Digital model generation method for tourist attraction display

Abstract

The invention discloses a digital model generation method for tourist attraction display, and the method comprises the steps: obtaining three-dimensional data of a tourist attraction, carrying out the high-precision modeling processing of the three-dimensional data, and obtaining an initial three-dimensional model; based on the initial three-dimensional model, reducing geometric surfaces of the initial three-dimensional model by adopting a surface reduction algorithm 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 finesses based on the simplified three-dimensional model and the multi-level model detail parameters; performing compression processing on texture data in the multi-stage model by adopting a texture compression algorithm to obtain a compressed texture model; and determining a final three-dimensional model from the multi-level model according to the performance index of the compressed texture model, and loading the final three-dimensional model into a memory of a display platform for real-time rendering and display. The problems of large data volume and low rendering efficiency of the three-dimensional model of the scenic area are solved.

Description

Technical Field

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

[0002] There is a technical contradiction in the process of generating digital models of tourist attractions: the balance between the refinement of the model and the performance of the display. On the one hand, in order to improve the realism and immersion of the model, high-precision 3D modeling and texture mapping of the terrain, buildings, vegetation, etc. of the scenic area are required, which requires the model to have a large number of geometric faces and high-resolution textures. On the other hand, overly refined models will lead to a decrease in rendering performance and affect the smoothness of the display. Especially when displaying on mobile devices or web pages, due to the limitations of device performance and network transmission speed, overly large model files will lead to slow loading, jamming and other problems, seriously affecting the user experience.

[0003] Therefore, how to ensure the precision of the model while taking into account the 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 scenarios and target platforms during the modeling and optimization process. By reducing the number of faces and simplifying the texture, 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 techniques, 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 spot display 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 scenic spot 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 acquire an initial three-dimensional model;

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

[0008] According to preset display platform performance parameters, setting multi-level model detail 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;

[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 three-dimensional 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 three-dimensional 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] According to the pre-processed three-dimensional vegetation data of the scenic area, a rule-based tree modeling algorithm is used to generate a realistic three-dimensional vegetation model according to the density and distribution characteristics of the point cloud data;

[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 sophistication 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 detail levels is achieved through a skinning algorithm to obtain the multi-level model.

[0026] Optionally, using a texture compression algorithm to compress the texture data in the multi-level model 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, a texture compression algorithm is used to compress the texture data;

[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, then 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 the final three-dimensional model from the multi-level model according to the performance index 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 basis of the evaluation model of the compressed texture model and the performance indicator data of the display platform, a decision tree algorithm is used to screen out candidate models that meet the rendering requirements from the multi-level 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 surface is simplified by a surface reduction algorithm, and the texture is 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 precision. 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 three-dimensional model, realizes multi-level simplification and optimization of the model, improves real-time rendering and display effects, and enables the three-dimensional model of the scenic area to be smoothly displayed on platforms with different performances. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0043] Figure 1 The present invention is a flowchart 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 in the embodiments of the present 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 professional terms of this embodiment are explained below:

[0047] A 3D digital model is a digital image created in computer graphics for a surface or object, representing the form of a physical entity in three-dimensional space. It is built using 3D software and can reflect the size, shape and texture of an object. It is an important part of the design process for many creative professions (such as engineers, designers, architects, game designers, etc.).

[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 mesh: Made up of vertices, edges, and faces, 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 surface: 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 such as CAD to define a set of instructions and rules to create and control the shape, size, and other features 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: specially 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] 3D digital models are widely used in many fields:

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

[0061] Game Development: 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 a tourist attraction. Figure 1 As shown, the specific steps include:

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

[0067] Based on the initial three-dimensional model, a surface reduction algorithm is used to reduce the geometric surfaces of the initial three-dimensional model to obtain a simplified three-dimensional 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, the multi-level model with different degrees of refinement is 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 index 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 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 three-dimensional 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 three-dimensional 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] According to the pre-processed three-dimensional vegetation data of the scenic area, a rule-based tree modeling algorithm is used to generate a realistic three-dimensional vegetation model according to the density and distribution characteristics of the point cloud data;

[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 3D model of a scenic spot begins with a variety of data acquisition technologies. Laser scanning technology can quickly obtain high-precision point cloud data. For example, using a ground laser scanner to perform a full-scale scan of a scenic spot can capture detailed information such as terrain undulations and building outlines. At the same time, photogrammetry technology obtains high-resolution images through drone aerial photography, which can not only provide surface texture information, but also supplement the blind area data of laser scanning. The preprocessing of raw data is crucial for subsequent modeling. Denoising can filter out abnormal points, such as using statistical outlier detection methods to remove points that do not conform to the characteristics of the ground object. Data fusion unifies data from different sources into the same coordinate system, such as using the ICP algorithm to align laser point clouds and image point clouds to improve data consistency and integrity. Terrain modeling is the basis for 3D reconstruction of scenic spots. The Poisson reconstruction algorithm can effectively process irregularly distributed point cloud data and generate a continuous and smooth surface model by solving the Poisson equation. For complex terrain, local detail optimization can be combined with spherical harmonic functions to better represent features such as steep mountains or gullies. The 3D reconstruction of buildings requires a combination of feature extraction and model matching. First, the plane, edge and other features of the building are extracted by the RANSAC algorithm, and then matched with the predefined building model library. For example, for traditional Chinese buildings, parametric models of typical components such as roofs and eaves can be pre-established, and accurate fitting of actual buildings can be achieved by adjusting parameters. Vegetation modeling needs to take into account its natural growth characteristics. The L-System algorithm can simulate the branch structure of trees and generate realistic trunks and branches by defining a series of growth rules. For large areas of vegetation in scenic spots, a random tree model can be used to randomly generate tree models of different types and sizes at appropriate locations according to the density distribution of point cloud data to restore the overall distribution characteristics of vegetation. Model integration is a key step in achieving overall consistency of the scenic spot. Through mesh simplification algorithms, such as edge collapse method, the geometric complexity can be reduced while maintaining the visual effect of the model. At the same time, using mesh optimization techniques, such as Laplacian smoothing, the discontinuity of the model surface can be eliminated to improve the overall visual quality. Finally, realistic rendering injects vitality into the scenic area model. The propagation of natural light is simulated by the global illumination algorithm, such as path tracing technology, which can accurately calculate the effects of direct and indirect lighting. Combined with physically based material models, such as BRDF functions, the optical properties of different surface materials can be accurately restored. In addition, using procedural texture generation technology, detailed textures can be automatically generated for large areas of the surface, improving the realism and detail performance 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; a face reduction algorithm is used to process the initial 3D model to reduce unnecessary geometric faces; in 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 application requirements, and if not, the simplification threshold is adjusted and 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, virtual simulation, etc., to improve system performance and user experience.

[0083] Further, generating a multi-level model with different levels of sophistication 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 according to the performance parameter;

[0086] Extract and simplify the skeleton wireframe of the 3D model, and 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 are obtained, including CPU frequency, memory capacity, GPU model, etc., and the threshold of the detail parameters of the multi-level model is determined according to these performance parameters. For the initial three-dimensional model, the edge collapse algorithm is used to simplify the model, and the first-level simplified three-dimensional model is obtained according to the threshold determined in the first step. For the simplified three-dimensional model in the previous step, the surface reconstruction algorithm is used to fuse the adjacent triangles and delete some details to obtain the second-level simplified three-dimensional model. According to the topological structure of the three-dimensional model, the skeleton wireframe is extracted, and the number of nodes and edges in the skeleton wireframe is reduced by the wireframe simplification algorithm to obtain a low-detail skeleton model. The texture of the simplified model in the second step is baked to transfer the surface detail information of the model to the texture map, reducing the number of model faces while retaining the visual details. The skeleton model is applied to the simplified mesh model, and the skinning algorithm is used to realize the animation control of the model at different detail levels. The rendering performance of the generated multi-level model on the target platform is judged. If it meets the display requirements, the final multi-level model is output, otherwise it returns to the second step, adjusts the simplification parameters, and regenerates the model.

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

[0090] Acquire a multi-level model to be compressed, and extract 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 the amount of texture data while maintaining visual quality. First, texture data is extracted from the multi-level model version, which includes color maps, normal maps, and specular maps. The feature analysis of texture data involves factors such as resolution, color depth, and texture type. For example, for a high-precision building model, it may contain a diffuse map with a resolution of 4096x4096 and a normal map with a resolution of 2048x2048. It is crucial to choose a suitable compression algorithm based on texture characteristics. For photo-realistic textures, lossy compression algorithms such as JPEG can be used; while for cartoon-style textures, lossless compression such as PNG is more suitable. In addition, there are special GPU texture compression formats such as DXT1 and DXT5. For example, the skin texture of a game character model may be suitable for DXT5 compression because it can well preserve the alpha channel information. During the compression process, it is necessary to balance the compression rate and image quality. Taking DXT1 compression as an example, it can compress 32-bit pixels to 4 bits with a compression ratio of up to 8:1, but it may cause some color distortion. For areas that require high quality, such as character faces, a lower compression rate can be used; for less noticeable areas, such as background textures, a higher compression rate can be used. The compressed texture data needs to be reintegrated into the model. This process involves updating the UV coordinates and material parameters. For example, if the original model used multiple 4K textures, it may become multiple 2K textures after compression. At this time, the UV mapping needs to be adjusted accordingly to ensure that the textures are applied correctly. Judging the compression effect is a key step in the compression process. Assume that the total amount of texture data of the original model is 100MB, and it is reduced to 40MB after compression. At the same time, visual inspection confirms that the visual quality is acceptable, then this compression is successful. If the amount of data after compression is not significantly reduced, or the visual quality is significantly reduced, it is necessary to adjust the compression parameters and reprocess. Parameter adjustment is an iterative process. For example, for JPEG compression, you can gradually reduce the quality parameter, starting from 90, and reduce it by 5-10 each time until you find a balance. For DXT compression, you can try different preprocessing options, such as whether to enable alpha premultiplication. The final output compressed texture model should achieve 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, and the quality loss is almost indistinguishable to the naked eye when viewed at 1080p resolution. 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, 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. According to the determination result, obtaining the compressed texture model 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, then 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] Further, according to the performance index of the compressed texture model, determining the final three-dimensional model from the multi-level 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 basis of the evaluation model of the compressed texture model and the performance indicator data of the display platform, a decision tree algorithm is used to screen out candidate models that meet the rendering requirements from the multi-level model versions;

[0102] Through the support vector machine algorithm, the geometric complexity, texture quality, rendering efficiency, etc. of the candidate models 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 related attribute data of the compressed texture model are obtained as the basis for evaluating the model performance. The performance indicator data of the display platform, including CPU, GPU performance, memory capacity, etc., are obtained to evaluate the rendering capability of the platform. According to the compressed texture model attributes and the display platform performance, the decision tree algorithm is used to screen out candidate models that meet the rendering requirements from the multi-level model version. The geometric complexity, texture quality, rendering efficiency, etc. of the candidate models are comprehensively evaluated by the support vector machine algorithm to obtain the performance score. The candidate models are sorted according to the performance score, 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 graphics memory of the display platform, and the vertex, texture and other data are optimized to improve the rendering efficiency. The graphics rendering interface of the display platform is called, the 3D model data is passed in, the model is rendered in real time using the GPU hardware acceleration capability, and the rendering result is presented to the user to achieve high-quality and high-performance model display.

[0105] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope 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 acquire an initial three-dimensional model; Based on the initial three-dimensional model, using a surface reduction algorithm to reduce geometric surfaces of the initial three-dimensional model to obtain a simplified three-dimensional model; According to preset display platform performance parameters, setting multi-level model detail 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; Using a texture compression algorithm to compress the texture data in the multi-level model to obtain a compressed texture model; 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.

2. A 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.

3. A method for generating a digital model for displaying a tourist attraction according to claim 2, characterized in that: 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 three-dimensional 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 three-dimensional grid 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; According to 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; The reconstructed three-dimensional models of terrain, buildings and vegetation are integrated to obtain an initial three-dimensional model.

4. A method for generating a digital model for displaying a tourist attraction according to claim 1, characterized in that: Obtaining a simplified 3D model includes: Set the simplification threshold, control the simplification degree, and obtain the simplified 3D model; 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.

5. The method for generating a digital model for displaying a tourist attraction according to claim 1, characterized in that: 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; A low-detail skeleton model is applied to the mesh model, and animation control of the model at different detail levels is achieved through a skinning algorithm to obtain the multi-level model.

6. A method for generating a digital model for displaying a tourist attraction according to claim 1, characterized in that: 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, a texture compression algorithm is used to compress the texture data; Reintegrate the compressed texture data into the multi-level model to obtain an initial compressed texture model; 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.

7. A method for generating a digital model for displaying a tourist attraction according to claim 6, characterized in that: 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, then 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, the parameters of the texture compression algorithm are adjusted.

8. The method for generating a digital model for displaying a tourist attraction according to claim 1, characterized in that: 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 basis of the evaluation model of the compressed texture model and the performance indicator data of the display platform, a decision tree algorithm is used to screen out candidate models that meet the rendering requirements from the multi-level 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.

Images