the OSM data, automatically executing the complete process from generating the building outline geometry to generating the roof for each polygon. It outputs all building models and their correct spatial distribution relationships, thereby efficiently generating large-scale building clusters and forming the basic layout of urban architecture.

[0064] Based on the road centerline vector in the OSM data and the road generation rules in the configuration file, road plan data, median strips, and road equipment are generated to construct a complete road network.

[0065] Figure 9 This diagram illustrates the generation of road plan data provided in an embodiment of this application. The procedural modeling engine reads road centerline vectors and road attribute information, such as road right-of-way width and road grade, from OSM data obtained from a geographic information resource database. It automatically constructs road plan geometry using procedural algorithms, including intelligent identification and processing of road intersections. It can automatically generate intersection plans of different shapes, such as crossroads and T-junctions, based on topological relationships, ensuring seamless fitting of road surfaces at each intersection, making the road plan data more accurate and complete.

[0066] Based on the road centerline data and preset cross-sectional configuration parameters, including the number of lanes, width, and functional zones, the procedural modeling engine programmatically generates three-dimensional geometric models of road infrastructure such as the road centerline, the median strip between motor vehicle lanes and non-motor vehicle lanes, and the central divider (vehicle lane). This process ensures the standardization and accuracy of the road cross-sectional structure, providing a foundation for the realistic representation of the road.

[0067] Figure 10 This is a schematic diagram of the placement of traffic lights at an intersection provided in an embodiment of this application. Based on the road geometry network, the procedural modeling engine automatically deploys road facilities through spatial calculation rules. At road intersections, the precise spatial coordinates and orientations of facilities such as traffic lights are calculated and determined based on the intersection points of road centerlines, road direction vectors, and width information. According to the semantic rules of facility locations defined in the configuration information, such as "facilities with traffic lights" and "facilities with fire hydrants," in designated areas, such as sidewalks and green belts, sampling algorithms such as Poisson distribution are used to avoid overlap, and the corresponding facility model assets are instantiated and assembled into the scene to complete the construction of the entire road equipment system, making the road scene more complete and realistic. Figure 11 This is a schematic diagram of a road divider provided in an embodiment of this application. Figure 11 It is specifically generated by parameter configuration. Figure 12This is a schematic diagram of the generated building outline geometry provided in an embodiment of this application. Different colors represent different configurations.

[0068] Based on DEM and water ripple data, basic terrain and water meshes are generated in the Gaea terrain engine. The negative terrain structure of rivers and lakes is sculpted through Boolean subtraction operations, and terrain meshes are generated. The terrain meshes are then locally optimized and material assigned in the procedural modeling engine.

[0069] Based on remote sensing image data, green areas, soil areas, and building areas are distinguished according to the color of the image data, and the regions are attached to the terrain mesh in a semantic form so that the terrain mesh can be locally optimized and material assigned in the procedural modeling engine.

[0070] Figure 13 This diagram illustrates the terrain mesh generation method used in this application embodiment. The original DEM data is standardized and preprocessed in GlobalMapper, converting it into an elevation grayscale image (PNG format) that meets the input requirements of the Gaea terrain engine. Subsequently, the elevation grayscale image is imported into the Gaea terrain engine to generate the basic terrain mesh. Simultaneously, water ripple data, processed in Photoshop to extract semantic features, is imported into Gaea to generate a separate water body mesh. Through Boolean subtraction operations within Gaea, the water body mesh is subtracted from the basic terrain mesh, thereby sculpting realistic negative terrain structures such as rivers and lakes, generating a high-precision terrain mesh with hydrological characteristics, providing a foundation for natural landscape construction.

[0071] Procedural modeling engines (such as Houdini) seamlessly import terrain project files (.tor) output by Gaea through their dedicated data interface node (gaea_tor_processor), fully inheriting their mesh and attribute data. During terrain mesh processing, local optimizations are performed on the terrain, such as smoothing steep slope edges and repairing mesh penetrations, to ensure that the terrain's geometry blends seamlessly with subsequently generated building and road models, making the terrain more natural and realistic.

[0072] Figure 14 This is a schematic diagram of scene terrain rendering provided in an embodiment of this application. Based on attributes such as terrain slope, altitude, and surface humidity, and the color of image data, attributes such as green areas, soil areas, and building areas are distinguished. In the procedural modeling engine, PBR (Physically Based Rendering) materials from the material resource library, such as rocks, soil, and grassland, are mapped layer by layer. This provides a physical basis for biological community simulation, while also giving the terrain surface richer visual effects and enhancing the realism of the scene.

[0073] S104: Import the generated scene data into the real-time rendering engine to perform dynamic environmental interaction and biological community simulation, and generate urban scenes.

[0074] The generated scene data is imported into a real-time rendering engine to perform dynamic environmental interaction and biological community simulation, generating urban scenes, including the following:

[0075] In the procedural modeling engine, based on terrain data, roads and terrain are integrated through geometric projection and terrain shaping technology, and buildings and terrain are integrated through spatial positioning and surface adaptation technology, thus merging the procedurally generated terrain, road network and building clusters to generate a unified scene.

[0076] Specifically, the procedural modeling engine can be the Houdini engine. The road centerline is sampled and snapped based on terrain elevation using ray projection nodes to ensure the road path accurately conforms to the surface undulations. Subsequently, the linear geometry is converted into road surface polygons with actual width and input as a mask into the height field system. The specified area is flattened using height field projection nodes, supplemented by blurring or erosion operations, making the road area smoother and more naturally integrated with the natural terrain while maintaining flatness. The source point data for building distribution is also fixed to the terrain surface via ray projection. To achieve a correct fit between the building posture and the sloping terrain, surface normal data needs to be obtained through attribute passing, and rotation matrices or quaternion orientations (orient attributes) are calculated accordingly. This ensures that when replicating the building model, each instance automatically aligns with the surface normal at its location, ensuring the building model is stably placed on the terrain. Through this method, the procedurally generated terrain, road network, and building clusters are integrated into an organic whole, constructing a unified and reasonable scene foundation.

[0077] By utilizing the particle and material systems of a real-time rendering engine, a water simulation based on a particle fluid system is created to simulate physically driven water effects and vegetation wind effects.

[0078] Specifically, the real-time rendering engine can be Unreal Engine. For various water bodies in the scene, such as rivers, lakes, and pools, a particle-based simulation is created. This system simulates the basic motion of water by solving simplified Navier-Stokes equations. By adjusting particle parameters such as velocity, viscosity, and gravity, realistic water surface ripples, waves, eddies, and water flow direction are generated programmatically. Simultaneously, the particle system has a collision query module that can detect the interaction between water particles and dynamic objects in the scene (such as vehicles, pedestrians, and rainfall) in real time. When a collision is detected, new particle generation rules are dynamically triggered, such as a vehicle splashing water when driving over a puddle or raindrops creating spreading ripples, thus achieving real-time, physically accurate feedback from the water body to the environment. Furthermore, the data generated by the particle simulation is output to the rendering target in real time. The water material reads this rendering target data and dynamically modulates the surface normal map, specular highlights, and transparency, making the water surface exhibit extremely realistic interactive effects, rather than simple static texture animation. In the real-time rendering engine, a directional wind field Actor (a virtual object used to simulate and control the direction of the wind field) is created to define the direction, speed, and intensity of the wind. This wind field serves as a global physical parameter in the scene, providing a unified dynamic basis for all vegetation. For vegetation assets such as tree branches and grass, dedicated materials based on world position offsets are created. In these materials, a combination of sine waves and noise functions is used, with the root of the branch as a pivot point, to dynamically calculate the bending amplitude and direction of each leaf or branch under wind influence. The material's input parameters, such as wind intensity and swaying frequency, are bound to the global wind field parameters to ensure that all vegetation in the scene responds consistently and harmoniously to the wind. To avoid the "paper-thin" effect caused by overly uniform vegetation movement, hierarchical and randomization factors are introduced into the material calculations. For example, the branches and leaves at the top of the tree canopy sway more than those at the bottom, and different branches and leaves on the same tree have random initial phases of swaying. In addition, the computational load is optimized according to the LOD level (Level of Detail) of the vegetation. A more complex wind calculation model is used for foreground vegetation, while simplified vertex animation is used for background vegetation, so as to ensure both effect and rendering performance.

[0079] By utilizing the procedural content generation framework of a real-time rendering engine, vegetation and surface rock elements are procedurally distributed based on the layered data of terrain attributes to generate ecologically reasonable biological communities. Biological community simulation is then performed to generate urban scenes.

[0080] Specifically, after importing the base scene generated in the procedural modeling engine (Houdini engine) into the real-time rendering engine (Unreal Engine), the primary task is to prepare the data and build the foundation for procedural content generation. This typically requires correctly transferring key attribute data such as height, slope, road area masking, and building area masking generated in the procedural modeling engine to the real-time rendering engine through the Landscape layer (terrain layer) or vertex shader pass. Simultaneously, a corresponding Procedural Content Generation Volume (PCG Volume) is created to cover the target generation area and associated with the Landscape in the scene, establishing the physical boundaries and coordinate references for procedural generation, laying the data foundation for subsequent content distribution. The main construction phase of the PCG graph (Procedural Content Generation Graph) focuses on the design of content distribution logic based on surface sampling and attribute-driven principles. Uniformly or randomly distributed point clouds are generated on the Landscape surface using Surface Sampler nodes (nodes in Unreal Engine) as potential content placement locations. Subsequently, the Get Landscape Data node is used to read terrain attributes, and distribution rules are implemented through Density Filter or Filter ByRange nodes. For example, rocks are distributed based on slope and height thresholds, or occlusion data imported from the procedural modeling engine is used to remove generation points in road and building areas using attribute query nodes, ensuring that vegetation and environmental elements do not penetrate artificial structures. After establishing the basic distribution rules, biome segmentation and distance-based generation strategies are further introduced to enhance the ecological rationality and hierarchy of the scene. By combining multiple attributes such as height, slope, and occlusion, different biomes are segmented using Select by Range nodes, such as forests, grasslands, and rocky mountains, and different combinations of vegetation and props are assigned to each area. For dynamic elements such as animals or insects, generation is usually combined with biome type and specific conditions, such as being far from roads or near water sources, and Blueprint Spawner nodes can be used instead of static meshes to integrate more complex behavioral logic.

[0081] Figure 15 This is a schematic diagram of a non-urban area generated according to an embodiment of this application. Figure 16This is a schematic diagram of the generated urban area provided in the embodiments of this application. The final stage focuses on system optimization and debugging, as well as the coordination of the overall ecosystem. Rendering efficiency is significantly improved by applying Hierarchical Instancing nodes, and visual fine-tuning is performed on the generated results, such as adjusting the random rotation and scaling of vegetation to align with ground normals, ensuring a natural feel. The real-time debugging capabilities provided by the procedural content generation framework (PCG framework) of the real-time rendering engine (Unreal Engine) are utilized to dynamically adjust rule parameters and preview changes in real time, enabling rapid iteration. This non-destructive workflow allows for continuous optimization and enrichment of the scene's environmental details while maintaining the basic structure of the procedural modeling engine (Houdini engine), ultimately forming a highly efficient and realistic procedural ecosystem, thereby generating a complete and realistic urban scene.

[0082] This application also provides a scene generation device 1700 based on multi-source GIS data fusion, such as... Figure 17 As shown, the device includes: an acquisition module 1701, a construction module 1702, a generation module 1703, and a rendering module 1704.

[0083] The acquisition module 1701 is used to acquire GIS data, preprocess the GIS data, and store the preprocessed GIS data in a geographic information resource database. The GIS data includes OSM data, DEM data, satellite remote sensing image data, and water ripple data.

[0084] Module 1702 is used to build a structured resource library. Semantic parameters are added to the 3D models in the structured resource library via configuration files, creating 3D model resources with structured semantics. The resources in the structured resource library include parametric building components, road components, vegetation, materials, and effects.

[0085] The generation module 1703 is used to generate scene data based on 3D model resources with structured semantics and GIS data in the geographic information resource library. It performs building, road generation and terrain processing in the procedural modeling engine through configuration files.

[0086] The rendering module 1704 is used to import the generated scene data into the real-time rendering engine to perform dynamic environment interaction and biological community simulation, and generate urban scenes.

[0087] Some modules in the apparatus described in this application can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, classes, etc., that perform a specific task or implement a specific abstract data type. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0088] The apparatus or module described in the above embodiments can be implemented by a computer chip or physical entity, or by a product with a certain function. For ease of description, the above apparatus is described by dividing it into various modules according to their functions. When implementing the embodiments of this application, the functions of each module can be implemented in one or more software and / or hardware. Of course, a module that implements a certain function can also be implemented by combining multiple sub-modules or sub-units.

[0089] The methods, apparatus, or modules described in this application can be implemented in a computer-readable program code manner. The controller can be implemented in any suitable manner, for example, as a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicon Labs C8051F320. A memory controller can also be implemented as part of the control logic of a memory. Those skilled in the art will also recognize that, in addition to implementing the controller in purely computer-readable program code manner, the same functionality can be achieved by logically programming the method steps to make the controller take the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the means included within it for implementing various functions can also be considered as structures within the hardware component. Alternatively, the device used to implement various functions can be viewed as either a software module that implements the method or a structure within a hardware component.

[0090] like Figure 18As shown, this application embodiment also provides a scene generation server based on multi-source GIS data fusion, including a memory 1801 and a processor 1802; the memory 1801 is used to store computer-executable instructions; the processor 1802 is used to execute computer-executable instructions to implement the scene generation method based on multi-source GIS data fusion described above in this application embodiment.

[0091] This application also provides a computer-readable storage medium storing executable instructions, which, when executed by a computer, enable the scene generation method based on multi-source GIS data fusion described above in this application embodiment.

[0092] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary hardware. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product, or it can be embodied in the process of data migration. The computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, mobile terminal, server, or network device, etc.) to execute the methods described in the embodiments of this application.

[0093] The various embodiments described in this specification are presented in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on its differences from other embodiments. All or part of this application can be used in numerous general-purpose or special-purpose computer system environments or configurations.

[0094] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of this application.

Claims

1. A scene generation method based on multi-source GIS data fusion, characterized in that, include: Acquire GIS data, preprocess the GIS data, and store the preprocessed GIS data in a geographic information resource database; the GIS data includes OSM data, DEM data, satellite remote sensing image data, and water ripple data; A structured resource library is constructed, and semantic parameters are added to the 3D models in the structured resource library through configuration files to create 3D model resources with structured semantics. The resources in the structured resource library include parametric building components, road components, vegetation, materials, and special effects. Based on 3D model resources with structured semantics and GIS data in the geographic information resource library, scene data is generated by generating buildings, roads and processing terrain through configuration files in the procedural modeling engine. The generated scene data is imported into a real-time rendering engine to perform dynamic environmental interaction and biological community simulation, thereby generating urban scenes.

2. The scene generation method based on multi-source GIS data fusion according to claim 1, characterized in that, The process of acquiring GIS data, preprocessing the GIS data, and storing the preprocessed GIS data in a geographic information resource database includes: Download OSM data for the target area from the OpenStreetMap platform, including geometric information and attribute labels for key vector elements of road networks and building outlines; Obtain DEM data of the target area from the geospatial data cloud platform, including terrain and elevation information; Satellite remote sensing image data of the target area is obtained through remote sensing data processing software, including the true color of the ground surface or infrared multispectral information. Obtain urban electronic map data of the target area through map source download tools, including thematic maps of water system geographical features; GIS software is used to perform coordinate system 1, geometric clipping, and semantic information extraction on GIS data to obtain preprocessed GIS data.

3. The scene generation method based on multi-source GIS data fusion according to claim 1, characterized in that, The construction of the structured resource library involves adding semantic parameters to the 3D models in the structured resource library through configuration files to construct 3D model resources with structured semantics, including: When constructing the structured resource repository, a hybrid architecture consisting of object storage and spatial database is adopted to store the resources; among them, unstructured digital asset files are stored in object storage, and metadata and their spatial relationship indexes are stored in the spatial database. The parameters required to be generated in the structured repository are set through a standardized JSON format configuration file; The contents of the configuration file are parsed and loaded through the Python script interface embedded in the programmatic modeling engine. The parameter set obtained by parsing the Python script is injected into the global attribute system of the procedural modeling engine in the form of key-value pairs; The procedural modeling engine initiates a dual-condition query request to the spatial database based on attributes and spatial location, according to the parameters in the configuration file. Based on the query conditions, the spatialized database returns a unique identifier for the matching asset, the path to the object storage, and the associated set of metadata attributes; Based on the returned object storage path, load the required unstructured digital asset files from the object storage to construct a 3D model resource with structured semantics.

4. The scene generation method based on multi-source GIS data fusion according to claim 1, characterized in that, The scene data, generated by the procedural modeling engine using configuration files, is based on structured semantic 3D model resources and GIS data from a geographic information resource database. This process involves building and road generation, terrain processing, and other data creation. The scene data includes: Building outline geometry is generated by extracting building outline polygons from OSM data and building heights from configuration files. Based on the building outline geometry and building generation rules in the configuration files, a procedural modeling engine is used to generate floor models, facade details, floor slabs, bases, and roofs to generate building clusters. Based on the road centerline vector in OSM data and the road generation rules in the configuration file, road plan data, medians and road equipment are generated to construct a complete road network. Based on DEM data and water ripple data, a basic terrain mesh and a water body mesh are generated in the Gaea terrain engine. The negative terrain structure of the river channel and lake is sculpted through Boolean subtraction operation, and a terrain mesh is generated. Based on remote sensing image data, green areas, soil areas, and building areas are distinguished according to the color of the image data, and the regions are attached to the terrain mesh in a semantic form so that the terrain mesh can be locally optimized and material assigned in the procedural modeling engine.

5. The scene generation method based on multi-source GIS data fusion according to claim 1, characterized in that, The process of importing the generated scene data into a real-time rendering engine for dynamic environmental interaction and biological community simulation to generate urban scenes includes: In the procedural modeling engine, based on terrain data, roads and terrain are integrated through geometric projection and terrain shaping technology, and buildings and terrain are integrated through spatial positioning and surface adaptation technology. The procedurally generated terrain, road network and building clusters are integrated to generate a unified scene. By utilizing the particle and material systems of a real-time rendering engine, a water simulation based on a particle fluid system is created to simulate physically driven water effects and vegetation wind effects. By utilizing the procedural content generation framework of a real-time rendering engine, vegetation and surface rock elements are procedurally distributed based on the layered data of terrain attributes to generate ecologically reasonable biological communities. Biological community simulation is then performed to generate urban scenes.

6. A scene generation device based on multi-source GIS data fusion, characterized in that, include: The acquisition module is used to acquire GIS data, preprocess the GIS data, and store the preprocessed GIS data in the geographic information resource database; the GIS data includes OSM data, DEM data, satellite remote sensing image data, and water ripple data. The building module is used to construct a structured resource library. It adds semantic parameters to the 3D models in the structured resource library through configuration files to construct 3D model resources with structured semantics. The resources in the structured resource library include parametric building components, road components, vegetation, materials, and special effects. The generation module is used to generate scene data based on 3D model resources with structured semantics and GIS data in the geographic information resource library. It performs building, road generation and terrain processing in the procedural modeling engine through configuration files. The rendering module is used to import the generated scene data into the real-time rendering engine to perform dynamic environmental interaction and biological community simulation, and generate urban scenes.

7. A scene generation server based on multi-source GIS data fusion, characterized in that, Including memory and processor; The memory is used to store computer-executable instructions; The processor is configured to execute the computer-executable instructions to implement the method according to any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores executable instructions, which, when executed by a computer, enable the implementation of the method as described in any one of claims 1-5.

Images