A digital model generation method for tourist attraction display

Abstract

The present invention discloses a method for generating digital models for displaying tourist attractions, comprising the following steps: obtaining three-dimensional data of the tourist attraction, performing high-precision modeling on the three-dimensional data to obtain an initial three-dimensional model; based on the initial three-dimensional model, using a face reduction algorithm to reduce the geometric faces of the initial three-dimensional model to obtain a simplified three-dimensional model; setting multi-level model detail parameters according to preset display platform performance parameters, and generating multi-level models with different degrees of precision based on the simplified three-dimensional model and the multi-level model detail parameters; compressing the texture data in the multi-level model using a texture compression algorithm to obtain a compressed texture model; and determining a final three-dimensional model from the multi-level model based on the performance indicators of the compressed texture model, loading the final three-dimensional model into the memory of the display platform for real-time rendering and display. The present invention solves the problem of large data volumes and low rendering efficiency in scenic area three-dimensional models.

Description

Technical Field

[0001] The present invention belongs to the technical field of digital models, and in particular relates to a method for generating digital models for display in tourist attractions. Background Art

[0002] The process of generating digital models of tourist attractions presents a technical contradiction: the balance between the model's level of detail and display performance. On the one hand, to enhance the model's realism and immersion, high-precision 3D modeling and texture mapping are required for the scenic area's terrain, buildings, vegetation, and other features. This requires the model to have a large number of geometric faces and high-resolution textures. On the other hand, overly detailed models can lead to reduced rendering performance, affecting the smoothness of the display. Especially when displaying on mobile devices or web pages, due to the limitations of device performance and network transmission speeds, overly large model files can lead to slow loading and lag, seriously affecting the user experience.

[0003] Therefore, how to ensure the precision of the model while taking into account display performance has become a key technical challenge. This requires reasonable control and balance of the complexity of the model according to the specific application scenario and target platform during the modeling and optimization process. By reducing the number of faces, simplifying the texture, etc., the data volume and computational complexity of the model can be reduced as much as possible without significantly affecting the visual effect. At the same time, it is also necessary to use various performance optimization technologies, such as multi-level details (LOD), frustum culling, texture compression, etc., to further improve rendering efficiency and loading speed. Only by finding the best balance between precision and performance can high-quality and high-performance digital scenic area displays be achieved. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention proposes a digital model generation method for tourist attraction display, which can achieve high-quality and high-performance digital scenic spot display.

[0005] The present invention proposes a method for generating a digital model for displaying a tourist attraction, comprising:

[0006] Acquire three-dimensional data of a tourist attraction, perform high-precision modeling processing on the three-dimensional data, and obtain an initial three-dimensional model;

[0007] Based on the initial three-dimensional model, using a surface reduction algorithm to reduce the geometric surfaces of the initial three-dimensional model to obtain a simplified three-dimensional model;

[0008] Setting multi-level model detail parameters according to preset display platform performance parameters, and generating multi-level models with different degrees of refinement based on the simplified three-dimensional model and the multi-level model detail parameters;

[0009] Using a texture compression algorithm to compress the texture data in the multi-level model to obtain a compressed texture model;

[0010] According to the performance index of the compressed texture model, a final three-dimensional model is determined from the multi-level model, and the final three-dimensional model is loaded into the memory of the display platform for real-time rendering and display.

[0011] Optionally, the three-dimensional data includes: terrain, buildings and vegetation of the scenic area.

[0012] Optionally, performing high-precision modeling processing on the three-dimensional data to obtain an initial three-dimensional model includes:

[0013] Preprocessing the acquired three-dimensional data to obtain preprocessed three-dimensional data;

[0014] Based on the pre-processed 3D data of the scenic area terrain, a surface reconstruction algorithm, such as Poisson reconstruction and spherical harmonic function, is used to generate a high-precision 3D grid model of the scenic area terrain;

[0015] Based on the pre-processed 3D data of scenic area buildings, by extracting characteristic lines and characteristic surfaces, combined with the pre-defined building model library, the model matching and parametric modeling methods are used to achieve high-precision 3D reconstruction of the buildings;

[0016] Based on the pre-processed 3D vegetation data of the scenic area and the density and distribution characteristics of the point cloud data, a rule-based tree modeling algorithm is used to generate a realistic 3D vegetation model;

[0017] The reconstructed three-dimensional models of terrain, buildings and vegetation are integrated to obtain an initial three-dimensional model.

[0018] Optionally, obtaining a simplified 3D model includes:

[0019] Set the simplification threshold, control the simplification degree, and obtain the simplified 3D model;

[0020] The simplified three-dimensional model is evaluated to determine whether it meets the application requirements. If not, the simplification threshold is adjusted and the model is simplified again. If it meets the application requirements, the simplified three-dimensional model is obtained.

[0021] Optionally, generating multi-level models with different levels of refinement includes:

[0022] Obtain the performance parameters of the display platform, including CPU frequency, memory capacity, and GPU model;

[0023] determining a detail parameter threshold of a multi-level model according to the performance parameter;

[0024] Extracting the simplified three-dimensional model skeleton wireframe, and reducing the number of nodes and edges in the skeleton wireframe through a wireframe simplification algorithm to obtain a low-detail skeleton model;

[0025] A low-detail skeleton model is applied to the mesh model, and animation control of the model at different levels of detail is achieved through a skinning algorithm to obtain the multi-level model.

[0026] Optionally, compressing the texture data in the multi-level model using a texture compression algorithm to obtain a compressed texture model includes:

[0027] Acquire a multi-level model to be compressed, and extract texture data of the multi-level model to be compressed;

[0028] According to the characteristics of the texture data, the texture data is compressed using a texture compression algorithm;

[0029] Reintegrate the compressed texture data into the multi-level model to obtain an initial compressed texture model;

[0030] It is determined whether the texture data volume of the initial compressed texture model is less than the texture data volume of the original multi-level model version, and the compressed texture model is obtained according to the determination result.

[0031] Optionally, determining whether the amount of texture data of the initial compressed texture model is less than the amount of texture data of the original multi-level model version, and obtaining the compressed texture model according to the determination result includes:

[0032] If the amount of texture data of the initial compressed texture model is less than the amount of texture data of the original multi-level model version, the compressed texture model is output;

[0033] If the amount of texture data of the compressed texture model is not less than that of the original multi-level model version, the parameters of the texture compression algorithm are adjusted.

[0034] Optionally, determining a final three-dimensional model from the multi-level model according to a performance indicator of the compressed texture model includes:

[0035] Obtaining the geometric complexity and texture quality-related attribute data of the compressed texture model as a basis for evaluating the model performance;

[0036] Obtain performance indicator data of the display platform, including CPU, GPU performance, and memory capacity, to evaluate the platform's rendering capabilities;

[0037] Based on the performance evaluation criteria of the compressed texture model and the performance indicator data of the display platform, a decision tree algorithm is used to screen candidate models that meet rendering requirements from multiple levels of model versions;

[0038] Through the support vector machine algorithm, the geometric complexity, texture quality, and rendering efficiency of the candidate models are comprehensively evaluated to obtain a performance score;

[0039] The candidate models are sorted according to their performance scores, and the model with the highest score is selected as the final 3D model.

[0040] Compared with the prior art, the present invention has the following advantages and technical effects:

[0041] The present invention performs high-precision modeling on the three-dimensional data of the scenic area's terrain, buildings, and vegetation. After obtaining the initial three-dimensional model, the geometric surfaces are simplified through a face reduction algorithm, and the textures are downsampled and compressed to reduce the amount of model data. At the same time, multi-level model details are set according to the performance parameters of the display platform to generate multi-level model versions with different degrees of fineness. Finally, based on the geometric complexity, texture quality, and platform performance indicators, the final version is selected from the multi-level model and loaded into the display platform memory. The present invention solves the problems of large data volume and low rendering efficiency of the scenic area's three-dimensional model, realizes multi-level simplification and optimization of the model, improves real-time rendering and display effects, and enables the scenic area's three-dimensional model to be smoothly displayed on platforms with different performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The accompanying drawings, which constitute part of this application, are intended to provide a further understanding of this application. The exemplary embodiments and descriptions of this application are intended to explain this application and do not constitute an improper limitation on this application. In the accompanying drawings:

[0043] Figure 1 This is a flow chart of a method for generating a digital model for displaying a tourist attraction according to an embodiment of the present invention. DETAILED DESCRIPTION

[0044] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0045] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0046] The following is an explanation of the professional terms of this embodiment:

[0047] A 3D digital model is a digital representation of a surface or object created in computer graphics. It represents a physical entity in three-dimensional space. Constructed using 3D software, it captures the object's size, shape, and texture. It is a crucial part of the design process for many creative professions, including engineers, designers, architects, and game designers.

[0048] Definition and purpose of 3D digital model:

[0049] A 3D digital model is an object or scene in a 3D space represented by a computer data structure. Common 3D data models include the following:

[0050] Polygonal meshes: Made up of vertices, edges, and faces, they are commonly used in computer graphics and game development.

[0051] Voxel model: discretizes three-dimensional space into a regular voxel grid, which is used in fields such as medical imaging and geological data.

[0052] Parametric surfaces: defined by mathematical functions or parametric equations, commonly used in CAD and architectural design.

[0053] Implicit surface: defined by implicit equations, suitable for representing complex mathematical shapes.

[0054] Implementation of 3D digital model:

[0055] Parametric modeling: Using tools like CAD to define a set of instructions and rules to create and control the shape, size, and other characteristics of a model. When the parameters change, the model automatically updates.

[0056] Polygonal modeling: Create the required 3D model by adding or removing points, edges, faces, or adjusting their positions. It is suitable for film, television, animation, games and other fields.

[0057] Surface modeling: specifically used to create curved objects, suitable for processing complex surfaces such as digital products, cars, airplanes, etc.

[0058] Application areas of 3D digital models:

[0059] Three-dimensional digital models are widely used in many fields:

[0060] Architecture: used for architectural design, urban planning, etc.

[0061] Game Development: Used for the creation of characters, scenes, and items.

[0062] Film and television animation: used for special effects production and character modeling.

[0063] Medicine: used for medical image processing and virtual surgery simulation.

[0064] Geology: For geological data analysis and simulation.

[0065] This embodiment proposes a method for generating a digital model for displaying tourist attractions. Figure 1 As shown, the specific steps include:

[0066] Obtain 3D data of tourist attractions, perform high-precision modeling on the 3D data, and obtain an initial 3D model;

[0067] Based on the initial 3D model, a surface reduction algorithm is used to reduce the geometric surfaces of the initial 3D model to obtain a simplified 3D model;

[0068] According to the preset display platform performance parameters, the multi-level model detail parameters are set, and based on the simplified three-dimensional model and the multi-level model detail parameters, multi-level models with different degrees of precision are generated;

[0069] Using a texture compression algorithm to compress texture data in a multi-level model to obtain a compressed texture model;

[0070] According to the performance indicators of the compressed texture model, the final three-dimensional model is determined from the multi-level model, and the final three-dimensional model is loaded into the memory of the display platform for real-time rendering and display.

[0071] Furthermore, the three-dimensional data includes: terrain, buildings and vegetation of the scenic area.

[0072] Furthermore, high-precision modeling processing is performed on the three-dimensional data to obtain an initial three-dimensional model, including:

[0073] Preprocessing the acquired three-dimensional data to obtain preprocessed three-dimensional data;

[0074] Based on the pre-processed 3D data of the scenic area terrain, a surface reconstruction algorithm, such as Poisson reconstruction and spherical harmonic function, is used to generate a high-precision 3D grid model of the scenic area terrain.

[0075] Based on the pre-processed 3D data of scenic area buildings, by extracting characteristic lines and characteristic surfaces, combined with the pre-defined building model library, the model matching and parametric modeling methods are used to achieve high-precision 3D reconstruction of the buildings;

[0076] Based on the pre-processed 3D vegetation data of the scenic area and the density and distribution characteristics of the point cloud data, a rule-based tree modeling algorithm is used to generate a realistic 3D vegetation model;

[0077] The reconstructed three-dimensional models of terrain, buildings and vegetation are integrated to obtain an initial three-dimensional model.

[0078] Specifically, the construction of a scenic area's 3D model begins with a variety of data acquisition technologies. Laser scanning technology can quickly acquire high-precision point cloud data. For example, a terrestrial laser scanner can scan a scenic area in all directions, capturing details such as topographic relief and building outlines. Photogrammetry, meanwhile, uses drones to capture high-resolution aerial images, which not only provide surface texture information but also complement the blind spots of laser scanning. Preprocessing of raw data is crucial for subsequent modeling. Denoising can filter out outliers, such as using statistical outlier detection methods to eliminate points that do not conform to the terrain's characteristics. Data fusion unifies data from different sources into a common coordinate system, such as using the ICP algorithm to register laser point clouds with image point clouds, improving data consistency and integrity. Terrain modeling is the foundation of scenic area 3D reconstruction. The Poisson reconstruction algorithm effectively processes irregularly distributed point cloud data, generating a continuous and smooth surface model by solving the Poisson equation. For complex terrain, spherical harmonic functions can be combined to optimize local details to better represent features such as steep hills or ravines. 3D reconstruction of buildings requires a combination of feature extraction and model matching. First, the RANSAC algorithm is used to extract features such as building planes and edges, which are then matched against a predefined library of architectural models. For example, for traditional Chinese architecture, parametric models of typical components like roofs and eaves can be pre-built, allowing for precise fit to the actual building by adjusting the parameters. Vegetation modeling, on the other hand, requires consideration of its natural growth characteristics. The L-System algorithm simulates the branching structure of trees, defining a series of growth rules to generate realistic trunks and branches. For large areas of vegetation within a scenic area, a random tree model can be employed. Based on the density distribution of point cloud data, tree models of varying species and sizes are randomly generated at appropriate locations to reproduce the overall distribution of vegetation. Model integration is a key step in achieving overall consistency within the scenic area. Mesh simplification algorithms, such as edge collapse, can reduce geometric complexity while maintaining the visual quality of the model. Mesh optimization techniques, such as Laplacian smoothing, can eliminate surface discontinuities in the model, improving overall visual quality. Finally, photorealistic rendering brings the scenic area model to life. Global illumination algorithms, such as path tracing, simulate the propagation of natural light, accurately calculating the effects of direct and indirect lighting. Combined with physically based material models, such as the BRDF function, the optical properties of different surface materials can be accurately restored. Furthermore, using procedural texture generation technology, richly detailed textures can be automatically generated for large surfaces, enhancing the realism and detail of the model.

[0079] Furthermore, obtaining the simplified three-dimensional model includes:

[0080] Set the simplification threshold, control the simplification degree, and obtain the simplified 3D model;

[0081] The simplified 3D model is evaluated to determine whether it meets the application requirements. If not, the simplification threshold is adjusted and the model is simplified again. If it meets the application requirements, the simplified 3D model is obtained.

[0082] Specifically, an initial 3D model is obtained, which contains a large number of geometric faces and a large amount of data. The initial 3D model is processed by a face reduction algorithm to reduce unnecessary geometric faces. During the face reduction process, a simplification threshold is set to control the degree of simplification to balance model accuracy and data volume. The simplified model is evaluated to determine whether it meets the application requirements. If not, the simplification threshold is adjusted and the model is simplified again. The simplified model is optimized, such as eliminating duplicate faces and repairing topological errors, to improve model quality. The simplified and optimized model is saved in a target format to obtain a simplified 3D model with smaller data volume and higher efficiency. The simplified 3D model is applied to target scenarios, such as 3D visualization and virtual simulation, to improve system performance and user experience.

[0083] Furthermore, generating multi-level models with different levels of refinement includes:

[0084] Obtain the performance parameters of the display platform, including CPU frequency, memory capacity, and GPU model;

[0085] Determine the detail parameter threshold of the multi-level model based on the performance parameter;

[0086] Extract and simplify the skeleton wireframe of the 3D model. Use the wireframe simplification algorithm to reduce the number of nodes and edges in the skeleton wireframe to obtain a low-detail skeleton model.

[0087] Apply a low-detail skeleton model to the mesh model, and use the skinning algorithm to achieve animation control of the model at different levels of detail to obtain a multi-level model.

[0088] Specifically, the performance parameters of the display platform, including CPU frequency, memory capacity, and GPU model, are obtained. The thresholds for detail parameters of the multi-level model are determined based on these performance parameters. The initial 3D model is simplified using an edge collapse algorithm. Based on the thresholds determined in the first step, a first-level simplified 3D model is obtained. A surface reconstruction algorithm is used to fuse adjacent triangles and remove some details from the simplified 3D model from the previous step, resulting in a second-level simplified 3D model. Based on the topological structure of the 3D model, a skeleton wireframe is extracted. A wireframe simplification algorithm is used to reduce the number of nodes and edges within the skeleton wireframe, resulting in a low-detail skeleton model. Texture baking is performed on the simplified model from the second step, transferring surface detail information to the texture map, reducing the number of facets while preserving visual detail. The skeleton model is applied to the simplified mesh model, and a skinning algorithm is used to animate the model at different levels of detail. The rendering performance of the generated multi-level model on the target platform is evaluated. If it meets the display requirements, the final multi-level model is output. Otherwise, the process returns to the second step, where the simplification parameters are adjusted and the model is regenerated.

[0089] Furthermore, a texture compression algorithm is used to compress the texture data in the multi-level model to obtain a compressed texture model including:

[0090] Obtaining a multi-level model to be compressed, and extracting texture data of the multi-level model to be compressed;

[0091] According to the characteristics of texture data, texture compression algorithm is used to compress texture data;

[0092] Reintegrate the compressed texture data into the multi-level model to obtain an initial compressed texture model;

[0093] It is determined whether the texture data volume of the initial compressed texture model is less than the texture data volume of the original multi-level model version, and the compressed texture model is obtained according to the determination result.

[0094] Specifically, texture compression is a key step in 3D model optimization, aiming to reduce texture data volume while maintaining visual quality. First, texture data is extracted from the multi-level model version, including color maps, normal maps, and specular maps. Texture data feature analysis involves factors such as resolution, color depth, and texture type. For example, a high-resolution architectural model might contain a diffuse map with a resolution of 4096x4096 and a normal map with a resolution of 2048x2048. Choosing an appropriate compression algorithm based on texture characteristics is crucial. For photorealistic textures, lossy compression algorithms such as JPEG are suitable, while for cartoon-style textures, lossless compression algorithms such as PNG are more suitable. In addition, specialized GPU texture compression formats such as DXT1 and DXT5 exist. For example, the skin texture of a game character model might be well-suited to DXT5 compression because it effectively preserves alpha channel information. The compression process requires a balance between compression rate and image quality. For example, DXT1 compression can compress 32-bit pixels to 4 bits, achieving a compression ratio of up to 8:1, but this may result in some color distortion. For areas requiring high quality, such as a character's face, a lower compression ratio can be used; for less noticeable areas, such as background textures, a higher compression ratio can be used. The compressed texture data needs to be reintegrated into the model. This process involves updating UV coordinates and material parameters. For example, if the original model used multiple 4K textures, these might become multiple 2K textures after compression. UV mapping needs to be adjusted accordingly to ensure the textures are properly applied. Judging the compression effectiveness is a key step in the compression process. For example, if the original model's texture data totals 100MB and is compressed down to 40MB, and visual inspection confirms that the visual quality is acceptable, then the compression is successful. If the compressed data size doesn't decrease significantly, or if the visual quality degrades significantly, the compression parameters should be adjusted and reprocessed. Parameter adjustment is an iterative process. For example, for JPEG compression, you can gradually reduce the quality parameter, starting at 90 and decreasing it by 5-10 points at a time until you find a balance. For DXT compression, you can experiment with different preprocessing options, such as whether or not to enable alpha premultiplication. The final output compressed texture model should strike a balance between file size and visual quality. A successful example might be: the original model texture data volume is 200MB, which is reduced to 50MB after compression. When viewed at 1080p resolution, the naked eye can hardly tell the quality loss. This optimization not only reduces storage space, but also improves model loading speed and rendering performance, especially on resource-constrained platforms such as mobile devices.

[0095] Further, determining whether the texture data volume of the initial compressed texture model is smaller than the texture data volume of the original multi-level model version, and obtaining the compressed texture model according to the determination result includes:

[0096] If the amount of texture data of the initial compressed texture model is less than the amount of texture data of the original multi-level model version, the compressed texture model is output;

[0097] If the amount of texture data of the compressed texture model is not less than that of the original multi-level model version, the parameters of the texture compression algorithm are adjusted.

[0098] Furthermore, determining the final three-dimensional model from the multi-level model based on the performance index of the compressed texture model includes:

[0099] Obtain relevant attribute data such as geometric complexity and texture quality of the compressed texture model as a basis for evaluating model performance;

[0100] Obtain performance indicator data of the display platform, including CPU, GPU performance, and memory capacity, to evaluate the platform's rendering capabilities;

[0101] Based on the performance evaluation criteria of the compressed texture model and the performance indicator data of the display platform, a decision tree algorithm is used to screen candidate models that meet rendering requirements from multiple levels of model versions;

[0102] Through the support vector machine algorithm, the geometric complexity, texture quality, rendering efficiency, etc. of the candidate model are comprehensively evaluated to obtain a performance score;

[0103] The candidate models are sorted according to their performance scores, and the model with the highest score is selected as the final 3D model.

[0104] Specifically, the geometric complexity, texture quality, and other relevant attribute data of the compressed texture model are obtained as a basis for evaluating model performance. Performance indicator data of the display platform, including CPU, GPU performance, and memory capacity, is obtained to evaluate the platform's rendering capabilities. Based on the compressed texture model attributes and the display platform's performance, a decision tree algorithm is used to screen candidate models that meet rendering requirements from multiple levels of model versions. A support vector machine algorithm is used to comprehensively evaluate the geometric complexity, texture quality, and rendering efficiency of the candidate models to obtain a performance score. The candidate models are sorted according to their performance scores, and the model with the highest score is selected as the final 3D model version. The final 3D model version data is loaded into the display platform's graphics memory, and vertex, texture, and other data are optimized to improve rendering efficiency. The display platform's graphics rendering interface is called, the 3D model data is passed in, the model is rendered in real time using the GPU hardware acceleration capabilities, and the rendering results are presented to the user, achieving high-quality, high-performance model display.

[0105] The above are merely preferred embodiments of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A method for generating a digital model for displaying a tourist attraction, characterized in that: include: Acquire three-dimensional data of a tourist attraction, perform high-precision modeling processing on the three-dimensional data, and obtain an initial three-dimensional model; Performing high-precision modeling processing on the three-dimensional data to obtain an initial three-dimensional model includes: Preprocessing the acquired three-dimensional data to obtain preprocessed three-dimensional data; Based on the pre-processed 3D data of the scenic area terrain, a surface reconstruction algorithm, including Poisson reconstruction and spherical harmonic function, is used to generate a high-precision 3D mesh model of the scenic area terrain; Based on the pre-processed 3D data of scenic area buildings, by extracting characteristic lines and characteristic surfaces, combined with the pre-defined building model library, the model matching and parametric modeling methods are used to achieve high-precision 3D reconstruction of the buildings; Based on the density and distribution characteristics of the pre-processed scenic area vegetation 3D data and point cloud data, a rule-based tree modeling algorithm is used to generate a realistic 3D vegetation model; Integrate the reconstructed three-dimensional models of terrain, buildings, and vegetation to obtain an initial three-dimensional model; Based on the initial three-dimensional model, using a surface reduction algorithm to reduce the geometric surfaces of the initial three-dimensional model to obtain a simplified three-dimensional model; Setting multi-level model detail parameters according to preset display platform performance parameters, and generating multi-level models with different degrees of refinement based on the simplified three-dimensional model and the multi-level model detail parameters; Using a texture compression algorithm to compress the texture data in the multi-level model to obtain a compressed texture model; The texture data in the multi-level model is compressed using a texture compression algorithm to obtain a compressed texture model, including: Acquire a multi-level model to be compressed, and extract texture data of the multi-level model to be compressed; According to the characteristics of the texture data, the texture data is compressed using a texture compression algorithm; Reintegrate the compressed texture data into the multi-level model to obtain an initial compressed texture model; Determine whether the texture data volume of the initial compressed texture model is less than the texture data volume of the original multi-level model version, and obtain the compressed texture model based on the determination result; Determining whether the texture data volume of the initial compressed texture model is less than the texture data volume of the original multi-level model version, and obtaining the compressed texture model according to the determination result includes: If the amount of texture data of the initial compressed texture model is less than the amount of texture data of the original multi-level model version, the compressed texture model is output; If the amount of texture data of the compressed texture model is not less than that of the original multi-level model version, adjusting the parameters of the texture compression algorithm; determining a final three-dimensional model from the multi-level model based on the performance indicators of the compressed texture model, and loading the final three-dimensional model into the memory of the display platform for real-time rendering and display; Obtaining a simplified 3D model includes: Set the simplification threshold, control the simplification degree, and obtain the simplified 3D model; Evaluate the simplified three-dimensional model to determine whether it meets the application requirements. If not, adjust the simplification threshold and simplify again. If it meets the application requirements, obtain the simplified three-dimensional model. Generating multi-level models with different levels of sophistication involves: Obtain the performance parameters of the display platform, including CPU frequency, memory capacity, and GPU model; determining a detail parameter threshold of a multi-level model according to the performance parameter; Extracting the simplified three-dimensional model skeleton wireframe, and reducing the number of nodes and edges in the skeleton wireframe through a wireframe simplification algorithm to obtain a low-detail skeleton model; Applying a low-detail skeleton model to a mesh model, and implementing animation control of the model at different levels of detail through a skinning algorithm to obtain the multi-level model; Determining a final three-dimensional model from the multi-level model according to the performance index of the compressed texture model includes: Obtaining the geometric complexity and texture quality-related attribute data of the compressed texture model as a basis for evaluating the model performance; Obtain performance indicator data of the display platform, including CPU, GPU performance, and memory capacity, to evaluate the platform's rendering capabilities; Based on the performance evaluation criteria of the compressed texture model and the performance indicator data of the display platform, a decision tree algorithm is used to screen candidate models that meet rendering requirements from multiple levels of model versions; Through the support vector machine algorithm, the geometric complexity, texture quality, and rendering efficiency of the candidate models are comprehensively evaluated to obtain a performance score; The candidate models are sorted according to their performance scores, and the model with the highest score is selected as the final 3D model.

2. The method for generating a digital model for displaying a tourist attraction according to claim 1, characterized in that: The three-dimensional data includes: terrain, buildings and vegetation of the scenic area.

Images