A grid simplification management method and system for multidimensional models
Through data analysis and visual analysis of multi-dimensional models, combined with neural network algorithms to perform meshing and hierarchical feature mapping, dynamically adjusting simplified parameters, solving the problems of low accuracy and efficiency in the simplified management of traditional multi-dimensional models, achieving high efficiency, Accurate grid simplified management.
Patent Information
- Application Number
- CN202411232126.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-04
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-09-04
AI Technical Summary
The grid simplification management method of traditional multi-dimensional models has problems such as low simplification accuracy or heavy tasks and low efficiency, which cannot meet user needs.
By obtaining the data analysis of the initial multidimensional model, geometric visual distribution data analysis and mapping processing are performed, combining self-organized mapping neural network and convolutional neural network algorithm, grid division and hierarchical feature mapping are performed, and simplified parameters are dynamically adjusted to achieve adaptive LOD Technology, optimization and simplification process.
It improves the grid simplified management accuracy and computing efficiency of multi-dimensional models, ensuring that the best visual experience and performance are provided in different visual ranges and application scenarios to meet user needs.
Smart Images

Figure CN118734099B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of neural network models, and in particular to a grid simplification management method and system for a multi-dimensional model. Background Art
[0002] Multidimensional models (such as 3D and 4D models) have been widely used in fields such as computer-aided engineering (CAE), computer graphics, geographic information systems (GIS), and scientific computing. With the advancement of data acquisition and generation technology, the resolution and complexity of these models are also increasing. High-resolution models can provide more details and higher accuracy, but they also bring a lot of data processing and storage problems. Especially in real-time applications, processing such a large amount of data often requires a lot of computing resources and time. Mesh simplification technology reduces the complexity and data volume of the model by reducing the number of vertices and patches in the model, thereby improving processing efficiency. However, the traditional mesh simplification management method of multidimensional models uniformly simplifies the mesh of the entire multidimensional model, which has the problem of too low simplification accuracy, or simplifies the meshes of the multidimensional model one by one, the mesh simplification task is too heavy and the simplification calculation efficiency is low, which cannot meet user needs. Summary of the invention
[0003] Based on this, the present invention provides a grid simplification management method and system for a multi-dimensional model to solve at least one of the above technical problems.
[0004] To achieve the above object, a grid simplification management method for a multidimensional model includes the following steps:
[0005] Step S1: obtaining an initial multidimensional model; performing multidimensional model data analysis processing based on the initial multidimensional model to generate multidimensional model analysis data, wherein the multidimensional model analysis data includes multidimensional model point cloud data, multidimensional model texture data and multidimensional model color data;
[0006] Step S2: performing geometric visual distribution data analysis based on the multidimensional model analysis data to generate geometric visual distribution data; performing geometric visual mapping processing based on the geometric visual distribution data to generate geometric visual data;
[0007] Step S3: performing mesh division management on the geometric visual data to generate mesh geometric visual data; performing mesh geometric visual feature weighting analysis processing on the mesh geometric visual data to generate weighted mesh geometric visual feature data;
[0008] Step S4: performing cluster feature analysis on the weighted grid geometric visual feature data according to a preset self-organizing map neural network clustering algorithm to generate clustered grid geometric visual feature data; performing clustered grid level feature mapping on the clustered grid geometric visual feature data to generate clustered grid level feature mapping data; performing grid level simplification matrix analysis based on the clustered grid level feature mapping data to generate a grid level simplification matrix;
[0009] Step S5: obtaining user historical interaction behavior data; performing a view distance simplification degree threshold analysis based on the user historical interaction behavior data to generate view distance simplification degree threshold data; performing adaptive LOD grid level simplification parameter prediction on the grid level simplification matrix and the view distance simplification degree threshold data based on a preset convolutional neural network algorithm to generate adaptive LOD grid level simplification parameters;
[0010] Step S6: Adaptively simplify and render the mesh geometric visual data using adaptive LOD mesh level simplification parameters to generate adjusted simplified mesh geometric visual data; and perform mesh simplification management operations on the multidimensional model based on the adjusted simplified mesh geometric visual data.
[0011] The present invention extracts point cloud data, texture data and color data from the multidimensional model by performing data analysis on the initial multidimensional model, so that the subsequent processing steps can accurately and comprehensively describe the geometric and visual features of the model, ensuring that key information is accurately captured and retained at the beginning of the simplification process. The geometric visual distribution data analysis and mapping processing can identify important geometric and visual features in the model, which not only helps to extract key information of the model, but also can maintain the visual consistency and important details of the model in the subsequent simplification process through the generation of geometric visual data, avoiding the loss of key geometric and visual information in the simplification process. By meshing and managing the geometric visual data, the complex multidimensional model is decomposed into multiple small grids that are easier to handle, thereby greatly improving the processing efficiency. At the same time, the weighted analysis of the mesh geometric visual features ensures that in the simplification process, important feature areas and details are retained. Through the weighted processing, the simplification process is made more targeted and accurate, ensuring that the simplified model retains important features while significantly reducing the amount of data. Cluster feature analysis can gather mesh data with similar features together to form clustered mesh geometric visual feature data, so as to more effectively manage and process complex multi-dimensional model data. Hierarchical feature mapping and simplified matrix analysis further organize and simplify these cluster data in a hierarchical structure, so that the complexity of the model can be flexibly adjusted under different accuracy requirements, improve processing efficiency and reduce unnecessary computing resource consumption, maintain the overall structure and key features of the model, and avoid the loss of important details. By introducing user historical interaction behavior data, the mesh simplification parameters can be dynamically adjusted according to the user's actual usage habits and needs. This adaptive LOD (Level of Detail) technology ensures that the most appropriate model details are presented at different viewing distances, so that users can get the best visual experience and performance in different scenarios and applications. Viewing distance simplification degree threshold analysis and adaptive parameter analysis enable the system to intelligently optimize the simplification process, reduce unnecessary detail loss, and improve computing efficiency and response speed. Adaptive simplified rendering adjustment enables the system to dynamically adjust and optimize the model according to real-time application requirements and user interaction, ensuring that the best performance and visual effects can be provided in various usage scenarios, which not only improves the efficiency of model use, but also greatly improves user experience. The execution of simplified management operations ensures the efficiency and practicality of multidimensional models in various applications, so that complex multidimensional data can be significantly simplified while maintaining important features. Therefore, the grid simplification management method of the multidimensional model of the present invention performs grid geometric visual division on the multidimensional model, and simplifies it considering the hierarchical strength of the grid that needs to be simplified, so that the grid simplification management accuracy and computational efficiency of the multidimensional model are guaranteed to meet user needs.
[0012] Preferably, step S1 comprises the following steps:
[0013] Step S11: obtaining an initial multidimensional model;
[0014] Step S12: performing multidimensional model denoising processing on the initial multidimensional model according to a preset denoising filter decision to generate a denoised multidimensional model;
[0015] Step S13: performing time stamp correction and synchronization processing on the denoised multidimensional model to generate a standard multidimensional model;
[0016] Step S14: performing multidimensional model data analysis processing according to the standard multidimensional model to generate multidimensional model analysis data.
[0017] The present invention ensures that the system can accurately obtain the initial multidimensional model data to be processed, provides a data foundation, and ensures the integrity and reliability of the data. Noise reduction processing effectively removes noise and interference in the multidimensional model and improves the quality of the data. By applying the preset noise reduction filter decision, it can effectively filter different types of noise, making the denoised model more accurate and clear, laying a good foundation for subsequent data analysis and processing. The timestamp correction synchronization processing ensures the consistency and synchronization of the multidimensional model data in the time dimension, eliminates the data errors and deviations caused by time inconsistency, and the generated standard multidimensional model has higher time consistency and accuracy, making subsequent processing more stable and reliable. By performing data analysis processing on the standard multidimensional model, the point cloud data, texture data and color data in the model can be accurately extracted to form detailed multidimensional model analysis data, ensuring the comprehensive extraction and accurate description of all key information, providing detailed data support for subsequent simplification processing, and improving the quality and accuracy of the simplification effect.
[0018] Preferably, the geometric visual distribution data includes geometric visual confidence distribution data and color interpolation distribution data, and step S2 includes the following steps:
[0019] Step S21: performing geometric vision confidence distribution evaluation processing according to the multi-dimensional model point cloud data and the multi-dimensional model texture data to generate geometric vision confidence distribution data;
[0020] Step S22: performing geometric visual association analysis on the multidimensional model point cloud data and the multidimensional model texture data to generate geometric visual association data, and performing geometric visual likelihood function analysis on the geometric visual association data to obtain a geometric visual likelihood function;
[0021] Step S23: performing geometric vision confidence distribution optimization processing on the geometric vision confidence distribution data according to the geometric vision likelihood function to generate optimized geometric vision confidence distribution data;
[0022] Step S24: performing color interpolation distribution analysis on the multidimensional model color data according to the optimized geometric vision confidence distribution data to generate color interpolation distribution data;
[0023] Step S25: Perform geometric vision mapping processing according to the optimized geometric vision confidence distribution data and the color interpolation distribution data to generate geometric vision data.
[0024] The present invention performs geometric visual confidence distribution evaluation on point cloud data and texture data, identifies and marks important geometric feature areas in the model, helps to distinguish high confidence and low confidence areas, so that in the subsequent simplification process, the high confidence areas can be retained, ensuring that the key details of the model are not ignored. Geometric visual association analysis can reveal the intrinsic connection between point cloud data and texture data, generate geometric visual association data, and quantify this connection through geometric visual likelihood function analysis, further enhance the understanding and description of geometric features, and provide a solid foundation for confidence distribution optimization. The confidence distribution data is optimized using the geometric visual likelihood function, and the geometric visual confidence distribution is accurately adjusted and optimized, so that the high confidence area is more prominent and the low confidence area is further filtered. In the subsequent simplification process, the important features of the model can be more accurately retained, and the quality of the simplification effect can be improved. Color interpolation distribution analysis uses the optimized geometric visual confidence distribution data to perform fine interpolation processing on the color data, which can better reflect the actual color and texture characteristics of the model and ensure the accuracy and consistency of color and texture during the simplification process. The geometric vision mapping process combines the optimized geometric vision confidence distribution data and color interpolation distribution data to generate geometric vision data, ensuring that the geometric and visual features of the model are fully preserved and accurately described during the simplification process.
[0025] Preferably, step S3 comprises the following steps:
[0026] Step S31: performing geometric topological structure analysis on the geometric visual data to generate geometric topological structure data;
[0027] Step S32: designing a geometric visual mesh division management decision based on the geometric topological structure data, and using the geometric visual mesh division management decision to perform mesh division management on the geometric visual data to generate mesh geometric visual data;
[0028] Step S33: performing feature analysis processing according to the mesh geometric visual data to generate mesh geometric visual feature data;
[0029] Step S34: performing feature weighting processing on the mesh geometric visual feature data according to a preset mesh geometric visual feature weighting decision to generate weighted mesh geometric visual feature data.
[0030] The present invention performs geometric topological structure analysis on geometric visual data, comprehensively understands the topological relationship and structural features of the model, helps to identify key connections and shape features in the model, provides important reference data for subsequent mesh division, and ensures the integrity and rationality of the structure during the simplification process. Mesh division decisions are made based on geometric topological structure data, so that the divided meshes have higher computing efficiency and processing performance while retaining the key geometric features of the model. The generated mesh geometric visual data is more structured, which improves the accuracy and efficiency of the overall simplification process. Feature analysis processing identifies and extracts important geometric and visual features in the mesh, generates mesh geometric visual feature data, ensures that during the simplification process, simplified hierarchical division is performed based on important features, reduces unnecessary detail loss, and improves the expressiveness and accuracy of the simplified model. Feature weighting processing assigns different weights to different geometric visual features according to the specific application requirements of the model, ensures that during the simplification process, important feature areas are retained first, while secondary areas are appropriately simplified, thereby achieving a more reasonable simplification effect and higher model quality.
[0031] Preferably, step S33 includes the following steps:
[0032] Extracting mesh geometric visual complexity index data from the mesh geometric visual data, and performing mesh geometric visual complexity feature evaluation processing according to the mesh geometric visual complexity index data to generate mesh geometric visual complexity feature data;
[0033] Identify mesh vertex data in the mesh geometric visual data, and perform mesh geometric visual gradient feature analysis on the mesh geometric visual data according to the mesh vertex data to generate mesh geometric visual gradient feature data;
[0034] The mesh geometric visual complexity feature data and the mesh geometric visual gradient feature data are mesh geometric visual feature integrated to generate mesh geometric visual feature data.
[0035] The present invention extracts complexity index data from mesh geometric visual data, quantifies and evaluates the geometric complexity features of the model, helps identify areas with higher complexity in the model, enables subsequent simplification processing to be optimized for these complex areas, ensures that the geometric details and structural rationality of the overall model can be maintained during the simplification process, and improves the accuracy and expressiveness of the simplified model. Identify mesh vertex data and perform geometric gradient feature analysis to identify gradient change areas in the model, help locate important geometric change points and edges in the model, and enable these key features to be retained preferentially during the simplification process, avoid losing important geometric information during the simplification process, and improve the visual consistency and realism of the simplified model. Integrate complexity feature data and gradient feature data, comprehensively consider the geometric complexity and gradient change features of the model, and fully retain the important geometric and visual features of the model, achieve a more optimized mesh simplification effect, and improve the overall quality and expressiveness of the simplified model.
[0036] Preferably, step S4 comprises the following steps:
[0037] Step S41: performing clustering feature analysis on the weighted grid geometric visual feature data according to a preset self-organizing map neural network clustering algorithm to generate clustered grid geometric visual feature data;
[0038] Step S42: performing grid cluster label analysis according to the cluster grid geometric visual feature data to generate grid cluster label data;
[0039] Step S43: setting the grid division level according to the grid clustering label data to generate grid division level data;
[0040] Step S44: performing clustered grid level feature mapping processing on the clustered grid geometric visual feature data according to the grid division level data to generate clustered grid level feature mapping data;
[0041] Step S45: performing grid level error matrix analysis according to the grid level feature mapping data to generate a grid level error matrix;
[0042] Step S46: Perform grid level simplification matrix analysis based on the grid level error matrix to generate a grid level simplification matrix.
[0043] The present invention classifies mesh geometric visual feature data with similar features together through cluster feature analysis to form clustered mesh geometric visual feature data, which is helpful to discover and identify areas with similar attributes in the model, making subsequent processing more efficient, and can retain the overall features of these similar areas during the simplification process, thereby improving the consistency of the simplification effect. The mesh cluster label analysis assigns a label to each cluster, identifies the features and attributes of different clusters, and the generated mesh cluster label data helps to accurately distinguish and process different types of clusters during the simplification process, making the simplification process more targeted and accurate, and further improving the simplification quality and expressiveness of the model. The mesh division level is set according to the cluster label data, and the model is divided into multiple hierarchical structures. The generated mesh division hierarchical data ensures that the hierarchical structure can be optimized during the simplification process, thereby achieving more flexible and efficient mesh simplification. Through hierarchical feature mapping processing, the clustered mesh geometric visual feature data is mapped to different hierarchical structures, and clustered mesh hierarchical feature mapping data is generated, ensuring that the feature information of each level can be retained during the simplification process, so that the simplified model can maintain high accuracy and expressiveness at different levels. The grid-level error matrix analysis quantifies and evaluates the simplification errors at different levels. The generated grid-level error matrix helps control the error range during the simplification process, ensuring that the errors of the simplified model at each level are minimized, and improving the accuracy and reliability of the simplification process. The simplification strategy is optimized based on the data of the error matrix, and the grid-level simplification matrix is generated, ensuring that the simplification process can minimize the amount of data while controlling the error, achieve efficient grid simplification, and improve the processing efficiency and application performance of the model.
[0044] Preferably, step S5 comprises the following steps:
[0045] Step S51: Obtaining user historical interaction behavior data;
[0046] Step S52: performing interaction distance and interaction visual correlation analysis based on the user's historical interaction behavior data to generate historical interaction visual distance correlation data;
[0047] Step S53: performing interactive sight distance fuzzy kernel analysis according to historical interactive sight distance associated data to generate historical interactive sight distance fuzzy kernel data;
[0048] Step S54: evaluating the threshold value of the degree of simplification of the sight distance based on the historical interactive sight distance blur kernel data, and generating the threshold data of the degree of simplification of the sight distance;
[0049] Step S55: Based on a preset convolutional neural network algorithm, adaptive LOD grid level simplification parameter prediction is performed on the grid level simplification matrix and the view distance simplification degree threshold data to generate adaptive LOD grid level simplification parameters.
[0050] The present invention collects the user's historical interactive behavior data, understands the user's actual operating habits and needs when using the model, and optimizes the user's specific usage situation, so that the subsequent simplification process is more in line with the user's usage habits, improving the user experience and satisfaction. The interactive distance and interactive visual association analysis can reveal the user's operating behavior and visual focus at different viewing distances. The generated historical interactive viewing distance association data helps understand the user's needs in different usage scenarios, ensuring that the simplified model can meet the user's actual needs at different viewing distances, and improving the adaptability and flexibility of the model. Through the interactive viewing distance fuzzy kernel analysis, the fuzzy relationship between different distances and vision in the user's interactive behavior is understood, and smoother and continuous viewing distance association data is generated, which helps to reduce noise and outliers in the analysis, improve the accuracy and reliability of the analysis results, and set the user to observe the multidimensional model according to the viewing distance for simplification intensity analysis. The viewing distance simplification degree threshold evaluation can determine the simplification degree threshold suitable for different viewing distances according to the user's historical interactive behavior and viewing distance fuzzy kernel data, ensuring that at different viewing distances, the simplified model can not only meet the performance requirements, but also provide sufficient details and visual effects, and optimize the user experience. By combining the grid level simplification matrix and the viewing distance simplification degree threshold data, adaptive LOD (Level of Detail) grid level simplification parameter analysis is performed, and the simplification parameters are dynamically adjusted so that the model can provide the best visual effects and performance under different viewing distances and application scenarios. The generated adaptive LOD grid level simplification parameters help to achieve efficient simplification and intelligent optimization of the model, thereby improving overall processing efficiency and user satisfaction.
[0051] Preferably, step S54 includes the following steps:
[0052] The fuzzy kernel residual is calculated according to the historical interactive line-of-sight fuzzy kernel data to generate the historical line-of-sight fuzzy kernel residual data;
[0053] Perform fitting residual processing on the historical line-of-sight blur kernel residual data to generate historical line-of-sight blur kernel fitting residual data;
[0054] The sight distance simplification degree threshold is evaluated based on the historical sight distance blur kernel fitting residual data to generate the sight distance simplification degree threshold data.
[0055] The present invention performs residual calculations on historical interactive line-of-sight blur kernel data, quantifies and evaluates deviations and anomalies in user interaction behaviors, and the generated blur kernel residual data helps to identify blur kernel strength, making the analysis results more accurate and stable. The fitting residual processing further smoothes and optimizes the blur kernel residual data, and the generated fitting residual data can more accurately reflect the laws and trends of user interaction behaviors. Through the fitting residual processing, the random errors in the analysis are reduced, the accuracy and effect of the simplification process are improved, and the simplified model is more in line with the actual needs of users. Based on the fitting residual data, the line-of-sight simplification degree threshold evaluation is performed, and the simplification threshold at different line-of-sight distances is accurately determined, ensuring that at different line-of-sight distances, the simplified model can meet performance requirements and provide sufficient details and visual effects. The generated line-of-sight simplification degree threshold data makes the model simplification process more intelligent and efficient, and improves the user experience and model adaptability.
[0056] Preferably, step S6 comprises the following steps:
[0057] Step S61: performing mesh adaptive simplification rendering processing on the mesh geometric visual data by using adaptive LOD mesh level simplification parameters to generate simplified mesh geometric visual data;
[0058] Step S62: performing global visual smoothing adjustment processing on the simplified mesh geometric visual data to generate adjusted simplified mesh geometric visual data;
[0059] Step S63: Execute the mesh simplification management operation of the multi-dimensional model according to the adjusted simplified mesh geometric visual data.
[0060] The present invention can dynamically adjust the degree of simplification according to specific application requirements and user interaction behaviors through adaptive LOD grid level simplification parameters, ensuring the best simplification effect under different viewing distances and scenes. Through adaptive simplified rendering processing, the simplified grid geometric visual data generated can not only meet the performance requirements, but also maintain sufficient visual details and accuracy, thereby improving the actual use effect and user experience of the model. The global visual smoothing adjustment process can effectively remove the visual noise and discontinuity that may be generated during the simplification process, making the simplified grid geometric visual data smoother and more consistent. The generated adjusted simplified grid geometric visual data not only improves the visual quality of the model, but also reduces visual fatigue and discomfort, and enhances the user's viewing experience and the overall aesthetics of the model. The final grid simplification management operation is based on the adjusted simplified grid geometric visual data, ensuring the efficiency and accuracy of the simplification process. The simplified multidimensional model not only has high computational efficiency and performance, but also maintains high visual quality and detail performance, providing a reliable and high-quality simplified model for practical applications.
[0061] This specification provides a grid simplification management system for a multidimensional model, which is used to execute the grid simplification management method for a multidimensional model as described above. The grid simplification management system for a multidimensional model includes:
[0062] A multidimensional model analysis module is used to obtain an initial multidimensional model; perform multidimensional model data analysis based on the initial multidimensional model to generate multidimensional model analysis data, wherein the multidimensional model analysis data includes multidimensional model point cloud data, multidimensional model texture data and multidimensional model color data;
[0063] A geometric vision mapping module is used to analyze geometric vision distribution data according to multi-dimensional model parsing data to generate geometric vision distribution data; perform geometric vision mapping processing according to the geometric vision distribution data to generate geometric vision data;
[0064] The mesh geometric visual feature analysis module is used to perform mesh division management on the geometric visual data and generate mesh geometric visual data; perform mesh geometric visual feature weighting analysis and processing based on the mesh geometric visual data and generate weighted mesh geometric visual feature data;
[0065] A grid level simplification matrix analysis module is used to perform clustering feature analysis on weighted grid geometric visual feature data according to a preset self-organizing map neural network clustering algorithm to generate clustered grid geometric visual feature data; perform clustered grid level feature mapping on the clustered grid geometric visual feature data to generate clustered grid level feature mapping data; perform grid level simplification matrix analysis based on the clustered grid level feature mapping data to generate a grid level simplification matrix;
[0066] An adaptive LOD grid level simplification parameter prediction module is used to obtain historical user interaction behavior data; perform a view distance simplification degree threshold analysis based on the historical user interaction behavior data to generate view distance simplification degree threshold data; perform adaptive LOD grid level simplification parameter prediction on the grid level simplification matrix and the view distance simplification degree threshold data based on a preset convolutional neural network algorithm to generate adaptive LOD grid level simplification parameters;
[0067] The multi-dimensional model mesh simplification module is used to perform mesh adaptive simplification rendering adjustment on mesh geometric visual data through adaptive LOD mesh level simplification parameters to generate adjusted simplified mesh geometric visual data; and execute mesh simplification management operations of the multi-dimensional model based on the adjusted simplified mesh geometric visual data.
[0068] The beneficial effect of the present application is that the present invention ensures that the basic data of the simplified processing is complete and accurate from obtaining the initial multidimensional model to generating detailed multidimensional model analysis data (including point cloud data, texture data, and color data), and the noise reduction filter decision and timestamp correction synchronization processing improve the quality and consistency of the model data, providing a reliable data basis for subsequent steps. Through geometric visual distribution data analysis and geometric visual mapping processing, important geometric structures and visual details in the model are identified and retained, and the visual consistency and expressiveness of the simplified model are improved. Cluster feature analysis, hierarchical feature mapping, and simplified matrix analysis realize intelligent clustering and hierarchical simplification of structured grid geometric visual data, ensure flexible adjustment of model complexity under different accuracy requirements, improve processing efficiency and reduce computing resource consumption, and use hierarchical error matrix analysis and simplified matrix analysis to ensure the accuracy and reliability of the simplification process. Considering the user's historical interactive behavior data to dynamically adjust the viewing distance simplification degree threshold and adaptive LOD parameters, the simplified model can better meet the actual user needs, making the simplification process more intelligent and personalized. Adaptive simplified rendering processing combined with global visual smoothing adjustment ensures that the simplified model can provide the best visual effects and performance at different viewing distances and scenes. The simplified management operation ensures the efficiency, accuracy and visual quality of the final model, providing a reliable solution for practical applications. The mesh simplification management method of the multi-dimensional model can efficiently and accurately simplify the mesh while maintaining the important geometric visual features of each mesh. BRIEF DESCRIPTION OF THE DRAWINGS
[0069] Figure 1 A schematic diagram of the steps of a simplified grid management method for a multi-dimensional model of the present invention;
[0070] Figure 2 for Figure 1 Detailed implementation steps of step S4 in FIG.
[0071] Figure 3 for Figure 1 Detailed implementation steps of step S5;
[0072] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION
[0073] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.
[0074] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.
[0075] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.
[0076] To achieve this, please refer to Figures 1 to 3 The present invention provides a method for generating a clinical nursing form, comprising the following steps:
[0077] Based on this, the present invention provides a grid simplification management method and system for a multi-dimensional model to solve at least one of the above technical problems.
[0078] To achieve the above object, a grid simplification management method for a multidimensional model includes the following steps:
[0079] Step S1: obtaining an initial multidimensional model; performing multidimensional model data analysis processing based on the initial multidimensional model to generate multidimensional model analysis data, wherein the multidimensional model analysis data includes multidimensional model point cloud data, multidimensional model texture data and multidimensional model color data;
[0080] Step S2: performing geometric visual distribution data analysis based on the multidimensional model analysis data to generate geometric visual distribution data; performing geometric visual mapping processing based on the geometric visual distribution data to generate geometric visual data;
[0081] Step S3: performing mesh division management on the geometric visual data to generate mesh geometric visual data; performing mesh geometric visual feature weighting analysis processing on the mesh geometric visual data to generate weighted mesh geometric visual feature data;
[0082] Step S4: performing cluster feature analysis on the weighted grid geometric visual feature data according to a preset self-organizing map neural network clustering algorithm to generate clustered grid geometric visual feature data; performing clustered grid level feature mapping on the clustered grid geometric visual feature data to generate clustered grid level feature mapping data; performing grid level simplification matrix analysis based on the clustered grid level feature mapping data to generate a grid level simplification matrix;
[0083] Step S5: obtaining user historical interaction behavior data; performing a view distance simplification degree threshold analysis based on the user historical interaction behavior data to generate view distance simplification degree threshold data; performing adaptive LOD grid level simplification parameter prediction on the grid level simplification matrix and the view distance simplification degree threshold data based on a preset convolutional neural network algorithm to generate adaptive LOD grid level simplification parameters;
[0084] Step S6: Adaptively simplify and render the mesh geometric visual data using adaptive LOD mesh level simplification parameters to generate adjusted simplified mesh geometric visual data; and perform mesh simplification management operations on the multidimensional model based on the adjusted simplified mesh geometric visual data.
[0085] The present invention extracts point cloud data, texture data and color data from the multidimensional model by performing data analysis on the initial multidimensional model, so that the subsequent processing steps can accurately and comprehensively describe the geometric and visual features of the model, ensuring that key information is accurately captured and retained at the beginning of the simplification process. The geometric visual distribution data analysis and mapping processing can identify important geometric and visual features in the model, which not only helps to extract key information of the model, but also can maintain the visual consistency and important details of the model in the subsequent simplification process through the generation of geometric visual data, avoiding the loss of key geometric and visual information in the simplification process. By meshing and managing the geometric visual data, the complex multidimensional model is decomposed into multiple small grids that are easier to handle, thereby greatly improving the processing efficiency. At the same time, the weighted analysis of the mesh geometric visual features ensures that in the simplification process, important feature areas and details are retained. Through the weighted processing, the simplification process is made more targeted and accurate, ensuring that the simplified model retains important features while significantly reducing the amount of data. Cluster feature analysis can gather mesh data with similar features together to form clustered mesh geometric visual feature data, so as to more effectively manage and process complex multi-dimensional model data. Hierarchical feature mapping and simplified matrix analysis further organize and simplify these cluster data in a hierarchical structure, so that the complexity of the model can be flexibly adjusted under different accuracy requirements, improve processing efficiency and reduce unnecessary computing resource consumption, maintain the overall structure and key features of the model, and avoid the loss of important details. By introducing user historical interaction behavior data, the mesh simplification parameters can be dynamically adjusted according to the user's actual usage habits and needs. This adaptive LOD (Level of Detail) technology ensures that the most appropriate model details are presented at different viewing distances, so that users can get the best visual experience and performance in different scenarios and applications. Viewing distance simplification degree threshold analysis and adaptive parameter analysis enable the system to intelligently optimize the simplification process, reduce unnecessary detail loss, and improve computing efficiency and response speed. Adaptive simplified rendering adjustment enables the system to dynamically adjust and optimize the model according to real-time application requirements and user interaction, ensuring that the best performance and visual effects can be provided in various usage scenarios, which not only improves the efficiency of model use, but also greatly improves user experience. The execution of simplified management operations ensures the efficiency and practicality of multidimensional models in various applications, so that complex multidimensional data can be significantly simplified while maintaining important features. Therefore, the grid simplification management method of the multidimensional model of the present invention performs grid geometric visual division on the multidimensional model, and simplifies it considering the hierarchical strength of the grid that needs to be simplified, so that the grid simplification management accuracy and computational efficiency of the multidimensional model are guaranteed to meet user needs.
[0086] As an embodiment of the present invention, refer to Figure 1The above is a schematic flow chart of the steps of a grid simplification management method for a multidimensional model of the present invention. In this embodiment, the grid simplification management method for a multidimensional model includes the following steps:
[0087] Step S1: obtaining an initial multidimensional model; performing multidimensional model data analysis processing based on the initial multidimensional model to generate multidimensional model analysis data, wherein the multidimensional model analysis data includes multidimensional model point cloud data, multidimensional model texture data and multidimensional model color data;
[0088] In an embodiment of the present invention, an initial multidimensional model is obtained from a data source or an on-site scanning device to ensure the integrity and accuracy of the model data. The obtained initial multidimensional model is parsed to extract key data elements: point cloud data extraction: point cloud data is extracted from the multidimensional model, and these data include the position information of each point in the model; texture data extraction: texture information of the model surface is extracted, including line distribution, etc.; color data extraction: color characteristics of the model are analyzed and recorded, including color distribution and color depth.
[0089] Step S2: performing geometric visual distribution data analysis based on the multidimensional model analysis data to generate geometric visual distribution data; performing geometric visual mapping processing based on the geometric visual distribution data to generate geometric visual data;
[0090] In an embodiment of the present invention, statistical analysis and image processing techniques are used to perform geometric visual distribution analysis on point clouds and texture data in multidimensional model parsing data, identify key geometric structures and visual features, and generate geometric visual distribution data, which describes the distribution of geometric features in the model. Based on the geometric visual distribution data, mapping algorithms such as surface fitting or feature enhancement are applied to convert it into geometric visual data that is more suitable for simplified processing.
[0091] Step S3: performing mesh division management on the geometric visual data to generate mesh geometric visual data; performing mesh geometric visual feature weighting analysis processing on the mesh geometric visual data to generate weighted mesh geometric visual feature data;
[0092] In an embodiment of the present invention, the geometric visual data is meshed and processed, and an algorithm such as Delaunay triangulation or tetrahedronization is used to convert the continuous geometric body into a structure composed of multiple small meshes, each mesh containing the local geometry and visual information of the model. The geometric visual data of each mesh is analyzed, the importance of the mesh and the contribution to various visual influences are determined, and feature weighting is performed based on the importance of the mesh and the various visual influences. Key features are given higher weights to ensure that these key features are retained in the subsequent simplification process, thereby generating weighted mesh geometric visual feature data.
[0093] Step S4: performing cluster feature analysis on the weighted grid geometric visual feature data according to a preset self-organizing map neural network clustering algorithm to generate clustered grid geometric visual feature data; performing clustered grid level feature mapping on the clustered grid geometric visual feature data to generate clustered grid level feature mapping data; performing grid level simplification matrix analysis based on the clustered grid level feature mapping data to generate a grid level simplification matrix;
[0094] In an embodiment of the present invention, a self-organizing map neural network clustering algorithm is applied to weighted grid geometric visual feature data, and grids are grouped according to feature similarity to form clustered grid geometric visual feature data, and similar visual and geometric feature clusters in the identification model are identified. The clustered grid geometric visual feature data is mapped to the grid level to generate clustered grid level feature mapping data, such as defining the grid level based on the strength of the visual features of the clustering type, so as to facilitate accurate simplification processing according to the level in subsequent steps. The level feature mapping data is used to perform simplified matrix analysis, such as calculating the quaternary quadratic error matrix of each level to evaluate the simplification potential and impact, and finally generating a grid level simplified matrix to optimize the accuracy and efficiency of subsequent simplification steps.
[0095] Step S5: obtaining user historical interaction behavior data; performing a view distance simplification degree threshold analysis based on the user historical interaction behavior data to generate view distance simplification degree threshold data; performing adaptive LOD grid level simplification parameter prediction on the grid level simplification matrix and the view distance simplification degree threshold data based on a preset convolutional neural network algorithm to generate adaptive LOD grid level simplification parameters;
[0096] In an embodiment of the present invention, interaction data of users when using the model, such as viewing distance, frequently operated areas, etc., are collected. These data reflect the user's usage habits and needs. The viewing distance changes in the user interaction data are analyzed, and statistical analysis or machine learning models are used to identify the user's sensitivity to model details at different viewing distances, which is implemented by analyzing the relationship between the zoom level and the user's dwell time. According to the results of the viewing distance sensitivity analysis, the simplification degree threshold at different viewing distances is set. For example, for farther viewing distances, the degree of simplification can be higher to improve rendering performance; for close-up views, a higher level of detail is maintained to meet the user's demand for details. Combined with the grid level simplification matrix and the viewing distance simplification degree threshold, the optimal simplification parameters of each level are calculated, such as genetic algorithms or simulated annealing, to find parameter settings that maximize performance while maintaining visual quality, so as to generate adaptive LOD grid level simplification parameters to ensure that the model can automatically adjust the simplification level in different user interaction scenarios and optimize the balance between performance and visual effects.
[0097] Step S6: Adaptively simplify and render the mesh geometric visual data using adaptive LOD mesh level simplification parameters to generate adjusted simplified mesh geometric visual data; and perform mesh simplification management operations on the multidimensional model based on the adjusted simplified mesh geometric visual data.
[0098] In an embodiment of the present invention, adaptive LOD mesh level simplification parameters are used to dynamically simplify mesh geometric visual data, and a mesh simplification algorithm (such as polygon simplification technology, detail reduction technology) and adaptive LOD mesh level simplification parameters are used to perform real-time simplification processing on a selected mesh level, which involves reducing the number of vertices and triangles in the mesh while trying to keep the visual features of the mesh unchanged, applying any necessary visual enhancement technology, such as normal mapping or parallax mapping, to compensate for visual details that may be lost during the simplification process, adjusting the mesh to adapt to different viewing distances and performance requirements, generating and adjusting simplified mesh geometric visual data, and performing a final simplification management operation based on the adjusted mesh geometric visual data to ensure that the simplified model optimizes its performance and loading speed while maintaining the necessary visual quality to adapt to different usage environments and user needs.
[0099] Preferably, step S1 comprises the following steps:
[0100] Step S11: obtaining an initial multidimensional model;
[0101] Step S12: performing multidimensional model denoising processing on the initial multidimensional model according to a preset denoising filter decision to generate a denoised multidimensional model;
[0102] Step S13: performing time stamp correction and synchronization processing on the denoised multidimensional model to generate a standard multidimensional model;
[0103] Step S14: performing multidimensional model data analysis processing according to the standard multidimensional model to generate multidimensional model analysis data.
[0104] The present invention ensures that the system can accurately obtain the initial multidimensional model data to be processed, provides a data foundation, and ensures the integrity and reliability of the data. Noise reduction processing effectively removes noise and interference in the multidimensional model and improves the quality of the data. By applying the preset noise reduction filter decision, it can effectively filter different types of noise, making the denoised model more accurate and clear, laying a good foundation for subsequent data analysis and processing. The timestamp correction synchronization processing ensures the consistency and synchronization of the multidimensional model data in the time dimension, eliminates the data errors and deviations caused by time inconsistency, and the generated standard multidimensional model has higher time consistency and accuracy, making subsequent processing more stable and reliable. By performing data analysis processing on the standard multidimensional model, the point cloud data, texture data and color data in the model can be accurately extracted to form detailed multidimensional model analysis data, ensuring the comprehensive extraction and accurate description of all key information, providing detailed data support for subsequent simplification processing, and improving the quality and accuracy of the simplification effect.
[0105] In an embodiment of the present invention, the source of the initial multidimensional model is determined, including a 3D scanning device, a CAD system, GIS data or an existing 3D database, and the initial multidimensional model is imported from the selected source into the processing system. Ensure that no key data is lost during the import process, check the imported initial multidimensional model, verify the integrity and correctness of the data, check whether the model has missing parts or significant errors, and ensure that the basic data of the model is reliable to obtain the initial multidimensional model. According to the specific characteristics and intended use of the model, set appropriate noise reduction filter parameters, including: filter type (such as mean filter, median filter, Gaussian filter, etc.), filter strength and radius parameters. Select a suitable noise reduction algorithm to process the initial multidimensional model, such as using a median filter: traverse each point or vertex, calculate the median position of its neighborhood points, and replace the origin position with this position; if a Gaussian filter is used: calculate the Gaussian weighted average position of each point or vertex, update the point position to smooth the noise, and generate a noise reduction multidimensional model. Check the timestamp information of each part of the model data to confirm the time consistency of the data. If the model data comes from multiple scans or multiple devices, ensure that the timestamps of all data can be aligned. Use time interpolation or resampling technology to correct inconsistent timestamps, such as interpolating data at different times to align them to a unified time point, resampling data according to a unified time axis to ensure data synchronization, merging the corrected data, and generating a standard multidimensional model with consistent timestamps. Clarify the parsing target and required data types, including point cloud data, texture data, and color data, set parsing parameters such as parsing accuracy, extraction range, and feature recognition standards, and generate multidimensional model parsing data. Including: using point cloud processing algorithms to extract point cloud data from the model, including density analysis, noise removal, and feature point extraction. Analyze and extract texture information on the model surface, separate and store texture maps from the model, parse texture coordinates, and ensure that the maps are correctly aligned with the model surface. Extract and record the color features of the model, sample the color of each point on the model surface, record RGB values or other color space data, and analyze and store color distribution information, including color gradients, average colors, etc.
[0106] Preferably, the geometric visual distribution data includes geometric visual confidence distribution data and color interpolation distribution data, and step S2 includes the following steps:
[0107] Step S21: performing geometric vision confidence distribution evaluation processing according to the multi-dimensional model point cloud data and the multi-dimensional model texture data to generate geometric vision confidence distribution data;
[0108] Step S22: performing geometric visual association analysis on the multidimensional model point cloud data and the multidimensional model texture data to generate geometric visual association data, and performing geometric visual likelihood function analysis on the geometric visual association data to obtain a geometric visual likelihood function;
[0109] Step S23: performing geometric vision confidence distribution optimization processing on the geometric vision confidence distribution data according to the geometric vision likelihood function to generate optimized geometric vision confidence distribution data;
[0110] Step S24: performing color interpolation distribution analysis on the multidimensional model color data according to the optimized geometric vision confidence distribution data to generate color interpolation distribution data;
[0111] Step S25: Perform geometric vision mapping processing according to the optimized geometric vision confidence distribution data and the color interpolation distribution data to generate geometric vision data.
[0112] The present invention performs geometric visual confidence distribution evaluation on point cloud data and texture data, identifies and marks important geometric feature areas in the model, helps to distinguish high confidence and low confidence areas, so that in the subsequent simplification process, the high confidence areas can be retained, ensuring that the key details of the model are not ignored. Geometric visual association analysis can reveal the intrinsic connection between point cloud data and texture data, generate geometric visual association data, and quantify this connection through geometric visual likelihood function analysis, further enhance the understanding and description of geometric features, and provide a solid foundation for confidence distribution optimization. The confidence distribution data is optimized using the geometric visual likelihood function, and the geometric visual confidence distribution is accurately adjusted and optimized, so that the high confidence area is more prominent and the low confidence area is further filtered. In the subsequent simplification process, the important features of the model can be more accurately retained, and the quality of the simplification effect can be improved. Color interpolation distribution analysis uses the optimized geometric visual confidence distribution data to perform fine interpolation processing on the color data, which can better reflect the actual color and texture characteristics of the model and ensure the accuracy and consistency of color and texture during the simplification process. The geometric vision mapping process combines the optimized geometric vision confidence distribution data and color interpolation distribution data to generate geometric vision data, ensuring that the geometric and visual features of the model are fully preserved and accurately described during the simplification process.
[0113] In an embodiment of the present invention, a suitable geometric visual confidence evaluation algorithm is selected, such as an evaluation method based on statistics or a machine learning model, such as SVM or a neural network, and the geometric visual confidence is calculated for each point in the point cloud data. Based on the feature changes and consistency of its local neighborhood, a similar evaluation is performed on the texture data, and the continuity and local changes of the texture are considered. The confidence values of the point cloud data and the texture data of each point are respectively summarized to generate the overall geometric visual confidence distribution data. The relationship between the point cloud data and the texture data is analyzed, the geometric visual association data is extracted, the distance and direction difference between the point cloud point and its texture corresponding point is calculated, and the geometric visual likelihood function is defined by a Gaussian distribution or a Bayesian probability algorithm to quantify the probability distribution of the geometric visual association data, and the likelihood function value is calculated using the geometric visual association data to generate the geometric visual likelihood function. The optimization goal is determined, such as maximizing the consistency of the confidence or minimizing the fluctuation of the confidence, and a suitable optimization algorithm is selected, such as a gradient descent method, a genetic algorithm or other optimization techniques, and the geometric visual confidence distribution data is optimized and adjusted according to the geometric visual likelihood function to generate the optimized confidence distribution data. Select appropriate color interpolation methods, such as bilinear interpolation, cubic spline interpolation, etc., perform interpolation analysis on color data based on optimized geometric visual confidence distribution data, generate color interpolation distribution data, and ensure smooth transition and consistency of colors at different locations. Select appropriate geometric visual mapping algorithms, such as projection mapping, grid mapping, etc., apply mapping algorithms, map optimized geometric visual confidence distribution data and color interpolation distribution data to low-dimensional space for analysis, generate comprehensive geometric visual data, facilitate mesh simplification of multi-dimensional models, and ensure that the model is visually consistent and detailed.
[0114] Preferably, step S3 comprises the following steps:
[0115] Step S31: performing geometric topological structure analysis on the geometric visual data to generate geometric topological structure data;
[0116] Step S32: designing a geometric visual mesh division management decision based on the geometric topological structure data, and using the geometric visual mesh division management decision to perform mesh division management on the geometric visual data to generate mesh geometric visual data;
[0117] Step S33: performing feature analysis processing according to the mesh geometric visual data to generate mesh geometric visual feature data;
[0118] Step S34: performing feature weighting processing on the mesh geometric visual feature data according to a preset mesh geometric visual feature weighting decision to generate weighted mesh geometric visual feature data.
[0119] The present invention performs geometric topological structure analysis on geometric visual data, comprehensively understands the topological relationship and structural features of the model, helps to identify key connections and shape features in the model, provides important reference data for subsequent mesh division, and ensures the integrity and rationality of the structure during the simplification process. Mesh division decisions are made based on geometric topological structure data, so that the divided meshes have higher computing efficiency and processing performance while retaining the key geometric features of the model. The generated mesh geometric visual data is more structured, which improves the accuracy and efficiency of the overall simplification process. Feature analysis processing identifies and extracts important geometric and visual features in the mesh, generates mesh geometric visual feature data, ensures that during the simplification process, simplified hierarchical division is performed based on important features, reduces unnecessary detail loss, and improves the expressiveness and accuracy of the simplified model. Feature weighting processing assigns different weights to different geometric visual features according to the specific application requirements of the model, ensures that during the simplification process, important feature areas are retained first, while secondary areas are appropriately simplified, thereby achieving a more reasonable simplification effect and higher model quality.
[0120] In an embodiment of the present invention, the geometric visual data is analyzed to identify the basic topological structure of the model, such as vertices, edges, faces and bodies, and the connection relationship of each topological element is recorded to ensure complete topological structure data. Geometric features are extracted, including information such as edges, corners, and surfaces, and algorithms such as edge detection and curvature calculation are used to identify and mark key geometric features. The identified and extracted geometric features and their connection relationships are used to generate geometric topological structure data, providing a basis for subsequent meshing. Based on the geometric topological structure data, meshing rules and parameters are set, such as mesh size, shape and density, and appropriate meshing methods are selected, such as Delaunay triangulation, tetrahedron division, etc., and meshing strategies are optimized to ensure that meshing can both ensure the integrity of geometric features and improve computing efficiency. Mesh partitioning parameters are adjusted through simulation or testing to find the optimal partitioning decision, and the set meshing rules and optimized strategies are used to manage the meshing of geometric visual data, generate mesh geometric visual data, and ensure the uniformity and stability of the mesh. According to the mesh geometric visual data, the geometric features of each mesh unit are extracted, and the geometric properties of the mesh unit, such as area, volume, curvature, etc., are calculated. The visual features of the mesh unit, such as visual complexity, gradient information, etc., are extracted. Image processing algorithms, such as color histogram, texture analysis, gradient analysis, etc., are used to obtain visual feature data, integrate geometric and visual feature data, and generate complete mesh geometric visual feature data. According to the preset weighting decision, different feature weight rules are set for different geometric visual types. For example, various types of information in geometric complexity are given lower weights, and gradient information is given higher weights. According to the set weighting rules, the geometric visual feature data of each mesh unit are weighted to generate weighted mesh geometric visual feature data. Ensure that the weighted feature data can accurately reflect the important features of the model and provide a basis for subsequent simplification processing.
[0121] Preferably, step S33 includes the following steps:
[0122] Extracting mesh geometric visual complexity index data from the mesh geometric visual data, and performing mesh geometric visual complexity feature evaluation processing according to the mesh geometric visual complexity index data to generate mesh geometric visual complexity feature data;
[0123] Identify mesh vertex data in the mesh geometric visual data, and perform mesh geometric visual gradient feature analysis on the mesh geometric visual data according to the mesh vertex data to generate mesh geometric visual gradient feature data;
[0124] The mesh geometric visual complexity feature data and the mesh geometric visual gradient feature data are mesh geometric visual feature integrated to generate mesh geometric visual feature data.
[0125] The present invention extracts complexity index data from mesh geometric visual data, quantifies and evaluates the geometric complexity features of the model, helps identify areas with higher complexity in the model, enables subsequent simplification processing to be optimized for these complex areas, ensures that the geometric details and structural rationality of the overall model can be maintained during the simplification process, and improves the accuracy and expressiveness of the simplified model. Identify mesh vertex data and perform geometric gradient feature analysis to identify gradient change areas in the model, help locate important geometric change points and edges in the model, and enable these key features to be retained preferentially during the simplification process, avoid losing important geometric information during the simplification process, and improve the visual consistency and realism of the simplified model. Integrate complexity feature data and gradient feature data, comprehensively consider the geometric complexity and gradient change features of the model, and fully retain the important geometric and visual features of the model, achieve a more optimized mesh simplification effect, and improve the overall quality and expressiveness of the simplified model.
[0126] In the embodiment of the present invention, indicators for measuring the geometric complexity of the mesh are defined, such as the number of facets, vertex density, curvature change, etc. Common indicators include curvature (Gaussian curvature, mean curvature), facet density, vertex density, etc., and each mesh unit is traversed to calculate its geometric complexity indicators, such as using differential geometry methods to calculate the curvature of each mesh facet or vertex; counting the number of facets in each mesh unit; and counting the number of vertices in each mesh unit. Based on the calculated complexity indicators, the complexity characteristics of each mesh unit are evaluated, and the geometric complexity of the mesh unit is quantified using statistical analysis methods (such as mean, variance, standard deviation), and the evaluation results are integrated to generate geometric visual complexity feature data for each mesh unit. Identify all vertex data from the mesh geometric visual data, including vertex coordinates, normal vectors and other information. Perform gradient feature analysis on each vertex based on the identified vertex data, such as calculating the geometric gradient of each vertex. Common methods include finite difference method, gradient vector flow (GVF), etc. Analyze the direction and amplitude of the gradient of each vertex, identify the main direction and intensity of geometric changes, and generate geometric visual gradient feature data for each vertex based on the calculated gradient features. Determine the rules and methods for integrating complexity feature data and gradient feature data to ensure the unity and compatibility of the two types of feature data, match feature data for each mesh unit, associate the corresponding complexity feature data with gradient feature data, use spatial position and topological relationship to ensure the correct matching of feature data, and fuse the complexity feature data and gradient feature data according to the set integration rules to generate complete mesh geometric visual feature data, such as using weighted average or feature superposition, and normalize the integrated feature data to ensure the consistency of data format and structure, and generate the final mesh geometric visual feature data to provide complete feature information for subsequent simplification and optimization.
[0127] Preferably, step S4 comprises the following steps:
[0128] Step S41: performing clustering feature analysis on the weighted grid geometric visual feature data according to a preset self-organizing map neural network clustering algorithm to generate clustered grid geometric visual feature data;
[0129] Step S42: performing grid cluster label analysis according to the cluster grid geometric visual feature data to generate grid cluster label data;
[0130] Step S43: setting the grid division level according to the grid clustering label data to generate grid division level data;
[0131] Step S44: performing clustered grid level feature mapping processing on the clustered grid geometric visual feature data according to the grid division level data to generate clustered grid level feature mapping data;
[0132] Step S45: performing grid level error matrix analysis according to the grid level feature mapping data to generate a grid level error matrix;
[0133] Step S46: Perform grid level simplification matrix analysis based on the grid level error matrix to generate a grid level simplification matrix.
[0134] The present invention classifies mesh geometric visual feature data with similar features together through cluster feature analysis to form clustered mesh geometric visual feature data, which is helpful to discover and identify areas with similar attributes in the model, making subsequent processing more efficient, and can retain the overall features of these similar areas during the simplification process, thereby improving the consistency of the simplification effect. The mesh cluster label analysis assigns a label to each cluster, identifies the features and attributes of different clusters, and the generated mesh cluster label data helps to accurately distinguish and process different types of clusters during the simplification process, making the simplification process more targeted and accurate, and further improving the simplification quality and expressiveness of the model. The mesh division level is set according to the cluster label data, and the model is divided into multiple hierarchical structures. The generated mesh division hierarchical data ensures that the hierarchical structure can be optimized during the simplification process, thereby achieving more flexible and efficient mesh simplification. Through hierarchical feature mapping processing, the clustered mesh geometric visual feature data is mapped to different hierarchical structures, and clustered mesh hierarchical feature mapping data is generated, ensuring that the feature information of each level can be retained during the simplification process, so that the simplified model can maintain high accuracy and expressiveness at different levels. The grid-level error matrix analysis quantifies and evaluates the simplification errors at different levels. The generated grid-level error matrix helps control the error range during the simplification process, ensuring that the errors of the simplified model at each level are minimized, and improving the accuracy and reliability of the simplification process. The simplification strategy is optimized based on the data of the error matrix, and the grid-level simplification matrix is generated, ensuring that the simplification process can minimize the amount of data while controlling the error, achieve efficient grid simplification, and improve the processing efficiency and application performance of the model.
[0135] As an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 Detailed implementation step flow chart of step S4 in the embodiment, step S4 in this embodiment includes:
[0136] Step S41: performing clustering feature analysis on the weighted grid geometric visual feature data according to a preset self-organizing map neural network clustering algorithm to generate clustered grid geometric visual feature data;
[0137] In an embodiment of the present invention, a self-organizing map neural network clustering algorithm is selected to format the weighted grid geometric visual feature data into a data structure suitable for cluster analysis, usually a multidimensional grid vector space, and cluster the weighted grid geometric visual feature data to identify grid groups with similar features, such as performing cluster analysis on the values of the weighted grid geometric visual feature data, and storing the clustering results as clustered grid geometric visual feature data, in which each grid unit is marked as the cluster to which it belongs.
[0138] Step S42: performing grid cluster label analysis according to the cluster grid geometric visual feature data to generate grid cluster label data;
[0139] In an embodiment of the present invention, a unique label is assigned to each cluster to ensure that different clusters have clear identification. According to the cluster grid geometric visual feature data, the characteristics of each cluster label are statistically analyzed, such as calculating the cluster center, variance, etc. The analysis results are used to generate grid cluster label data with the corresponding grid simplified hierarchical information, providing a basis for subsequent hierarchical division.
[0140] Step S43: setting the grid division level according to the grid clustering label data to generate grid division level data;
[0141] In an embodiment of the present invention, hierarchical rules for grid division are set based on grid clustering label data, such as the granularity and detail level of each level and the hierarchical division interval of the corresponding clustering grid geometric visual feature data. According to the set rules, the clustering grids are hierarchically divided to ensure that the grids in each level have similar features and details. The hierarchical information of each grid is recorded, and grid division hierarchical data is generated to ensure the accuracy and consistency of the hierarchical division.
[0142] Step S44: performing clustered grid level feature mapping processing on the clustered grid geometric visual feature data according to the grid division level data to generate clustered grid level feature mapping data;
[0143] In an embodiment of the present invention, the rules and methods for clustered grid hierarchical feature mapping, such as mapping priority and strategy, are determined, and hierarchical feature mapping is performed on the clustered grid geometric visual feature data based on the grid division hierarchical data to ensure the transmission and integration of features between different levels. The mapping results are used to generate clustered grid hierarchical feature mapping data to ensure the accuracy and completeness of the feature mapping, so that the clustered grid geometric visual feature data has corresponding grid division hierarchical labels.
[0144] Step S45: performing grid level error matrix analysis according to the grid level feature mapping data to generate a grid level error matrix;
[0145] In an embodiment of the present invention, error measurement is performed based on vertex position error, normal vector error, curvature error, etc. in mesh level feature mapping data, and error calculation is performed between each adjacent level, such as using Hausdorff distance or root mean square error (RMSE) to calculate the normal vector error between meshes, and the calculated errors are used to construct an error matrix, such as a quaternary quadratic matrix, to evaluate mesh error information between different levels and represent the error relationship between the levels. The rows and columns of the error matrix represent different levels, respectively, and each element in the matrix represents the error value between the corresponding levels, thereby generating a mesh level error matrix.
[0146] Step S46: Perform grid level simplification matrix analysis based on the grid level error matrix to generate a grid level simplification matrix.
[0147] In an embodiment of the present invention, a simplification goal is set according to application requirements, such as minimizing errors, balancing computing resources and visual effects, optimizing rendering speed, etc. Suitable simplification strategies are selected, such as QEM (Quadratic Error Metric), multi-level mesh simplification (MLM), etc., and the priority and application scenario of the strategy are determined according to the set simplification goals. Initialize the parameters of the simplification algorithm, such as the initial error threshold, the number of iterations, etc. If the QEM algorithm is used, the simplification parameters are calculated according to the error matrix, and the simplification priority is calculated for each grid unit. The unit with the smallest error is simplified first, the error matrix is updated, and the simplification parameters are adjusted in real time to ensure that the overall error is controlled within the set range. Based on the optimization results, a grid-level simplification strategy is generated, the simplification parameters and processing steps of each level are recorded, and the simplification strategy is converted into a simplification matrix. The matrix elements represent the simplification decisions and parameters of different levels.
[0148] Preferably, step S5 comprises the following steps:
[0149] Step S51: Obtaining user historical interaction behavior data;
[0150] Step S52: performing interaction distance and interaction visual correlation analysis based on the user's historical interaction behavior data to generate historical interaction visual distance correlation data;
[0151] Step S53: performing interactive sight distance fuzzy kernel analysis according to historical interactive sight distance associated data to generate historical interactive sight distance fuzzy kernel data;
[0152] Step S54: evaluating the threshold value of the degree of simplification of the sight distance based on the historical interactive sight distance blur kernel data, and generating the threshold data of the degree of simplification of the sight distance;
[0153] Step S55: Based on a preset convolutional neural network algorithm, adaptive LOD grid level simplification parameter prediction is performed on the grid level simplification matrix and the view distance simplification degree threshold data to generate adaptive LOD grid level simplification parameters.
[0154] The present invention collects the user's historical interactive behavior data, understands the user's actual operating habits and needs when using the model, and optimizes the user's specific usage situation, so that the subsequent simplification process is more in line with the user's usage habits, improving the user experience and satisfaction. The interactive distance and interactive visual association analysis can reveal the user's operating behavior and visual focus at different viewing distances. The generated historical interactive viewing distance association data helps understand the user's needs in different usage scenarios, ensuring that the simplified model can meet the user's actual needs at different viewing distances, and improving the adaptability and flexibility of the model. Through the interactive viewing distance fuzzy kernel analysis, the fuzzy relationship between different distances and vision in the user's interactive behavior is understood, and smoother and continuous viewing distance association data is generated, which helps to reduce noise and outliers in the analysis, improve the accuracy and reliability of the analysis results, and set the user to observe the multidimensional model according to the viewing distance for simplification intensity analysis. The viewing distance simplification degree threshold evaluation can determine the simplification degree threshold suitable for different viewing distances according to the user's historical interactive behavior and viewing distance fuzzy kernel data, ensuring that at different viewing distances, the simplified model can not only meet the performance requirements, but also provide sufficient details and visual effects, and optimize the user experience. By combining the grid level simplification matrix and the viewing distance simplification degree threshold data, adaptive LOD (Level of Detail) grid level simplification parameter analysis is performed, and the simplification parameters are dynamically adjusted so that the model can provide the best visual effects and performance under different viewing distances and application scenarios. The generated adaptive LOD grid level simplification parameters help to achieve efficient simplification and intelligent optimization of the model, thereby improving overall processing efficiency and user satisfaction.
[0155] As an embodiment of the present invention, refer to Figure 3 As shown, Figure 1 FIG. 5 is a flow chart of detailed implementation steps of step S5, in this embodiment, step S5 includes:
[0156] Step S51: Obtaining user historical interaction behavior data;
[0157] In an embodiment of the present invention, a user behavior tracking tool is integrated into an application, and the interaction data of users when using a multidimensional model is collected through log records or a dedicated API. The interaction data items that need to be collected are defined, including but not limited to: viewing angle (viewing angle), zoom level (viewing distance), user dwell time (viewing time in a specific area), interaction frequency (operation frequency), etc., to obtain historical user interaction behavior data.
[0158] Step S52: performing interaction distance and interaction visual correlation analysis based on the user's historical interaction behavior data to generate historical interaction visual distance correlation data;
[0159] In an embodiment of the present invention, the collected user interaction behavior data is preprocessed, including data cleaning, deduplication, outlier processing, etc., to ensure data quality, calculate the user's interaction distance under different operations, quantify the change in distance between the user and the model, use statistical methods (such as mean, variance, etc.) to analyze the user's behavior patterns at different viewing distances, analyze the user's visual focus at different viewing distances, use heat maps or other visualization methods to identify user high-frequency interaction areas, and analyze the correlation information of interactive visual resolution at the interaction distance, combine the viewing distance and visual focus point data, and generate historical interaction viewing distance correlation data.
[0160] Step S53: performing interactive sight distance fuzzy kernel analysis according to historical interactive sight distance associated data to generate historical interactive sight distance fuzzy kernel data;
[0161] In an embodiment of the present invention, a suitable blur kernel type is selected to define a blur kernel function, such as a Gaussian blur kernel, a triangular blur kernel, etc., which is used to describe the interactive behavior pattern of users at different viewing distances. The historical interactive viewing distance associated data is used to calculate the blur kernel function, and the blur kernel data of the user at different viewing distances is generated. Through blur kernel analysis, the blur boundaries and change trends of the user's interactive behavior are identified, and the calculated blur kernel data is fused with the user's interactive behavior data to generate historical interactive viewing distance blur kernel data.
[0162] Step S54: evaluating the threshold value of the degree of simplification of the sight distance based on the historical interactive sight distance blur kernel data, and generating the threshold data of the degree of simplification of the sight distance;
[0163] In an embodiment of the present invention, a setting standard for a viewing distance simplification degree threshold is defined, and the simplification degree thresholds at different viewing distances are calculated based on historical interactive viewing distance fuzzy kernel data in combination with user interaction behaviors and blur kernel analysis results. The user's requirements for model details at different viewing distances are quantified, and the corresponding simplification thresholds are set. The viewing distance simplification degree thresholds are adjusted through multiple simulations and iterative optimizations to ensure the best simplification effect. The calculated and optimized viewing distance simplification degree thresholds are recorded to generate viewing distance simplification degree threshold data.
[0164] Step S55: Based on a preset convolutional neural network algorithm, adaptive LOD grid level simplification parameter prediction is performed on the grid level simplification matrix and the view distance simplification degree threshold data to generate adaptive LOD grid level simplification parameters.
[0165] In an embodiment of the present invention, a grid level simplification matrix is integrated with viewing distance simplification degree threshold data to form a data basis for parameter analysis, an adaptive LOD (Level of Detail) model is defined, LOD simplification parameters of different levels are determined, and an optimization algorithm (such as genetic algorithm, particle swarm optimization, etc.) is used to optimize the adaptive LOD parameters, with the goal of minimizing errors and maximizing visual effects and performance balance. Based on the integrated data and the optimization algorithm, the adaptive LOD grid level simplification parameters are analyzed and calculated, and the optimal LOD simplification parameters of each grid level at different viewing distances are generated. The analysis and calculation results are recorded, and the adaptive LOD grid level simplification parameters are generated to ensure that the model details can be dynamically adjusted in actual applications.
[0166] Preferably, step S54 includes the following steps:
[0167] The fuzzy kernel residual is calculated according to the historical interactive line-of-sight fuzzy kernel data to generate the historical line-of-sight fuzzy kernel residual data;
[0168] Perform fitting residual processing on the historical line-of-sight blur kernel residual data to generate historical line-of-sight blur kernel fitting residual data;
[0169] The sight distance simplification degree threshold is evaluated based on the historical sight distance blur kernel fitting residual data to generate the sight distance simplification degree threshold data.
[0170] The present invention performs residual calculations on historical interactive line-of-sight blur kernel data, quantifies and evaluates deviations and anomalies in user interaction behaviors, and the generated blur kernel residual data helps to identify blur kernel strength, making the analysis results more accurate and stable. The fitting residual processing further smoothes and optimizes the blur kernel residual data, and the generated fitting residual data can more accurately reflect the laws and trends of user interaction behaviors. Through the fitting residual processing, the random errors in the analysis are reduced, the accuracy and effect of the simplification process are improved, and the simplified model is more in line with the actual needs of users. Based on the fitting residual data, the line-of-sight simplification degree threshold evaluation is performed, and the simplification threshold at different line-of-sight distances is accurately determined, ensuring that at different line-of-sight distances, the simplified model can meet performance requirements and provide sufficient details and visual effects. The generated line-of-sight simplification degree threshold data makes the model simplification process more intelligent and efficient, and improves the user experience and model adaptability.
[0171] In an embodiment of the present invention, blur kernel residual calculation is performed based on historical interactive line-of-sight blur kernel data, and a suitable mathematical method is selected to calculate the residual, such as a difference method, a least square method, etc., to measure the deviation between the blur kernel data and a reference model (such as an ideal blur kernel shape). The selected residual calculation method is applied to the historical interactive line-of-sight blur kernel data, and the difference between each data point and the ideal model is analyzed. The calculated residual value reflects the degree of deviation between the user's actual interactive behavior and the preset blur kernel behavior model. The calculated residual value of each data point is sorted and summarized to generate historical line-of-sight blur kernel residual data. According to the characteristics of the residual data, a suitable data fitting method is selected, such as polynomial fitting, curve fitting, etc., in order to reduce errors and smooth data through the fitting process. The selected fitting method is applied to the historical line-of-sight blur kernel residual data, and the fitting parameters are adjusted until the optimal fitting effect is achieved, so as to reduce the influence of random fluctuations and outliers of the data, and generate historical line-of-sight blur kernel fitting residual data. According to the distribution characteristics of the fitting residual data, select appropriate statistical methods to evaluate the threshold of the degree of line-of-sight simplification, such as statistical tests based on confidence intervals, outlier analysis, etc. Apply the selected evaluation method to analyze the fitting residual data, determine the threshold of the degree of line-of-sight simplification that can best represent the characteristics of user interaction behavior, adjust the threshold until the optimal threshold that can balance maintaining user experience quality and optimizing performance is found, organize and record the evaluated line-of-sight simplification thresholds, and generate line-of-sight simplification threshold data.
[0172] Preferably, step S6 comprises the following steps:
[0173] Step S61: performing mesh adaptive simplification rendering processing on the mesh geometric visual data by using adaptive LOD mesh level simplification parameters to generate simplified mesh geometric visual data;
[0174] Step S62: performing global visual smoothing adjustment processing on the simplified mesh geometric visual data to generate adjusted simplified mesh geometric visual data;
[0175] Step S63: Execute the mesh simplification management operation of the multi-dimensional model according to the adjusted simplified mesh geometric visual data.
[0176] The present invention can dynamically adjust the degree of simplification according to specific application requirements and user interaction behaviors through adaptive LOD grid level simplification parameters, ensuring the best simplification effect under different viewing distances and scenes. Through adaptive simplified rendering processing, the simplified grid geometric visual data generated can not only meet the performance requirements, but also maintain sufficient visual details and accuracy, thereby improving the actual use effect and user experience of the model. The global visual smoothing adjustment process can effectively remove the visual noise and discontinuity that may be generated during the simplification process, making the simplified grid geometric visual data smoother and more consistent. The generated adjusted simplified grid geometric visual data not only improves the visual quality of the model, but also reduces visual fatigue and discomfort, and enhances the user's viewing experience and the overall aesthetics of the model. The final grid simplification management operation is based on the adjusted simplified grid geometric visual data, ensuring the efficiency and accuracy of the simplification process. The simplified multidimensional model not only has high computational efficiency and performance, but also maintains high visual quality and detail performance, providing a reliable and high-quality simplified model for practical applications.
[0177] In an embodiment of the present invention, a simplification engine is configured according to adaptive LOD mesh level simplification parameters, which specify the degree of simplification of meshes at different levels, including the proportion of vertices and facets to be retained at each level, and a simplification algorithm (such as polygon simplification technology) is applied to dynamically adjust the complexity of the mesh according to the adaptive LOD mesh level simplification parameters, including reducing the number of vertices and simplifying the mesh structure, while retaining important visual features as much as possible, and generating simplified mesh geometric visual data, each data set including the simplified mesh and its corresponding visual attributes (such as texture, color). A global visual smoothing strategy is defined, such as using a smoothing filter (Gaussian filter, mean filter, etc.) to smooth the edges of the simplified mesh to reduce visual artifacts or edge motion noise, smoothing the simplified mesh geometric visual data, especially in highly simplified areas, to improve visual quality and reduce hard edges, and saving the adjusted simplified mesh geometric visual data to ensure that all processed data has good visual coherence and smoothness. Specific tasks of the mesh simplification management operation are configured, such as regularly updating simplified mesh data, monitoring simplification effects, adjusting simplification parameters, etc., and the mesh simplification management operation is performed according to the adjusted simplified mesh geometric visual data. This includes data maintenance, updating simplification parameters, and real-time adjustment of simplification levels to adapt to different user needs and hardware environments, monitoring the execution of grid simplification management jobs, and adjusting management policies and simplification parameters based on monitoring data to ensure optimized performance and user experience.
[0178] This specification provides a grid simplification management system for a multidimensional model, which is used to execute the grid simplification management method for a multidimensional model as described above. The grid simplification management system for a multidimensional model includes:
[0179] A multidimensional model analysis module is used to obtain an initial multidimensional model; perform multidimensional model data analysis based on the initial multidimensional model to generate multidimensional model analysis data, wherein the multidimensional model analysis data includes multidimensional model point cloud data, multidimensional model texture data and multidimensional model color data;
[0180] A geometric vision mapping module is used to analyze geometric vision distribution data according to multi-dimensional model parsing data to generate geometric vision distribution data; perform geometric vision mapping processing according to the geometric vision distribution data to generate geometric vision data;
[0181] The mesh geometric visual feature analysis module is used to perform mesh division management on the geometric visual data and generate mesh geometric visual data; perform mesh geometric visual feature weighting analysis and processing based on the mesh geometric visual data and generate weighted mesh geometric visual feature data;
[0182] A grid level simplification matrix analysis module is used to perform clustering feature analysis on weighted grid geometric visual feature data according to a preset self-organizing map neural network clustering algorithm to generate clustered grid geometric visual feature data; perform clustered grid level feature mapping on the clustered grid geometric visual feature data to generate clustered grid level feature mapping data; perform grid level simplification matrix analysis based on the clustered grid level feature mapping data to generate a grid level simplification matrix;
[0183] An adaptive LOD grid level simplification parameter prediction module is used to obtain historical user interaction behavior data; perform a view distance simplification degree threshold analysis based on the historical user interaction behavior data to generate view distance simplification degree threshold data; perform adaptive LOD grid level simplification parameter prediction on the grid level simplification matrix and the view distance simplification degree threshold data based on a preset convolutional neural network algorithm to generate adaptive LOD grid level simplification parameters;
[0184] The multi-dimensional model mesh simplification module is used to perform mesh adaptive simplification rendering adjustment on mesh geometric visual data through adaptive LOD mesh level simplification parameters to generate adjusted simplified mesh geometric visual data; and execute mesh simplification management operations of the multi-dimensional model based on the adjusted simplified mesh geometric visual data.
[0185] The beneficial effect of the present application is that the present invention ensures that the basic data of the simplified processing is complete and accurate from obtaining the initial multidimensional model to generating detailed multidimensional model analysis data (including point cloud data, texture data, and color data), and the noise reduction filter decision and timestamp correction synchronization processing improve the quality and consistency of the model data, providing a reliable data basis for subsequent steps. Through geometric visual distribution data analysis and geometric visual mapping processing, important geometric structures and visual details in the model are identified and retained, and the visual consistency and expressiveness of the simplified model are improved. Cluster feature analysis, hierarchical feature mapping, and simplified matrix analysis realize intelligent clustering and hierarchical simplification of structured grid geometric visual data, ensure flexible adjustment of model complexity under different accuracy requirements, improve processing efficiency and reduce computing resource consumption, and use hierarchical error matrix analysis and simplified matrix analysis to ensure the accuracy and reliability of the simplification process. Considering the user's historical interactive behavior data to dynamically adjust the viewing distance simplification degree threshold and adaptive LOD parameters, the simplified model can better meet the actual user needs, making the simplification process more intelligent and personalized. Adaptive simplified rendering processing combined with global visual smoothing adjustment ensures that the simplified model can provide the best visual effects and performance at different viewing distances and scenes. The simplified management operation ensures the efficiency, accuracy and visual quality of the final model, providing a reliable solution for practical applications. The mesh simplification management method of the multi-dimensional model can efficiently and accurately simplify the mesh while maintaining the important geometric visual features of each mesh.
[0186] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.
[0187] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.
Claims
1. A method for simplifying and managing a multidimensional model grid, characterized in that: The following steps are involved: Step S1: obtaining an initial multidimensional model; Perform multidimensional model data analysis based on the initial multidimensional model to generate multidimensional model analysis data, wherein the multidimensional model analysis data includes multidimensional model point cloud data, multidimensional model texture data and multidimensional model color data; Step S2: performing geometric visual distribution data analysis based on the multidimensional model parsing data to generate geometric visual distribution data; Perform geometric vision mapping processing according to the geometric vision distribution data to generate geometric vision data; Wherein, step S2 comprises: Step S21: performing geometric vision confidence distribution evaluation processing according to the multi-dimensional model point cloud data and the multi-dimensional model texture data to generate geometric vision confidence distribution data; Step S22: performing geometric visual association analysis on the multidimensional model point cloud data and the multidimensional model texture data to generate geometric visual association data, and performing geometric visual likelihood function analysis on the geometric visual association data to obtain a geometric visual likelihood function; Step S23: performing geometric vision confidence distribution optimization processing on the geometric vision confidence distribution data according to the geometric vision likelihood function to generate optimized geometric vision confidence distribution data; Step S24: performing color interpolation distribution analysis on the multidimensional model color data according to the optimized geometric vision confidence distribution data to generate color interpolation distribution data; Step S25: performing geometric vision mapping processing according to the optimized geometric vision confidence distribution data and the color interpolation distribution data to generate geometric vision data; Step S3: performing grid division management on the geometric vision data to generate grid geometric vision data; performing grid geometric vision feature weighting analysis processing according to the grid geometric vision data to generate weighted grid geometric vision feature data; Step S4: performing cluster feature analysis on the weighted grid geometric visual feature data according to a preset self-organizing map neural network clustering algorithm to generate clustered grid geometric visual feature data; performing clustered grid level feature mapping on the clustered grid geometric visual feature data to generate clustered grid level feature mapping data; performing grid level simplification matrix analysis based on the clustered grid level feature mapping data to generate a grid level simplification matrix; Step S5: obtaining user historical interaction behavior data; performing a view distance simplification degree threshold analysis based on the user historical interaction behavior data to generate view distance simplification degree threshold data; performing adaptive LOD grid level simplification parameter prediction on the grid level simplification matrix and the view distance simplification degree threshold data based on a preset convolutional neural network algorithm to generate adaptive LOD grid level simplification parameters; Wherein step S5 comprises: Step S51: Obtaining user historical interaction behavior data; Step S52: performing interaction distance and interaction visual correlation analysis based on the user's historical interaction behavior data to generate historical interaction visual distance correlation data; Step S53: performing interactive sight distance fuzzy kernel analysis according to historical interactive sight distance associated data to generate historical interactive sight distance fuzzy kernel data; Step S54: evaluating the threshold value of the degree of simplification of the sight distance based on the historical interactive sight distance blur kernel data, and generating the threshold data of the degree of simplification of the sight distance; Step S54 includes: performing blur kernel residual calculation according to the historical interactive sight distance blur kernel data to generate historical sight distance blur kernel residual data; performing fitting residual processing on the historical sight distance blur kernel residual data to generate historical sight distance blur kernel fitting residual data; performing sight distance simplification degree threshold evaluation according to the historical sight distance blur kernel fitting residual data to generate sight distance simplification degree threshold data; Step S55: performing adaptive LOD grid level simplification parameter prediction on the grid level simplification matrix and the view distance simplification degree threshold data based on a preset convolutional neural network algorithm to generate adaptive LOD grid level simplification parameters; Step S6: Adaptively simplify and render the mesh geometric visual data using adaptive LOD mesh level simplification parameters to generate adjusted simplified mesh geometric visual data; and perform mesh simplification management operations on the multidimensional model based on the adjusted simplified mesh geometric visual data.
2. The method for managing mesh simplification of a multidimensional model according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: obtaining an initial multidimensional model; Step S12: performing multidimensional model denoising processing on the initial multidimensional model according to a preset denoising filter decision to generate a denoised multidimensional model; Step S13: performing time stamp correction and synchronization processing on the denoised multidimensional model to generate a standard multidimensional model; Step S14: performing multidimensional model data analysis processing according to the standard multidimensional model to generate multidimensional model analysis data.
3. The method for managing mesh simplification of a multidimensional model according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: performing geometric topological structure analysis on the geometric visual data to generate geometric topological structure data; Step S32: designing a geometric visual mesh division management decision based on the geometric topological structure data, and using the geometric visual mesh division management decision to perform mesh division management on the geometric visual data to generate mesh geometric visual data; Step S33: performing feature analysis processing according to the mesh geometric visual data to generate mesh geometric visual feature data; Step S34: performing feature weighting processing on the mesh geometric visual feature data according to a preset mesh geometric visual feature weighting decision to generate weighted mesh geometric visual feature data.
4. The method for managing mesh simplification of a multidimensional model according to claim 3, characterized in that: Step S33 includes the following steps: Extracting mesh geometric visual complexity index data from the mesh geometric visual data, and performing mesh geometric visual complexity feature evaluation processing according to the mesh geometric visual complexity index data to generate mesh geometric visual complexity feature data; Identify mesh vertex data in the mesh geometric visual data, and perform mesh geometric visual gradient feature analysis on the mesh geometric visual data according to the mesh vertex data to generate mesh geometric visual gradient feature data; The mesh geometric visual complexity feature data and the mesh geometric visual gradient feature data are mesh geometric visual feature integrated to generate mesh geometric visual feature data.
5. The method for managing mesh simplification of a multidimensional model according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: performing clustering feature analysis on the weighted grid geometric visual feature data according to a preset self-organizing map neural network clustering algorithm to generate clustered grid geometric visual feature data; Step S42: performing grid cluster label analysis according to the cluster grid geometric visual feature data to generate grid cluster label data; Step S43: setting the grid division level according to the grid clustering label data to generate grid division level data; Step S44: performing clustered grid level feature mapping processing on the clustered grid geometric visual feature data according to the grid division level data to generate clustered grid level feature mapping data; Step S45: performing grid level error matrix analysis according to the grid level feature mapping data to generate a grid level error matrix; Step S46: Perform grid level simplification matrix analysis based on the grid level error matrix to generate a grid level simplification matrix.
6. The method for managing mesh simplification of a multidimensional model according to claim 1, characterized in that: Step S6 includes the following steps: Step S61: performing mesh adaptive simplification rendering processing on the mesh geometric visual data by using adaptive LOD mesh level simplification parameters to generate simplified mesh geometric visual data; Step S62: performing global visual smoothing adjustment processing on the simplified mesh geometric visual data to generate adjusted simplified mesh geometric visual data; Step S63: Execute the mesh simplification management operation of the multi-dimensional model according to the adjusted simplified mesh geometric visual data.
7. A grid simplification management system for a multidimensional model, characterized in that: A method for executing a mesh simplification management method for a multidimensional model as claimed in claim 1, wherein the mesh simplification management system for a multidimensional model comprises: A multidimensional model analysis module is used to obtain an initial multidimensional model; perform multidimensional model data analysis based on the initial multidimensional model to generate multidimensional model analysis data, wherein the multidimensional model analysis data includes multidimensional model point cloud data, multidimensional model texture data and multidimensional model color data; A geometric vision mapping module is used to perform geometric vision confidence distribution evaluation processing based on multidimensional model point cloud data and multidimensional model texture data to generate geometric vision confidence distribution data; perform geometric vision association analysis on multidimensional model point cloud data and multidimensional model texture data to generate geometric vision association data, and perform geometric vision likelihood function analysis on the geometric vision association data to obtain a geometric vision likelihood function; perform geometric vision confidence distribution optimization processing on the geometric vision confidence distribution data according to the geometric vision likelihood function to generate optimized geometric vision confidence distribution data; perform color interpolation distribution analysis on the multidimensional model color data according to the optimized geometric vision confidence distribution data to generate color interpolation distribution data; perform geometric vision mapping processing based on the optimized geometric vision confidence distribution data and color interpolation distribution data to generate geometric vision data; a grid geometric vision feature analysis module is used to perform grid division management on the geometric vision data to generate grid geometric vision data; perform grid geometric vision feature weighting analysis processing based on the grid geometric vision data to generate weighted grid geometric vision feature data; A grid level simplification matrix analysis module is used to perform clustering feature analysis on weighted grid geometric visual feature data according to a preset self-organizing map neural network clustering algorithm to generate clustered grid geometric visual feature data; perform clustered grid level feature mapping on the clustered grid geometric visual feature data to generate clustered grid level feature mapping data; perform grid level simplification matrix analysis based on the clustered grid level feature mapping data to generate a grid level simplification matrix; The adaptive LOD grid level simplification parameter prediction module is used to obtain the user's historical interaction behavior data; perform interaction distance and interaction visual correlation analysis based on the user's historical interaction behavior data to generate historical interaction sight distance correlation data; perform interaction sight distance blur kernel analysis based on the historical interaction sight distance correlation data to generate historical interaction sight distance blur kernel data; perform blur kernel residual calculation based on the historical interaction sight distance blur kernel data to generate historical sight distance blur kernel residual data; perform fitting residual processing on the historical sight distance blur kernel residual data to generate historical sight distance blur kernel fitting residual data; perform sight distance simplification degree threshold evaluation based on the historical sight distance blur kernel fitting residual data to generate sight distance simplification degree threshold data; perform adaptive LOD grid level simplification parameter prediction on the grid level simplification matrix and the sight distance simplification degree threshold data based on the preset convolutional neural network algorithm to generate adaptive LOD grid level simplification parameters; The multi-dimensional model mesh simplification module is used to perform mesh adaptive simplification rendering adjustment on mesh geometric visual data through adaptive LOD mesh level simplification parameters to generate adjusted simplified mesh geometric visual data; and execute mesh simplification management operations of the multi-dimensional model based on the adjusted simplified mesh geometric visual data.
Citation Information
Patent Citations
Natural image super pixel segmentation method based on graph model
CN105913451A
Automatic generation method of color LOD model based on block decomposition
CN109118588A
Wind power prediction method based on weighted fuzzy clustering
CN114219179A
BIM model processing method and device, storage medium and electronic device
CN115238345A
Augmented reality visualization technical scheme in cable line inspection process
CN118397228A
