Data lightweight processing method and system for simulation model

Abstract

The present invention relates to the field of data processing, and in particular to a method and system for lightweight data processing of a simulation model. The method comprises the following steps: obtaining the material complexity of each region based on the simulation model to be processed; quantifying the table structure key quantity of the simulation model geometric structure points based on the material complexity of each region to obtain the table structure point key value; performing structural simplification processing on the simulation model to be processed based on the table structure point key value, thereby obtaining a high dynamic frame region model and a low dynamic frame region model; performing dynamic detail rendering on the high dynamic frame region model and the low dynamic frame region model to obtain a frame detail rendering model and a resampling model; performing floating point encoding on the frame detail rendering model to generate a frame detail encoding sequence; performing posture change discrete frequency calculation on the resampling model to generate an entropy encoding sequence. The present invention realizes efficient and reliable lightweight processing of simulation models.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to a method and system for lightweight data processing of a simulation model. Background Art

[0002] In today's wave of industrial automation and digitalization, the application of various simulation models is becoming increasingly widespread. As a digital carrier of virtual reality, simulation models play an important role in many fields such as product design, process planning, control and debugging. As the complexity of simulation models continues to increase, the amount of model data is also growing exponentially, which brings huge challenges to storage, transmission and processing. Traditional simulation model data processing methods usually rely on manual analysis and manual optimization, which are inefficient and difficult to meet the needs of modern industrial automation. For a complex three-dimensional simulation model, its geometric shape, material properties, motion state and other parameters are numerous. If the original digital description method is used for storage and transmission, it will inevitably take up a lot of storage space and bandwidth resources. At the same time, in actual applications, not all model details require high-precision expression, and the accuracy can often be appropriately reduced in exchange for higher processing efficiency. Therefore, there is an urgent need for an intelligent simulation model lightweight processing method that can automatically optimize the storage and transmission of model data for different application scenarios to achieve efficient and reliable simulation model processing. Summary of the invention

[0003] In order to solve the above technical problems, the present invention proposes a data lightweight processing method and system for a simulation model to solve at least one of the above technical problems.

[0004] To achieve the above object, the present invention provides a method for lightweight processing of simulation model data, comprising the following steps:

[0005] Step S1: obtaining the material complexity of each region based on the simulation model to be processed; quantifying the key values ​​of the table structure of the geometric structure points of the simulation model based on the material complexity of each region to obtain the key values ​​of the table structure points;

[0006] Step S2: Simplifying the structure of the simulation model to be processed based on the key values ​​of the table structure points, thereby obtaining a high dynamic frame region model and a low dynamic frame region model;

[0007] Step S3: performing dynamic detail rendering on the high dynamic frame region model and the low dynamic frame region model to obtain a frame detail rendering model and a resampling model;

[0008] Step S4: performing floating point encoding on the frame detail rendering model to generate a frame detail encoding sequence; performing posture change discrete frequency calculation on the resampling model to generate an entropy encoding sequence;

[0009] Step S5: extracting multi-source characteristic parameters of the simulation model based on the simulation model to be processed; performing heterogeneous parameter encoding on the multi-source characteristic parameters of the simulation model to obtain characteristic parameter compression encoding;

[0010] Step S6: Adaptively optimize the encoding parameters of the entropy coding sequence, the frame detail coding sequence and the feature parameter compression coding to obtain an adaptive coding fine-tuning strategy; and perform a simulation model lightweight processing operation based on the adaptive coding fine-tuning strategy.

[0011] The present invention obtains the material complexity of each area based on the simulation model to be processed to understand the material complexity of different areas in the model, which is helpful for subsequent structural simplification and data lightweight processing. Based on the material complexity of each area, the table structure key value of the simulation model geometric structure points is quantified to determine which geometric structure points are more important to the surface structure. The simulation model to be processed is structurally simplified based on the key value of the table structure point to reduce redundant information and complexity in the model. The high dynamic frame area model and the low dynamic frame area model select the model with high dynamic range or low dynamic range according to different needs to achieve flexible application and lightweight processing of data. Dynamic detail rendering is performed on the high dynamic frame area model and the low dynamic frame area model to increase the realism and details of the model and improve the visual effect. The frame detail rendering model and the resampling model select different models according to different needs and scenarios to achieve flexible rendering and resampling. The frame detail rendering model is floating-point encoded to compress and encode the data of the rendering model. Reduce the space required for storage and transmission, calculate the discrete frequency of posture changes of the resampled model to extract the posture change information of the model, generate frame detail coding sequence and entropy coding sequence to further compress and encode the data of the model, realize lightweight processing and optimization of data, extract multi-source feature parameters of the simulation model based on the simulation model to be processed, extract various feature information of the model, including geometric shape, material properties, etc., perform heterogeneous parameter encoding on the multi-source feature parameters of the simulation model, uniformly encode different types of parameters, reduce the amount of data stored and transmitted, perform adaptive coding parameter optimization on entropy coding sequence, frame detail coding sequence and feature parameter compression coding, adjust the coding parameters according to the characteristics and requirements of the data, further optimize the data compression effect, fine-tune the coding according to different situations and requirements, improve the compression rate and effect of the data, execute the simulation model lightweight processing operation, perform data lightweight processing on the simulation model according to the adaptive coding fine-tuning strategy, reduce the amount of data stored and transmitted, and improve processing efficiency and performance.

[0012] Preferably, step S1 comprises the following steps:

[0013] Step S11: performing regional material complexity identification on the simulation model to be processed to obtain the material complexity of each region;

[0014] Step S12: performing topological cutting on the simulation model to be processed to obtain a plurality of cut simulation models;

[0015] Step S13: performing section contour recognition on a plurality of cutting simulation models, thereby generating section contour lines;

[0016] Step S14: identifying geometric structure points of multiple cutting simulation models based on the section contour lines, and marking the geometric structure points of the simulation models;

[0017] Step S15: quantifying the key values ​​of the table structure points of the simulation model based on the material complexity of each area to obtain the key values ​​of the table structure points.

[0018] The present invention understands the material complexity of different regions in the model by performing regional material complexity identification on the simulation model to be processed, which is helpful for subsequent data lightweight processing and optimization, obtains the material complexity of each region to provide a basis for subsequent geometric structure point identification and table structure key quantity quantification, performs topological cutting on the simulation model to divide the model into multiple independent parts, which is convenient for subsequent processing and analysis, obtains multiple cutting simulation models to independently process each part, improves the efficiency and flexibility of data processing, performs section contour identification on multiple cutting simulation models to extract the section contour line of each model, obtains the geometric shape information of the model, generates section contour lines to provide a basis for subsequent geometric structure point identification and table structure key quantity quantification, performs geometric structure point identification on multiple cutting simulation models based on the section contour lines to identify geometric structure points in the model, such as corner points, boundary points, etc., marks the geometric structure points of the simulation model to provide accurate key point positions for subsequent table structure key quantity quantification, performs table structure key quantity quantification on the geometric structure points of the simulation model based on the material complexity of each region to determine which geometric structure points are more important to the surface structure, obtains the key values ​​of the table structure points to provide a basis for subsequent data lightweight processing, gives priority to retaining the data of the key points, and reduces the amount of unimportant data.

[0019] Preferably, step S2 comprises the following steps:

[0020] Step S21: Simplify the structure of the simulation model to be processed based on the key values ​​of the table structure points to obtain a simplified simulation model;

[0021] Step S22: performing regional division on the simplified simulation model to obtain a regional division grid;

[0022] Step S23: Calculate the dynamic frame density of the area division grid to obtain the dynamic frame density of different areas;

[0023] Step S24: performing regional dynamic frame priority analysis on the simplified simulation model based on the dynamic frame densities of different regions, thereby obtaining a high dynamic frame regional model and a low dynamic frame regional model.

[0024] The present invention reduces redundant information and complexity in the model by performing structural simplification processing on the simulation model to be processed based on the key values ​​of the table structure points, improves the efficiency of data processing and transmission, obtains a simplified simulation model, reduces the data volume and calculation load of the model, adapts to the environment with limited resources, such as mobile devices or network transmission, divides the simplified simulation model into different areas, improves the efficiency of data processing and rendering, obtains the area division grid, independently processes each area, realizes parallel calculation and rendering, improves performance and effect, calculates the dynamic frame density of the area division grid, determines the density of the dynamic frame according to the characteristics and requirements of each area, realizes fine-grained control of different areas, obtains the dynamic frame density of different areas, flexibly processes and optimizes each area according to the difference in dynamic frame density, realizes data lightweight and performance improvement, performs regional dynamic frame priority analysis on the simplified simulation model based on the dynamic frame density of different areas, determines which areas need higher dynamic frame density, improves the details and realism of key areas, obtains a high dynamic frame area model and a low dynamic frame area model, selects to use a high dynamic range or a low dynamic range model according to different requirements, and realizes flexible application and lightweight processing of data.

[0025] Preferably, the specific steps of step S21 are:

[0026] The key value of the table structure point is compared based on the preset feature point threshold. When the preset feature point threshold is greater than the key value of the table structure point, it is determined to be a non-key geometric structure point;

[0027] When the preset feature point threshold is less than the key value of the table structure point, it is determined to be a key geometric structure point;

[0028] Set restricted areas for key geometric structure points to obtain key structure protection areas;

[0029] Identify the locations of non-critical geometric structure points, perform structural deletion processing on the simulation model to be processed based on the locations of non-critical geometric structure points and critical structure protection areas, and obtain an optimized simulation model;

[0030] Perform model boundary detection on the optimized simulation model to obtain model boundary topological cracks;

[0031] Perform surface reshaping and anti-aliasing processing based on the topological cracks at the model boundary to obtain a reshaped simulation model;

[0032] Calculating the structural error of the simulation model to be processed based on the reshaped simulation model, thereby obtaining the model structural error;

[0033] The reshaped simulation model is locally refined and optimized according to the model structure error to obtain a simplified simulation model.

[0034] The present invention compares the key values ​​of table structure points based on a preset feature point threshold value to divide geometric structure points into two categories: key and non-key, so as to facilitate subsequent processing and optimization, determine non-key geometric structure points to identify points with less impact on the model, and provide a basis for data lightweight processing, set restricted areas for key geometric structure points to determine the protection area of ​​the key structure to ensure that it is not over-simplified or deleted, and obtain the key structure protection area to retain important geometric details in subsequent processing, improve the fidelity and quality of the data, identify the position of non-key geometric structure points to determine the position information of non-key points in the simulation model to be processed, and provide an accurate target for subsequent structure deletion processing, perform structure deletion processing on the simulation model to be processed based on the position of non-key geometric structure points and the key structure protection area to remove non-key points, reduce the data volume and complexity of the model, perform model boundary detection on the optimized simulation model to identify topological cracks on the model boundary, that is, the existence of discontinuous or incorrectly connected boundary parts The model boundary topological cracks are obtained to provide key information for subsequent surface reshaping and anti-aliasing processing, ensure the integrity and quality of the model boundary, and perform surface reshaping and anti-aliasing processing based on the model boundary topological cracks to repair or improve the discontinuity and jagged edges of the model surface, improve the appearance quality of the model, and obtain a reshaped simulation model to make the model surface smoother and more realistic, improve the visual effect and rendering quality, and perform structural error calculation on the processed simulation model based on the reshaped simulation model to evaluate the structural differences and errors between the reshaped model and the original model. The model structural error is obtained to provide quantitative guidance for subsequent local refinement and optimization, ensure the structural consistency between the optimized model and the original model, and perform local refinement and optimization on the reshaped simulation model according to the model structural error. Further refinement and adjustment are performed on the areas with large errors to improve the accuracy and quality of the model, and a simplified simulation model is obtained to minimize the structural error while maintaining the overall lightweight of the model, providing a high-quality simplified model.

[0035] Preferably, the specific steps of step S3 are:

[0036] Step S31: Calculate the dynamic resolution of the high dynamic frame region model and the low dynamic frame region model to obtain the dynamic frame resolution;

[0037] Step S32: performing dynamic behavior detail analysis on the high dynamic frame region model based on the dynamic frame resolution to obtain high dynamic frame behavior detail data;

[0038] Step S33: performing dynamic detail rendering on the high dynamic frame region model based on the high dynamic frame behavior detail data to obtain a frame detail rendering model;

[0039] Step S34: performing spline interpolation resampling processing on the low dynamic frame region model based on the dynamic frame resolution to generate a resampled model.

[0040] The present invention calculates the dynamic resolution of the high dynamic frame area model and the low dynamic frame area model, determines the appropriate resolution according to the dynamic frame density of each area, realizes dynamic resolution adaptation, and obtains the dynamic frame resolution to provide guidance for subsequent rendering and resampling processing, ensuring the balance of details and performance in different areas. Based on the dynamic frame resolution, the high dynamic frame area model is subjected to dynamic behavior detail analysis to identify important behavior details in the high dynamic frame area, such as dynamic objects, particle effects, etc., and the high dynamic frame behavior detail data is obtained to provide key information for subsequent dynamic detail rendering, ensuring the realism and dynamics of the high dynamic frame area. Dynamic detail rendering is performed on the high dynamic frame area model. Special rendering processing is performed on the behavioral details in the high dynamic frame area to highlight the details and dynamic effects. The frame detail rendering model retains the details and characteristics of the high dynamic frame area while keeping the overall model lightweight, thereby improving the visual effect and sense of reality. Spline interpolation resampling processing is performed on the low dynamic frame area model based on the dynamic frame resolution. The low dynamic frame area model is resampled to an appropriate resolution according to the dynamic frame resolution to reduce data redundancy and distortion. The resampled model is generated. While keeping the overall model lightweight, the appearance quality and smoothness of the low dynamic frame area are guaranteed, thereby improving the rendering effect and performance.

[0041] Preferably, the specific steps of step S34 are:

[0042] Based on the dynamic frame resolution, the low dynamic frame area model is used to identify the motion displacement and obtain the dynamic displacement curve;

[0043] Extract sampling points from the dynamic displacement curve to obtain a low-frame point displacement curve;

[0044] Performing original frame posture change analysis on the low-dynamic frame region model to obtain frame posture change data;

[0045] Perform spline interpolation resampling based on the frame posture change data to obtain low-frame posture data;

[0046] Based on the low-frame posture data and low-frame point displacement curve, the low-frame dynamic frame area model is reconstructed to generate a resampling model.

[0047] The present invention analyzes the dynamic displacement in the low dynamic frame area by performing action displacement recognition on the low dynamic frame area model based on the dynamic frame resolution, that is, the position change of the object in time, and obtains the dynamic displacement curve to provide the motion information of the low dynamic frame area, providing a basis for the subsequent reconstruction and resampling processing, extracts the sampling points of the dynamic displacement curve to extract a part of discrete sampling points from the continuous dynamic displacement curve, reduces the redundancy and complexity of the data, obtains the low frame point displacement curve as the displacement information of the low dynamic frame area model in time, and provides key data for the subsequent reconstruction and resampling, performs the original frame posture change analysis on the low dynamic frame area model to identify the posture change of the model in the low dynamic frame area, that is, the morphological change of the object in space, and obtains the frame The posture change data provides the morphological information of the low-dynamic frame area and provides key data for subsequent reconstruction and resampling processing. Spline interpolation resampling is performed based on the frame posture change data. Data interpolation and resampling are performed according to the frame posture change data to fill in the missing or discontinuous parts of the low-dynamic frame area model and obtain low-frame posture data. The morphological changes of the low-dynamic frame area model are made smoother and continuous, thereby improving the appearance quality and realism of the model. Low-frame reconstruction is performed on the low-dynamic frame area model based on the low-frame posture data and the low-frame point displacement curve. The posture and morphological changes of the model are adjusted according to the low-frame posture data and the point displacement curve to generate a resampled model. While keeping the overall lightweight of the model, the morphological and dynamic information of the low-dynamic frame area is retained, thereby improving the appearance quality and realism of the model.

[0048] Preferably, the specific steps of step S4 are:

[0049] Step S41: performing floating point encoding on the frame detail rendering model to obtain a floating point encoded bit stream;

[0050] Step S42: performing coding offset position identification on the floating point coding bit stream to obtain frame coding offset data;

[0051] Step S43: encoding and compressing the floating point coding bit stream based on the frame coding offset data to generate a frame detail coding sequence;

[0052] Step S44: calculating the discrete frequency of posture change of the resampling model to obtain conditional entropy;

[0053] Step S45: constructing a Huffman coding table based on conditional entropy;

[0054] Step S46: Perform traversal entropy coding on the resampling model based on the Huffman coding table to generate an entropy coding sequence.

[0055] The present invention converts the model data represented by floating-point numbers into binary coding form by performing floating-point coding on the frame detail rendering model, thereby reducing the storage space and transmission cost of the data, obtaining a floating-point coded bit stream to represent the data of the frame detail rendering model in the form of a bit stream, providing input for subsequent coding and compression processing, performing coding offset position identification on the floating-point coded bit stream to analyze the position information of different codes in the bit stream, that is, identifying the start and end positions of the codes, obtaining frame coding offset data to provide the position information of the coded data, providing key data for subsequent coding, compression and decoding, encoding and compressing the floating-point coded bit stream based on the frame coding offset data, compressing the bit stream according to the coding offset data, reducing redundant data and coding length, generating a frame detail coding sequence to represent the data of the frame detail rendering model in a more compact manner, and further reducing the storage space of the data. and transmission cost, calculate the discrete frequency of posture changes of the resampling model, analyze the frequency of posture changes, that is, the degree of change of the model under different postures, and obtain the conditional entropy to measure the amount of information of the resampling model in posture changes, which provides a basis for subsequent encoding and compression, and constructs a Huffman coding table based on the conditional entropy. According to the size of the conditional entropy, a Huffman tree and a coding table are constructed, so that symbols with higher frequencies obtain shorter codes. The constructed Huffman coding table provides coding rules for subsequent entropy coding, improves coding efficiency and compression ratio, and performs traversal entropy coding on the resampling model based on the Huffman coding table. Encode the data of the resampling model according to the Huffman coding table, represent common symbols with shorter codes, and generate entropy coding sequences to represent the data of the resampling model in a more compact way, further reducing the storage space and transmission cost of the data, and realizing lightweight processing of data.

[0056] Preferably, the specific steps of step S5 are:

[0057] Step S51: extracting multimodal parameters of the simulation model to be processed to obtain multi-source characteristic parameters of the simulation model;

[0058] Step S52: performing heterogeneous parameter fusion processing on multi-source characteristic parameters of the simulation model to obtain heterogeneous fusion parameters of the simulation model;

[0059] Step S53: performing attribute quantization compression coding on the heterogeneous fusion parameters of the simulation model to obtain feature parameter compression coding.

[0060] The present invention extracts multiple feature parameters from different data sources by performing multimodal parameter extraction on the simulation model to be processed, including parameters in terms of morphology, texture, motion, etc., to obtain multi-source feature parameters of the simulation model. A parameter set of multiple different features is obtained to provide input for subsequent heterogeneous parameter fusion and compression coding, and heterogeneous parameter fusion processing is performed on the multi-source feature parameters of the simulation model. Feature parameters from different data sources are fused to generate comprehensive heterogeneous feature parameters to obtain heterogeneous fusion parameters of the simulation model. Multiple feature parameters are represented in a comprehensive manner, and contributions of different features are comprehensively considered to improve the comprehensiveness and integrity of parameter expression. Attribute quantization compression coding is performed on the heterogeneous fusion parameters of the simulation model. Parameters are quantized, and continuous parameter values ​​are converted into discrete coding representations to obtain feature parameter compression coding. The redundancy and complexity of parameter data are reduced, and the parameters are represented in a more compact manner, which reduces storage space and transmission costs, and realizes lightweight data processing.

[0061] Preferably, the specific steps of step S6 are:

[0062] Step S61: Performing structural integration processing on the entropy coding sequence and the frame detail coding sequence to obtain the simulation model integrated structural coding;

[0063] Step S62: Adaptively optimize the encoding parameters of the feature parameter compression encoding and the simulation model integrated structure encoding to obtain an adaptive encoding fine-tuning strategy;

[0064] Step S63: Based on the adaptive coding fine-tuning strategy, the simulation model to be processed is stored in a hierarchical coding manner, and a hierarchical coding simulation model is constructed to perform lightweight processing of the simulation model.

[0065] The present invention integrates the entropy coding sequence and the frame detail coding sequence by structural integration processing to form a unified coding structure, and obtains the simulation model integrated structure coding, which combines the entropy coding sequence and the frame detail coding sequence into a whole coding result, improves the consistency and integrity of the coding, and performs adaptive coding parameter optimization on the feature parameter compression coding and the simulation model integrated structure coding, optimizes and adjusts the coding parameters according to the characteristics of the data and the coding requirements, obtains an adaptive coding fine-tuning strategy to obtain a more adaptable coding parameter setting, improves the coding efficiency and compression ratio, and performs hierarchical coding storage on the simulation model to be processed based on the adaptive coding fine-tuning strategy. According to the adaptive coding strategy, the simulation model data is divided into different levels for coding and storage, and a hierarchical coding simulation model is constructed to organize and store the simulation model data by level, so as to realize hierarchical management and lightweight processing of data, and further reduce storage space and transmission costs.

[0066] In this specification, a data lightweight processing system for a simulation model is provided, which is used to execute the data lightweight processing method for the simulation model as described above, including:

[0067] A key quantification module is used to obtain the material complexity of each area based on the simulation model to be processed; based on the material complexity of each area, the table structure key quantification is performed on the geometric structure points of the simulation model to obtain the key values ​​of the table structure points;

[0068] A structure simplification module is used to simplify the structure of the simulation model to be processed based on the key values ​​of the table structure points, so as to obtain a high dynamic frame area model and a low dynamic frame area model;

[0069] A detail rendering module, used for performing dynamic detail rendering on a high dynamic frame region model and a low dynamic frame region model to obtain a frame detail rendering model and a resampling model;

[0070] The step-by-step encoding module is used to perform floating-point encoding on the frame detail rendering model to generate a frame detail encoding sequence; perform discrete frequency calculation on the posture change of the resampling model to generate an entropy encoding sequence;

[0071] A heterogeneous parameter encoding module is used to extract multi-source characteristic parameters of a simulation model based on a simulation model to be processed; heterogeneous parameter encoding is performed on the multi-source characteristic parameters of the simulation model to obtain characteristic parameter compression encoding;

[0072] The coding fine-tuning module is used to adaptively optimize the coding parameters of the entropy coding sequence, the frame detail coding sequence and the characteristic parameter compression coding to obtain an adaptive coding fine-tuning strategy; based on the adaptive coding fine-tuning strategy, the simulation model lightweight processing operation is executed.

[0073] The present invention obtains the material complexity of each area based on the simulation model to be processed, and quantitatively evaluates the complexity of different areas by analyzing the material properties and structural characteristics of the model to obtain the material complexity of each area, and provides a quantitative measurement of the complexity of different areas in the model, so as to provide a basis for subsequent key quantification, and performs table structure key quantification on the geometric structure points of the simulation model based on the material complexity of each area. According to the distribution of material complexity, the structural points in the model are key quantified to determine which points play a key role in the structure of the model, and the key values ​​of the table structure points are obtained to obtain the key point information of the model structure, so as to provide a basis for subsequent structural simplification, and dynamic detail rendering is performed on the high dynamic frame area model and the low dynamic frame area model. According to the dynamic properties of the model, the high dynamic area and the low dynamic area are subjected to detail rendering processing to enhance the realism and detail expression of the model, and a frame detail rendering model and a resampled model are obtained to obtain a model after detail rendering processing, so as to improve the visual quality and realism of the model, and the frame detail rendering model is subjected to floating point encoding to convert the model data represented by floating point numbers into a coded representation to reduce the storage space and transmission cost of the data, and the posture change discretization is performed on the resampled model. Frequency calculation calculates the discrete frequency according to the posture change of the model, provides input for the subsequent entropy coding, and obtains the frame detail coding sequence and the entropy coding sequence. The model data is represented in a coded manner, which compresses the data representation. The multi-source feature parameters of the simulation model are extracted based on the simulation model to be processed. A variety of feature parameters are extracted from the model, including parameters in terms of morphology, texture, motion, etc., and heterogeneous parameter encoding is performed on the multi-source feature parameters of the simulation model. The parameters of different features are encoded and represented in a unified manner, which improves the comprehensiveness and integrity of the parameters, and the feature parameter compression coding is obtained. The redundancy and complexity of parameter data are reduced, and lightweight processing of data is realized. The entropy coding sequence, frame detail coding sequence and feature parameter compression coding are adaptively optimized. According to the characteristics of the data and the coding requirements, the coding parameter settings are optimized to improve the coding efficiency and compression ratio, and the adaptive coding fine-tuning strategy is obtained to obtain a more adaptive coding parameter configuration, further improve the coding performance and data compression effect, and execute the simulation model lightweight processing operation. The simulation model is lightweight processed according to the adaptive coding fine-tuning strategy to reduce the storage and transmission costs of data and realize efficient data processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 A schematic diagram of the steps of a data lightweight processing method for a simulation model of the present invention;

[0075] Figure 2 Detailed implementation flow chart of step S1;

[0076] Figure 3 Detailed implementation flow chart of step S2;

[0077] Figure 4 Detailed implementation flow chart of step S3. DETAILED DESCRIPTION

[0078] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0079] The present application example provides a data lightweight processing method and system for a simulation model. The execution subject of the data lightweight processing method and system for the simulation model includes but is not limited to: mechanical equipment, data processing platform, cloud server node, network upload device, etc. equipped with the system can be regarded as the general computing node of the present application, and the data processing platform includes but is not limited to: at least one of an audio image management system, an information management system, and a cloud data management system.

[0080] See also Figures 1 to 4 The present invention provides a data lightweight processing method for a simulation model, and the data lightweight processing method for a simulation model comprises the following steps:

[0081] Step S1: obtaining the material complexity of each region based on the simulation model to be processed; quantifying the key values ​​of the table structure of the geometric structure points of the simulation model based on the material complexity of each region to obtain the key values ​​of the table structure points;

[0082] Step S2: Simplifying the structure of the simulation model to be processed based on the key values ​​of the table structure points, thereby obtaining a high dynamic frame region model and a low dynamic frame region model;

[0083] Step S3: performing dynamic detail rendering on the high dynamic frame region model and the low dynamic frame region model to obtain a frame detail rendering model and a resampling model;

[0084] Step S4: performing floating point encoding on the frame detail rendering model to generate a frame detail encoding sequence; performing posture change discrete frequency calculation on the resampling model to generate an entropy encoding sequence;

[0085] Step S5: extracting multi-source characteristic parameters of the simulation model based on the simulation model to be processed; performing heterogeneous parameter encoding on the multi-source characteristic parameters of the simulation model to obtain characteristic parameter compression encoding;

[0086] Step S6: Adaptively optimize the encoding parameters of the entropy coding sequence, the frame detail coding sequence and the feature parameter compression coding to obtain an adaptive coding fine-tuning strategy; and perform a simulation model lightweight processing operation based on the adaptive coding fine-tuning strategy.

[0087] The present invention obtains the material complexity of each area based on the simulation model to be processed to understand the material complexity of different areas in the model, which is helpful for subsequent structural simplification and data lightweight processing. Based on the material complexity of each area, the table structure key value of the simulation model geometric structure points is quantified to determine which geometric structure points are more important to the surface structure. The simulation model to be processed is structurally simplified based on the key value of the table structure point to reduce redundant information and complexity in the model. The high dynamic frame area model and the low dynamic frame area model select the model with high dynamic range or low dynamic range according to different needs to achieve flexible application and lightweight processing of data. Dynamic detail rendering is performed on the high dynamic frame area model and the low dynamic frame area model to increase the realism and details of the model and improve the visual effect. The frame detail rendering model and the resampling model select different models according to different needs and scenarios to achieve flexible rendering and resampling. The frame detail rendering model is floating-point encoded to compress and encode the data of the rendering model. Reduce the space required for storage and transmission, calculate the discrete frequency of posture changes of the resampled model to extract the posture change information of the model, generate frame detail coding sequence and entropy coding sequence to further compress and encode the data of the model, realize lightweight processing and optimization of data, extract multi-source feature parameters of the simulation model based on the simulation model to be processed, extract various feature information of the model, including geometric shape, material properties, etc., perform heterogeneous parameter encoding on the multi-source feature parameters of the simulation model, uniformly encode different types of parameters, reduce the amount of data stored and transmitted, perform adaptive coding parameter optimization on entropy coding sequence, frame detail coding sequence and feature parameter compression coding, adjust the coding parameters according to the characteristics and requirements of the data, further optimize the data compression effect, fine-tune the coding according to different situations and requirements, improve the compression rate and effect of the data, execute the simulation model lightweight processing operation, perform data lightweight processing on the simulation model according to the adaptive coding fine-tuning strategy, reduce the amount of data stored and transmitted, and improve processing efficiency and performance.

[0088] In the embodiment of the present invention, refer to Figure 1 , is a schematic flow chart of the steps of a data lightweight processing method for a simulation model of the present invention. In this example, the steps of the data lightweight processing method for the simulation model include:

[0089] Step S1: obtaining the material complexity of each region based on the simulation model to be processed; quantifying the key values ​​of the table structure of the geometric structure points of the simulation model based on the material complexity of each region to obtain the key values ​​of the table structure points;

[0090] In this embodiment, the simulation model data to be processed is read, including geometric structure, material information, etc., and the simulation model is partitioned into several areas. The partitioning is performed according to factors such as the complexity of the geometric structure and the similarity of material characteristics. The material information of each area is analyzed, and the material complexity index of the area is calculated. The material complexity takes into account factors such as the texture complexity, normal map complexity, and reflection property complexity of the material. The area where the point is located is determined, and the material complexity index of the area is obtained. According to the geometric characteristics of the point (such as curvature, normal change, etc.) and the material complexity of the area, the table structure criticality of the point is calculated. The table structure criticality reflects the importance of the point in the entire model. The table structure criticality value of each geometric structure point is obtained, and the key points of the geometric structure of the entire simulation model can be quantified.

[0091] Step S2: Simplifying the structure of the simulation model to be processed based on the key values ​​of the table structure points, thereby obtaining a high dynamic frame region model and a low dynamic frame region model;

[0092] In this embodiment, points with key values ​​higher than a preset threshold are classified as high dynamic frame areas, and points with key values ​​lower than a preset threshold are classified as low dynamic frame areas. The complete geometric structure of the high dynamic frame area is retained without simplification. The mesh topology of the high dynamic frame area is further optimized to improve the geometric quality. Polygon simplification, LOD and other technologies are used to downsample and simplify the geometric structure of the low dynamic frame area. According to different simplification algorithms and parameters, multiple simplified model versions are generated, and the simplified geometric details are retained to avoid visual distortion caused by over-simplification. For simplified points in the low dynamic frame area model, different simplification methods are used, such as mesh simplification algorithms or LOD (level of detail) technology, to simplify the connection relationship of these points. The simplification method is selected according to the task requirements and actual conditions, and it is necessary to balance the geometric shape and performance requirements of the model. The high dynamic frame area model that retains the complete structure and the simplified low dynamic frame area model are combined to form a final simulation model with simplified structure.

[0093] Step S3: performing dynamic detail rendering on the high dynamic frame region model and the low dynamic frame region model to obtain a frame detail rendering model and a resampling model;

[0094] In this embodiment, dynamic details are rendered on the high dynamic frame area model, and the low dynamic frame area model is further geometrically simplified using polygon simplification, LOD and other technologies. Dynamic texture mapping, normal mapping and other technologies are used to improve material details while maintaining geometric simplicity, and a resampled low dynamic frame area model is generated, which has low geometric complexity but retained details. The high dynamic frame area model rendered with dynamic details and the resampled low dynamic frame area model are combined to form a final dynamic detail rendering model.

[0095] Step S4: performing floating point encoding on the frame detail rendering model to generate a frame detail encoding sequence; performing posture change discrete frequency calculation on the resampling model to generate an entropy encoding sequence;

[0096] In this embodiment, the geometry, material, lighting and other detail information of the high dynamic frame area model is encoded as floating-point data, and an efficient floating-point encoding method, such as half-precision floating point and adaptive quantization, is adopted to compress the data volume to the maximum extent while ensuring the visual quality, generate a frame detail encoding sequence for subsequent data transmission and storage, and perform discrete frequency calculation of posture changes on the resampled model. By analyzing the changes of the model at different times or postures, the frequency or discreteness of the posture changes is calculated. The method includes using a rotation matrix, quaternion or other posture representation method to calculate the degree of posture change of the model in a continuous time period, and organizing the results of the discrete frequency calculation of the posture change into a sequence to form an entropy coding sequence. The sequence is a one-dimensional array or other data structure, and the posture change frequency information of the model is arranged in a specific order.

[0097] Step S5: extracting multi-source characteristic parameters of the simulation model based on the simulation model to be processed; performing heterogeneous parameter encoding on the multi-source characteristic parameters of the simulation model to obtain characteristic parameter compression encoding;

[0098] In this embodiment, the characteristic parameters of the simulation model are analyzed from the aspects of geometry, material, and lighting, including vertex coordinates, normals, texture coordinates, material properties, lighting parameters, etc. These characteristic parameters come from different data sources and show heterogeneous characteristics. Different encoding methods are used according to the data type, range, distribution characteristics, etc. of different characteristic parameters, such as floating-point compression encoding for geometric coordinates and discrete encoding for material properties. Corresponding compression algorithms are designed for different encoding methods, such as entropy coding, differential coding, vector quantization, etc., to finally generate characteristic parameter compression coding, which greatly reduces the data volume.

[0099] Step S6: Adaptively optimize the encoding parameters of the entropy coding sequence, the frame detail coding sequence and the feature parameter compression coding to obtain an adaptive coding fine-tuning strategy; and perform a simulation model lightweight processing operation based on the adaptive coding fine-tuning strategy.

[0100] In this embodiment, the characteristics and compression effect of each coding sequence are analyzed, and the coding parameters, such as quantization accuracy, entropy coding method, etc., are adjusted according to the target application scenario and performance requirements. The parameter configuration of each coding sequence is adaptively optimized by using machine learning and other technologies. The goal is to further reduce the overall data volume while ensuring visual quality. For different application scenarios and device performance, the optimal coding parameter configuration is formulated, and these coding parameter configurations are packaged into an adaptive coding fine-tuning strategy. This strategy can be used to guide the subsequent lightweight processing of the simulation model, compress and encode the entropy coding sequence, frame detail coding sequence and feature parameters generated previously, and further optimize the coding parameters according to the adaptive coding fine-tuning strategy, and finally output the lightweight processed simulation model data packet.

[0101] In this embodiment, refer to Figure 2 , is a flowchart of detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include:

[0102] Step S11: performing regional material complexity identification on the simulation model to be processed to obtain the material complexity of each region;

[0103] Step S12: performing topological cutting on the simulation model to be processed to obtain multiple cut simulation models;

[0104] Step S13: performing section contour recognition on a plurality of cutting simulation models, thereby generating section contour lines;

[0105] Step S14: identifying geometric structure points of multiple cutting simulation models based on the section contour lines, and marking the geometric structure points of the simulation models;

[0106] Step S15: quantifying the key values ​​of the table structure points of the simulation model based on the material complexity of each area to obtain the key values ​​of the table structure points.

[0107] In this embodiment, the material attribute information of the simulation model to be processed is analyzed, including texture mapping, material parameters, etc., and the model is divided into regions according to the complexity of the material attributes. The material complexity of each region is evaluated, such as the fineness of the texture mapping, the complexity of the material parameters, etc., to obtain the material complexity evaluation results of each region. According to the geometric structure and material characteristics of the model, a reasonable cutting position and method are determined, and the model is topologically cut to generate multiple cut simulation model sub-blocks to ensure the geometric integrity and material consistency of each sub-block after cutting. For each cut sub-block, the contour line information of its section is extracted, and image processing technology, such as edge detection, is used. , contour extraction, etc., obtain the 2D contour line of the section, map the extracted 2D contour line to the 3D space, and obtain the 3D contour line of the section; based on the section contour line, identify the geometric structure feature points of each sub-block, such as corner points, edge points, surface inflection points and other key geometric features, mark the position information of these geometric structure points in the model, and quantify the key values ​​of the geometric structure points according to the material complexity of the area; use numerical measurement, weight allocation or other methods to calculate the key values ​​of the geometric structure points; and calculate the key values ​​of the table structure points according to the results of the key quantization, which is the weighted average, maximum value or other defined calculation method of the key quantization results of each area.

[0108] In this embodiment, refer to Figure 3 , is a flowchart of detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include:

[0109] Step S21: Simplify the structure of the simulation model to be processed based on the key values ​​of the table structure points to obtain a simplified simulation model;

[0110] Step S22: performing regional division on the simplified simulation model to obtain a regional division grid;

[0111] Step S23: Calculate the dynamic frame density of the area division grid to obtain the dynamic frame density of different areas;

[0112] Step S24: performing regional dynamic frame priority analysis on the simplified simulation model based on the dynamic frame densities of different regions, thereby obtaining a high dynamic frame regional model and a low dynamic frame regional model.

[0113] In this embodiment, the structure points of the simulation model are sorted, and the structure points with lower critical values ​​are arranged in front, and the structure points with higher critical values ​​are arranged in the back. Starting from the sorted structure point list, the structure points with lower critical values ​​are gradually removed, and the structure of the model is simplified using simplification algorithms, such as edge collapse, vertex merging, etc. After removing a structure point each time, the topological structure and geometric features of the model are recalculated, and the geometric structure points with higher criticality are retained to ensure the details of the key areas. The geometric topological structure and material characteristics of the simplified model are analyzed, and the model is divided into regions according to the characteristics of geometric complexity, material complexity, etc., to generate representations of different regions. The grid data of the domain is used to perform regional dynamic frame priority analysis on the simplified simulation model. According to the dynamic frame density, the high dynamic frame area and the low dynamic frame area are determined. Each area is divided into grids for dynamic frame density calculation, including analyzing the degree of geometric change, motion trajectory, point cloud density, etc. in the area to evaluate the dynamic frame density. According to the results of the regional dynamic frame priority analysis, the relevant structures and features belonging to the high dynamic frame area in the simplified simulation model are extracted to generate a high dynamic frame regional model. According to the results of the regional dynamic frame priority analysis, the relevant structures and features belonging to the low dynamic frame area in the simplified simulation model are extracted to generate a low dynamic frame regional model.

[0114] In this embodiment, the specific steps of step S21 are:

[0115] The key value of the table structure point is compared based on the preset feature point threshold. When the preset feature point threshold is greater than the key value of the table structure point, it is determined to be a non-key geometric structure point;

[0116] When the preset feature point threshold is less than the key value of the table structure point, it is determined to be a key geometric structure point;

[0117] Set restricted areas for key geometric structure points to obtain key structure protection areas;

[0118] Identify the locations of non-critical geometric structure points, perform structural deletion processing on the simulation model to be processed based on the locations of non-critical geometric structure points and critical structure protection areas, and obtain an optimized simulation model;

[0119] Perform model boundary detection on the optimized simulation model to obtain model boundary topological cracks;

[0120] Perform surface reshaping and anti-aliasing processing based on the topological cracks at the model boundary to obtain a reshaped simulation model;

[0121] Calculating the structural error of the simulation model to be processed based on the reshaped simulation model, thereby obtaining the model structural error;

[0122] The reshaped simulation model is locally refined and optimized according to the model structure error to obtain a simplified simulation model.

[0123] In this embodiment, according to the needs and preset conditions, a feature point threshold is set to judge the key geometric structure points and non-key geometric structure points. The threshold is a value or a range. The key value of each table structure point is compared with the feature point threshold. If the preset feature point threshold is less than the key value of the table structure point, then the point is judged as a key geometric structure point. If the preset feature point threshold is greater than the key value of the table structure point, then the point is judged as a non-key geometric structure point. Based on the position and characteristics of the key geometric structure points, a restriction area is set to protect the key structure. According to the setting of the restriction area, the area containing the key geometric structure points is demarcated as the key structure protection area, which will be retained in the optimized simulation model to maintain the integrity of the key structure. Based on the position of the non-key geometric structure points and the key structure protection area, the simulation model to be processed is subjected to structural deletion processing. In the area where the non-key geometric structure points are located outside the key structure protection area, structural deletion is performed to optimize the simulation model. By deleting the non-key geometric structure points, an optimized simulation model is obtained, which will retain the key structure and the key structure. Protect the area and remove non-critical structures at the same time. Perform model boundary detection on the optimized simulation model, identify the boundary position of the model, perform topological analysis at the boundary position of the model, and detect whether there is a topological crack. Topological crack refers to the discontinuous or missing topological structure at the boundary of the model. For the topological cracks at the model boundary, perform surface reshaping, including filling the cracks and repairing the missing topological structure, so that the model boundary becomes smooth and continuous in geometry and topology. In order to improve the visual quality of the model, the reshaped simulation model is anti-aliased. Anti-aliasing reduces the jagged edges and edge artifacts of the model boundary. Use the comparison method between the reshaped simulation model and the simulation model to be processed to calculate the structural error between them. The structural error is the difference in geometric features, the mismatch of topological structure, etc. According to the calculation results of the structural error, determine the area that needs to be refined and optimized. The refinement optimization includes increasing the subdivision level, adding additional geometric details, etc. Through the local refinement optimization of the reshaped simulation model, a simplified simulation model is obtained. The model reduces unnecessary details on the basis of maintaining structural features to improve computational efficiency and reduce storage space.

[0124] In this embodiment, refer to Figure 4 , is a flowchart of detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include:

[0125] Step S31: Calculate the dynamic resolution of the high dynamic frame region model and the low dynamic frame region model to obtain the dynamic frame resolution;

[0126] Step S32: performing dynamic behavior detail analysis on the high dynamic frame region model based on the dynamic frame resolution to obtain high dynamic frame behavior detail data;

[0127] Step S33: performing dynamic detail rendering on the high dynamic frame region model based on the high dynamic frame behavior detail data to obtain a frame detail rendering model;

[0128] Step S34: performing spline interpolation resampling processing on the low dynamic frame region model based on the dynamic frame resolution to generate a resampled model.

[0129] In this embodiment, the model to be processed is divided into a high dynamic frame area and a low dynamic frame area. The high dynamic frame area usually contains key parts or key details that require higher resolution. Dynamic resolution calculation is performed on the high dynamic frame area model and the low dynamic frame area model. Based on different factors, such as visual importance, motion complexity, etc., the appropriate resolution of each area is determined. According to the dynamic resolution calculation result, the dynamic frame resolution corresponding to each area is obtained. The high dynamic frame area will have a higher resolution, while the low dynamic frame area will have a lower resolution. Dynamic behavior detail analysis is performed on the high dynamic frame area model, including detecting and tracking key points, analyzing motion trajectories, extracting action features, etc. Through the dynamic behavior detail analysis, high dynamic frame behavior detail data is obtained. These data include key The coordinates, speed, acceleration and other information of the points, as well as other action-related features, are used to render dynamic details of the high dynamic frame area model using the high dynamic frame behavior detail data, including adding details around key points, adjusting textures, changing material properties, etc. Through dynamic detail rendering, a frame detail rendering model is obtained, which adds more details and realism to the high dynamic frame area to enhance the visual effects and the realism of dynamic behavior. The low dynamic frame area model is subjected to spline interpolation resampling processing. Spline interpolation is a commonly used interpolation method, which generates new data points by curve fitting between known data points. Through spline interpolation resampling processing, a resampling model is generated, which adds additional data points to the low dynamic frame area to improve the representation of details and the smoothness of dynamic behavior.

[0130] In this embodiment, the specific steps of step S34 are:

[0131] Based on the dynamic frame resolution, the low dynamic frame area model is used to identify the motion displacement and obtain the dynamic displacement curve;

[0132] Extract sampling points from the dynamic displacement curve to obtain a low-frame point displacement curve;

[0133] Performing original frame posture change analysis on the low-dynamic frame region model to obtain frame posture change data;

[0134] Perform spline interpolation resampling based on the frame posture change data to obtain low-frame posture data;

[0135] Based on the low-frame posture data and low-frame point displacement curve, the low-frame dynamic frame area model is reconstructed to generate a resampling model.

[0136] In this embodiment, motion displacement recognition is performed for the low dynamic frame region model. It includes tracking the position of specific key points in the model and calculating their displacement in time. The key frame is matched with the adjacent frame to track the position change of the feature point between different frames. It is implemented using a feature matching algorithm, such as an optical flow method or a feature point matching algorithm. According to the matched feature points, their displacement in time is calculated. The displacement size is estimated by calculating the distance change or pixel displacement between the feature points. Through motion displacement recognition, a dynamic displacement curve of the low dynamic frame region model is obtained. Sampling points are extracted from the dynamic displacement curve according to the sampling frequency. Equally spaced sampling is used or adaptive sampling is performed according to the change of the curve. Computer vision technology, such as a posture estimation algorithm or a model fitting algorithm, is used to estimate the posture of the object in the low dynamic frame. The position and posture information of the key point are obtained. The posture information between adjacent frames is compared and analyzed to obtain frame posture change data. The angle change, rotation matrix or posture transformation matrix between the key points are calculated. The frame posture change data is interpolated using a spline interpolation method, such as a spline curve or a spline surface fitting. Fill the missing values ​​between the data to obtain smooth low frame posture data. The low-frame pose data is obtained based on the interpolation results. These data reflect the pose changes on the low-dynamic frame and are used in the subsequent reconstruction process. According to the low-frame pose data, the pose of the low-dynamic frame region model is adjusted. The pose of the model is adjusted according to the angle of pose change or the rotation matrix to adapt it to the requirements of the low frame. According to the low-frame point displacement curve, the displacement of the low-dynamic frame region model is adjusted. According to the value of the displacement curve, the key points of the model are translated to reflect the changes in point displacement. The adjusted pose and displacement are applied to the low-dynamic frame region model to generate a resampled model. The model is pose- and displacement-adjusted on the low frame to adapt to the requirements of low dynamics.

[0137] In this embodiment, the specific steps of step S4 are:

[0138] Step S41: performing floating point encoding on the frame detail rendering model to obtain a floating point encoded bit stream;

[0139] Step S42: performing coding offset position identification on the floating point coding bit stream to obtain frame coding offset data;

[0140] Step S43: encoding and compressing the floating point coding bit stream based on the frame coding offset data to generate a frame detail coding sequence;

[0141] Step S44: calculating the discrete frequency of posture change of the resampling model to obtain conditional entropy;

[0142] Step S45: constructing a Huffman coding table based on conditional entropy;

[0143] Step S46: Perform traversal entropy coding on the resampling model based on the Huffman coding table to generate an entropy coding sequence.

[0144] In this embodiment, the floating-point values ​​in the frame detail rendering model are encoded and converted into a binary bit stream. A floating-point encoding algorithm is used, such as the floating-point representation of the IEEE 754 standard, to convert the floating-point numbers into binary codes to obtain a floating-point encoded bit stream, which represents the binary codes of the floating-point values ​​in the frame detail rendering model. The floating-point encoded bit stream is scanned to identify the encoding offset position therein, which refers to the position in the bit stream that represents the starting point of the encoding of the floating-point values. The identified encoding offset position is recorded to obtain frame encoding offset data, which represents the encoding starting point position of each floating-point value in the floating-point encoded bit stream. The floating-point encoded bit stream is compressed and encoded according to the frame encoding offset data. The floating-point encoded bit stream is compressed using an encoding algorithm, such as differential encoding, Huffman encoding, or arithmetic encoding, to obtain a compressed frame detail encoding sequence, which represents the compressed floating-point encoded bit stream. The posture change data in the resampling model is discretized and divided into different discrete states. A discretization method, such as a histogram uniform segmentation or a clustering algorithm, is used to divide the posture change data into multiple discrete states. According to the discretized posture change data, the conditional entropy is calculated. The conditional entropy refers to the uncertainty or information content of the model under a given posture change state. The entropy calculation method in information theory, such as Shannon entropy, is used to calculate the conditional entropy. According to the discretized posture change state, the frequency of occurrence of each state is counted to obtain a frequency table, and the frequency of each state is recorded. Based on the frequency table, a Huffman tree is constructed using the Huffman coding algorithm. The Huffman tree is a tree structure used to represent the encoding of different states. According to the Huffman tree, a Huffman coding table is generated. The Huffman coding table records the encoding corresponding to each discrete state and the number of bits of the encoding. The posture change data in the resampled model is traversed, and each discrete state is encoded into a bit sequence according to the Huffman coding table. During the traversal process, the corresponding Huffman code is found according to the current state, and the code is added to the entropy coding sequence to obtain an entropy coding sequence, which represents the posture change data of the resampled model after Huffman coding compression.

[0145] In this embodiment, the specific steps of step S5 are:

[0146] Step S51: extracting multimodal parameters of the simulation model to be processed to obtain multi-source characteristic parameters of the simulation model;

[0147] Step S52: performing heterogeneous parameter fusion processing on multi-source characteristic parameters of the simulation model to obtain heterogeneous fusion parameters of the simulation model;

[0148] Step S53: performing attribute quantization compression coding on the heterogeneous fusion parameters of the simulation model to obtain feature parameter compression coding.

[0149] In this embodiment, a variety of different types of feature parameters, geometric shape parameters, vertex coordinates, surface parameters, material parameters, color, texture, reflection characteristics, motion parameters, position, rotation, speed are extracted from the simulation model to obtain the result of multi-modal parameter extraction, that is, multi-source feature parameters of the simulation model. These parameters represent the feature information of the simulation model to be processed in different modes. Heterogeneous parameter fusion processing is performed on the multi-source feature parameters of the simulation model, including fusing the feature parameters from different modes to obtain a more comprehensive and consistent feature representation, and using fusion algorithms such as feature weighting, feature splicing, feature transformation, etc. to fuse the multi-source feature parameters to perform heterogeneous parameter fusion processing on the simulation model. The heterogeneous fusion parameters of the model are attribute quantized. Attribute quantization converts continuous parameter values ​​into discrete attribute values ​​to reduce the representation space of the parameters. Quantization algorithms, such as clustering and histogram segmentation, are used to discretize the heterogeneous fusion parameters. The quantized simulation model heterogeneous fusion parameters are encoded and compressed. Coding algorithms, such as Huffman coding and arithmetic coding, are used to encode and compress the discretized parameters to reduce the number of representation bits and storage space of the parameters. The compressed coding results, i.e., the compressed coding of the feature parameters, are obtained. These codes represent the heterogeneous fusion parameters of the simulation model after attribute quantization and compression coding, effectively reducing the storage and transmission costs of the parameters.

[0150] In this embodiment, the specific steps of step S6 are:

[0151] Step S61: Performing structural integration processing on the entropy coding sequence and the frame detail coding sequence to obtain the simulation model integrated structural coding;

[0152] Step S62: Adaptively optimize the encoding parameters of the feature parameter compression encoding and the simulation model integrated structure encoding to obtain an adaptive encoding fine-tuning strategy;

[0153] Step S63: Based on the adaptive coding fine-tuning strategy, the simulation model to be processed is stored in a hierarchical coding manner, and a hierarchical coding simulation model is constructed to perform lightweight processing of the simulation model.

[0154] In this embodiment, an entropy coding sequence and a frame detail coding sequence that need to be structurally integrated are prepared. The entropy coding sequence is a sequence obtained by entropy encoding the result of feature parameter compression encoding, and the frame detail coding sequence is a sequence obtained by encoding the detail information of the simulation model, such as frame changes, motion, texture, etc. The entropy coding sequence and the frame detail coding sequence are structurally integrated, including structurally integrating the two sequences to better represent the structural information of the simulation model, using appropriate algorithms and techniques, such as sequence fusion, structural transformation, etc., to integrate the entropy coding sequence and the frame detail coding sequence to obtain an integrated structural coding, and obtain the result after the structural integration processing, that is, the integrated structural coding of the simulation model, which represents the structure of the simulation model after the entropy coding sequence and the frame detail coding sequence are structurally integrated. Information, adaptively optimize the coding parameters of feature parameter compression coding and integrated structure coding of simulation model, including adjusting coding parameters, optimizing coding algorithms, etc., to improve coding efficiency and quality, and obtain the result after adaptive coding parameter optimization, that is, adaptive coding fine-tuning strategy. This strategy represents the method and parameter setting for fine-tuning feature parameter compression coding and integrated structure coding of simulation model. Based on the adaptive coding fine-tuning strategy, hierarchical coding and storage are performed on the simulation model to be processed, including hierarchical coding of different parts or levels of the simulation model to achieve lightweight processing and storage optimization, and the result after hierarchical coding storage is obtained, that is, a hierarchical coded simulation model is constructed. The hierarchical coded simulation model is stored and processed more efficiently, while maintaining important feature parameters and structure, to achieve efficient storage and lightweight processing of the simulation model.

[0155] In this embodiment, a data lightweight processing system for a simulation model is provided, which is used to execute the data lightweight processing method for the simulation model as described above, including:

[0156] A key quantification module is used to obtain the material complexity of each area based on the simulation model to be processed; based on the material complexity of each area, the table structure key quantification is performed on the geometric structure points of the simulation model to obtain the key values ​​of the table structure points;

[0157] A structure simplification module is used to simplify the structure of the simulation model to be processed based on the key values ​​of the table structure points, so as to obtain a high dynamic frame area model and a low dynamic frame area model;

[0158] A detail rendering module, used for performing dynamic detail rendering on a high dynamic frame region model and a low dynamic frame region model to obtain a frame detail rendering model and a resampling model;

[0159] The step-by-step encoding module is used to perform floating-point encoding on the frame detail rendering model to generate a frame detail encoding sequence; perform discrete frequency calculation on the posture change of the resampling model to generate an entropy encoding sequence;

[0160] A heterogeneous parameter encoding module is used to extract multi-source characteristic parameters of a simulation model based on a simulation model to be processed; heterogeneous parameter encoding is performed on the multi-source characteristic parameters of the simulation model to obtain characteristic parameter compression encoding;

[0161] The coding fine-tuning module is used to adaptively optimize the coding parameters of the entropy coding sequence, the frame detail coding sequence and the characteristic parameter compression coding to obtain an adaptive coding fine-tuning strategy; based on the adaptive coding fine-tuning strategy, the simulation model lightweight processing operation is executed.

[0162] The present invention obtains the material complexity of each area based on the simulation model to be processed, and quantitatively evaluates the complexity of different areas by analyzing the material properties and structural characteristics of the model to obtain the material complexity of each area, and provides a quantitative measurement of the complexity of different areas in the model, so as to provide a basis for subsequent key quantification, and performs table structure key quantification on the geometric structure points of the simulation model based on the material complexity of each area. According to the distribution of material complexity, the structural points in the model are key quantified to determine which points play a key role in the structure of the model, and the key values ​​of the table structure points are obtained to obtain the key point information of the model structure, so as to provide a basis for subsequent structural simplification, and dynamic detail rendering is performed on the high dynamic frame area model and the low dynamic frame area model. According to the dynamic properties of the model, the high dynamic area and the low dynamic area are subjected to detail rendering processing to enhance the realism and detail expression of the model, and a frame detail rendering model and a resampled model are obtained to obtain a model after detail rendering processing, so as to improve the visual quality and realism of the model, and the frame detail rendering model is subjected to floating point encoding to convert the model data represented by floating point numbers into a coded representation to reduce the storage space and transmission cost of the data, and the posture change discretization is performed on the resampled model. Frequency calculation calculates the discrete frequency according to the posture change of the model, provides input for the subsequent entropy coding, and obtains the frame detail coding sequence and the entropy coding sequence. The model data is represented in a coded manner, which compresses the data representation. The multi-source feature parameters of the simulation model are extracted based on the simulation model to be processed. A variety of feature parameters are extracted from the model, including parameters in terms of morphology, texture, motion, etc., and heterogeneous parameter encoding is performed on the multi-source feature parameters of the simulation model. The parameters of different features are encoded and represented in a unified manner, which improves the comprehensiveness and integrity of the parameters, and the feature parameter compression coding is obtained. The redundancy and complexity of parameter data are reduced, and lightweight processing of data is realized. The entropy coding sequence, frame detail coding sequence and feature parameter compression coding are adaptively optimized. According to the characteristics of the data and the coding requirements, the coding parameter settings are optimized to improve the coding efficiency and compression ratio, and the adaptive coding fine-tuning strategy is obtained to obtain a more adaptive coding parameter configuration, further improve the coding performance and data compression effect, and execute the simulation model lightweight processing operation. The simulation model is lightweight processed according to the adaptive coding fine-tuning strategy to reduce the storage and transmission costs of data and realize efficient data processing.

[0163] 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.

[0164] The above 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 lightweight processing of simulation model data, characterized in that: The following steps are involved: Step S1: obtaining the material complexity of each area based on the simulation model to be processed; Based on the material complexity of each area, the key values ​​of the table structure points of the simulation model are quantified to obtain the key values ​​of the table structure points; Step S2: Simplifying the structure of the simulation model to be processed based on the key values ​​of the table structure points, thereby obtaining a high dynamic frame region model and a low dynamic frame region model; Step S3: performing dynamic detail rendering on the high dynamic frame region model and the low dynamic frame region model to obtain a frame detail rendering model and a resampling model; Step S4: performing floating-point encoding on the frame detail rendering model to generate a frame detail encoding sequence; performing posture change discrete frequency calculation on the resampling model to generate an entropy encoding sequence; the posture change discrete frequency calculation is specifically as follows: using a discretization method to divide the posture change data into multiple discrete states, and calculating conditional entropy based on the discretized posture change data, where conditional entropy refers to the uncertainty or information content of the model under a given posture change state; Step S5: extracting multi-source characteristic parameters of the simulation model based on the simulation model to be processed; performing heterogeneous parameter encoding on the multi-source characteristic parameters of the simulation model to obtain characteristic parameter compression encoding; Step S6: Adaptively optimize the encoding parameters of the entropy encoding sequence, the frame detail encoding sequence and the feature parameter compression encoding to obtain an adaptive encoding fine-tuning strategy; perform a simulation model lightweight processing operation based on the adaptive encoding fine-tuning strategy; The specific steps of step S1 are: Step S11: performing regional material complexity identification on the simulation model to be processed to obtain the material complexity of each region; Step S12: performing topological cutting on the simulation model to be processed to obtain a plurality of cut simulation models; Step S13: performing section contour recognition on a plurality of cutting simulation models, thereby generating section contour lines; Step S14: identifying geometric structure points of multiple cutting simulation models based on the section contour lines, and marking the geometric structure points of the simulation models; Step S15: quantifying the key values ​​of the table structure points of the simulation model based on the material complexity of each region to obtain the key values ​​of the table structure points; the key value quantification of the table structure is specifically: quantifying the key values ​​of the geometric structure points according to the material complexity of the region, using numerical measurement, weight distribution or other methods to calculate the key values ​​of the geometric structure points, and calculating the key values ​​of the table structure points according to the results of the key value quantification; The specific steps of step S2 are: Step S21: Simplify the structure of the simulation model to be processed based on the key values ​​of the table structure points to obtain a simplified simulation model; the structure simplification process is specifically as follows: sort the structure points of the simulation model, put the structure points with lower key values ​​in front, and put the structure points with higher key values ​​in the back, start from the sorted structure point list, gradually remove the structure points with lower key values, use the simplification algorithm to simplify the structure of the model, and recalculate the topological structure and geometric features of the model after removing a structure point each time, and retain the geometric structure points with higher key values; Step S22: performing regional division on the simplified simulation model to obtain a regional division grid; Step S23: Calculate the dynamic frame density of the area division grid to obtain the dynamic frame density of different areas; Step S24: performing regional dynamic frame priority analysis on the simplified simulation model based on dynamic frame densities of different regions, thereby obtaining a high dynamic frame regional model and a low dynamic frame regional model; The specific steps of step S3 are: Step S31: Calculate the dynamic resolution of the high dynamic frame region model and the low dynamic frame region model to obtain the dynamic frame resolution; Step S32: performing dynamic behavior detail analysis on the high dynamic frame region model based on the dynamic frame resolution to obtain high dynamic frame behavior detail data; Step S33: performing dynamic detail rendering on the high dynamic frame region model based on the high dynamic frame behavior detail data to obtain a frame detail rendering model; Step S34: performing spline interpolation resampling processing on the low dynamic frame region model based on the dynamic frame resolution to generate a resampled model.

2. The method for lightweight processing of simulation model data according to claim 1, characterized in that: The specific steps of step S21 are: The key value of the table structure point is compared based on the preset feature point threshold. When the preset feature point threshold is greater than the key value of the table structure point, it is determined to be a non-key geometric structure point; When the preset feature point threshold is less than the key value of the table structure point, it is determined to be a key geometric structure point; Set restricted areas for key geometric structure points to obtain key structure protection areas; Identify the locations of non-critical geometric structure points, perform structural deletion processing on the simulation model to be processed based on the locations of non-critical geometric structure points and critical structure protection areas, and obtain an optimized simulation model; Perform model boundary detection on the optimized simulation model to obtain model boundary topological cracks; Perform surface reshaping and anti-aliasing processing based on the topological cracks at the model boundary to obtain a reshaped simulation model; Calculating the structural error of the simulation model to be processed based on the reshaped simulation model, thereby obtaining the model structural error; The reshaped simulation model is locally refined and optimized according to the model structure error to obtain a simplified simulation model.

3. The method for lightweight processing of simulation model data according to claim 1, characterized in that: The specific steps of step S34 are: Based on the dynamic frame resolution, the low dynamic frame area model is used to identify the motion displacement and obtain the dynamic displacement curve; Extract sampling points from the dynamic displacement curve to obtain a low-frame point displacement curve; Performing original frame posture change analysis on the low-dynamic frame region model to obtain frame posture change data; Perform spline interpolation resampling based on the frame posture change data to obtain low-frame posture data; Based on the low-frame posture data and low-frame point displacement curve, the low-frame dynamic frame area model is reconstructed to generate a resampling model.

4. The method for lightweight processing of simulation model data according to claim 1, characterized in that: The specific steps of step S4 are: Step S41: performing floating point encoding on the frame detail rendering model to obtain a floating point encoded bit stream; Step S42: performing coding offset position identification on the floating point coding bit stream to obtain frame coding offset data; Step S43: encoding and compressing the floating point coding bit stream based on the frame coding offset data to generate a frame detail coding sequence; Step S44: calculating the discrete frequency of posture change of the resampling model to obtain conditional entropy; Step S45: constructing a Huffman coding table based on conditional entropy; Step S46: Perform traversal entropy coding on the resampling model based on the Huffman coding table to generate an entropy coding sequence.

5. The method for lightweight processing of simulation model data according to claim 1, characterized in that: The specific steps of step S5 are: Step S51: extracting multimodal parameters of the simulation model to be processed to obtain multi-source characteristic parameters of the simulation model; Step S52: performing heterogeneous parameter fusion processing on multi-source characteristic parameters of the simulation model to obtain heterogeneous fusion parameters of the simulation model; Step S53: Quantize and compress the simulation model heterogeneous fusion parameters. Get the characteristic parameter compression code.

6. The method for lightweight processing of simulation model data according to claim 1, characterized in that: The specific steps of step S6 are: Step S61: Performing structural integration processing on the entropy coding sequence and the frame detail coding sequence to obtain the simulation model integrated structural coding; Step S62: Adaptively optimize the encoding parameters of the feature parameter compression encoding and the simulation model integrated structure encoding to obtain an adaptive encoding fine-tuning strategy; Step S63: Based on the adaptive coding fine-tuning strategy, the simulation model to be processed is stored in a hierarchical coding manner, and a hierarchical coding simulation model is constructed to perform lightweight processing of the simulation model.

7. A data lightweight processing system for a simulation model, characterized in that: The data lightweight processing method for executing the simulation model according to claim 1 comprises: A key quantification module is used to obtain the material complexity of each area based on the simulation model to be processed; based on the material complexity of each area, the table structure key quantification is performed on the geometric structure points of the simulation model to obtain the key values ​​of the table structure points; A structure simplification module is used to simplify the structure of the simulation model to be processed based on the key values ​​of the table structure points, so as to obtain a high dynamic frame area model and a low dynamic frame area model; A detail rendering module, used for performing dynamic detail rendering on a high dynamic frame region model and a low dynamic frame region model to obtain a frame detail rendering model and a resampling model; The step-by-step encoding module is used to perform floating-point encoding on the frame detail rendering model to generate a frame detail encoding sequence; perform discrete frequency calculation on the posture change of the resampling model to generate an entropy encoding sequence; A heterogeneous parameter encoding module is used to extract multi-source characteristic parameters of a simulation model based on a simulation model to be processed; heterogeneous parameter encoding is performed on the multi-source characteristic parameters of the simulation model to obtain characteristic parameter compression encoding; The coding fine-tuning module is used to adaptively optimize the coding parameters of the entropy coding sequence, the frame detail coding sequence and the feature parameter compression coding to obtain an adaptive coding fine-tuning strategy; based on the adaptive coding fine-tuning strategy, the simulation model lightweight processing operation is executed.

Images