Digital visual optimization method and system based on artificial intelligence

Abstract

The invention relates to the field of digital visualization, in particular to a digital visual optimization method and system based on artificial intelligence, and the method comprises the steps: obtaining the real-time operation data and model data of a target object, reading the geometric grid information of a model through a 3D engine, simplifying the three-dimensional grid data, extracting key vertexes, and deleting redundant data. And optimizing the model structure. And training a machine learning model based on the historical operation data, predicting the future state of the target object, and displaying the operation state and the prediction result of the target object in real time through a three-dimensional visual interface. The user can optimize the operation parameters of the target object according to the prediction result, adjust and improve the system efficiency in real time, and realize intelligent management. According to the method and the system, the rendering efficiency and the system response speed of the three-dimensional grid data can be more effectively improved, efficient data synchronization, optimization and visual display are realized, and the method and the system have important technical significance and application value.

Description

Technical Field

[0001] The present invention relates to the field of digital visualization technology, and in particular to a digital visualization optimization method and system based on artificial intelligence. Background Art

[0002] With the rapid development of digital twin technology, particularly in areas such as intelligent manufacturing, equipment monitoring, and factory management, real-time acquisition and analysis of operational data has become crucial for improving efficiency and optimizing management. However, digital twin systems, faced with large amounts of data and complex 3D models, often require efficient data processing capabilities, particularly in the simplification and optimization of 3D mesh data. Traditional methods often fail to fully consider how to reduce data volume without losing key geometric and topological features. This not only increases the computational burden but also compromises the real-time responsiveness and visualization of virtual models.

[0003] Most current digital twin systems experience delays in real-time data synchronization and 3D visualization, resulting in insufficient real-time performance and accuracy. Furthermore, traditional 3D modeling and optimization technologies often struggle to balance rendering accuracy with computational efficiency. This is particularly true for digital twin systems used for large-scale city, factory, or equipment management. Ensuring model optimization while maintaining visual quality and performance is a pressing challenge. Summary of the Invention

[0004] In view of the above technical problems, the present invention provides an artificial intelligence-based digital visualization optimization method and system, which can more effectively improve the rendering efficiency and system response speed of three-dimensional mesh data, realize efficient data synchronization, optimization and visualization display, and has important technical significance and application value.

[0005] Other features and advantages of the present invention will become apparent from the following detailed description, or may be learned in part by practice of the present invention.

[0006] According to one aspect of the present invention, a digital visual optimization method based on artificial intelligence is proposed, the method comprising:

[0007] Acquire real-time operation data and model data of the target object, use a 3D engine to read the geometric mesh information of the model data, and count the number of vertices and triangular meshes in the three-dimensional mesh information of the model data; extract and save only the coordinate set of the vertex from the three-dimensional mesh information, and retain the coordinates of the vertices of the non-shrinkable edges according to the geometric and topological characteristics, and delete other coordinates so that the shape information of the model data is separated separately; delete the original triangular mesh, and evaluate the deleted triangular mesh based on curvature calculation and quadratic error matrix method to ensure that the shape change of the model data after deleting the triangular mesh is within a preset range; calculate the coordinate average of the three vertices of the triangular mesh based on the triangle barycentric coordinate algorithm, use the coordinate average as the new vertex, and reconstruct the new triangular mesh according to the shortest distance principle of the adjacent multiple vertices, and repeat the operation until the reduction value of the number of vertices in the three-dimensional mesh information reaches a threshold;

[0008] Establishing a corresponding digital twin virtual model according to the operating characteristics of the target object and the model data, and reading the operating data of the target object based on a sensor, connecting the operating data to and synchronously driving the digital twin virtual model, so that the digital twin virtual model is consistent with the actual operating status of the target object in real time;

[0009] Building and training a machine learning model for performance prediction based on the historically collected operating data, wherein the machine learning model is selected from support vector regression, artificial neural network, or a combination of the two; using the trained machine learning model to predict the future state of the target object, and evaluating the prediction accuracy through error analysis and statistical indicators;

[0010] The operating status and prediction results of the digital twin virtual model are displayed in real time through a three-dimensional visualization interface, and user interactive operation functions are provided to monitor and manage the target object, as well as optimize the operating parameters of the target object according to the prediction results.

[0011] Furthermore, the digital twin virtual model runs on a visualization platform, which also includes roads and buildings.

[0012] Furthermore, when the road is created, the centerline data of the road in vector format is obtained, and the centerline data includes the width information, speed limit value, and channel attributes of the road; after obtaining the centerline data, the centerline of each road is offset by a fixed distance, and spatial buffering is performed on all centerline polylines of each road to create a polygon around the centerline of each road, and adjacent polygons are merged to form the boundary area of ​​the road, and the boundary area is rasterized to convert the vector data into raster data, and the raster data is tiled into multiple small blocks, renamed and input into the visualization platform.

[0013] Furthermore, when the building is created, vector data of the coverage area containing the outline of the building and point cloud data of the building collected based on the optical radar are obtained, and the point cloud data includes the height information and three-dimensional structure of the building. The point cloud data is cropped into the vector data, and the average height of the building is determined by calculating the median value of the Z-axis height in the vector data of the coverage area. Based on the average height, the coverage area of ​​the building is stretched along the Z-axis direction to the average height, so that the two-dimensional outline data of the building is converted into three-dimensional grid data.

[0014] Furthermore, the target object is a mobile device or a fixed device.

[0015] According to another aspect of the present invention, there is provided a digital visual optimization system based on artificial intelligence, comprising:

[0016] A model structure optimization module, the model structure optimization module is used to obtain real-time operation data and model data of the target object, use a 3D engine to read the geometric mesh information of the model data, and count the number of vertices and triangular meshes in the three-dimensional mesh information of the model data; extract and save only the coordinate set of the vertex from the three-dimensional mesh information, and retain the coordinates of the vertices of the non-shrinkable edges based on geometric and topological characteristics, delete other coordinates, so that the shape information of the model data is separated separately; delete the original triangular mesh, and evaluate the deleted triangular mesh based on curvature calculation and quadratic error matrix method to ensure that the shape change of the model data after deleting the triangular mesh is within a preset range; calculate the coordinate average of the three vertices of the triangular mesh based on the triangle barycentric coordinate algorithm, use the coordinate average as the new vertex, and reconstruct the new triangular mesh according to the shortest distance principle of multiple adjacent vertices, and repeat the operation until the reduction value of the number of vertices in the three-dimensional mesh information reaches a threshold;

[0017] A model generation module, the model generation module is used to establish a corresponding digital twin virtual model based on the operating characteristics of the target object and the model data, and read the operating data of the target object based on a sensor, connect the operating data to the digital twin virtual model and synchronously drive the digital twin virtual model, so that the digital twin virtual model is consistent with the actual operating status of the target object in real time;

[0018] a prediction module, the prediction module being configured to construct and train a machine learning model for performance prediction based on the historically collected operating data, the machine learning model being selected from support vector regression, artificial neural network, or a combination thereof, predict the future state of the target object using the trained machine learning model, and evaluate the prediction accuracy through error analysis and statistical indicators;

[0019] A visualization display module is used to display the operating status and prediction results of the digital twin virtual model in real time through a three-dimensional visualization interface, and provide user interactive operation functions to monitor and manage the target object, and optimize the operating parameters of the target object according to the prediction results.

[0020] The technical solution of the present invention has the following beneficial effects:

[0021] This method extracts and retains key information from the target object's 3D mesh, removes redundant data, and uses curvature calculation and the quadratic error matrix method to evaluate the shape changes after model simplification. This ensures that the model's visualization and accuracy are maintained while reducing the number of vertices. This optimization method significantly reduces the computational workload, enabling the digital twin virtual model to maintain efficient rendering speed even when processing large amounts of data.

[0022] By combining the real-time operational data of the target object with optimized 3D mesh data, this invention achieves real-time synchronization between the target object's digital twin virtual model and its actual operational state, and displays this through optimized visualization. This process ensures a high degree of consistency between the digital twin model and the actual state, and displays the dynamic changes of the target object through real-time graphical rendering, providing users with clear and intuitive real-time feedback.

[0023] This invention uses a three-dimensional visualization interface to display the operating status and prediction results of the target object in real time, which not only improves the intuitiveness of data display but also optimizes the user interaction experience. Users can interact with the digital twin model through the interface, adjust operating parameters, and optimize the target object's status based on the prediction results. In this way, users can efficiently manage and monitor the target object, thereby improving the system's operability and decision-making support capabilities.

[0024] This invention uses historical operational data and machine learning models (support vector regression and artificial neural networks) to predict the future state of an object. Combined with error analysis and statistical indicator evaluation, this invention provides highly accurate predictions of future operational states and optimizes the object's operation based on these predictions. The optimized visualization intuitively illustrates the discrepancies between predictions and actual conditions, helping users quickly identify potential issues and make adjustments. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a flow chart of a digital visual optimization method based on artificial intelligence in an embodiment of this specification;

[0026] Figure 2 is a schematic diagram of a triangular mesh in an embodiment of this specification;

[0027] Figure 3 This is a structural block diagram of an artificial intelligence-based digital visual optimization system in an embodiment of this specification. DETAILED DESCRIPTION

[0028] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that the present invention will be more comprehensive and complete and the concepts of the example embodiments will be fully conveyed to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present invention. However, those skilled in the art will appreciate that the technical solutions of the present invention may be practiced while omitting one or more of the specific details, or that other methods, components, devices, steps, etc. may be employed. In other cases, well-known technical solutions are not shown or described in detail to avoid obscuring various aspects of the present invention.

[0029] The accompanying drawings are merely schematic illustrations of the present invention. Identical reference numerals in the drawings denote identical or similar components, and thus repetitive descriptions thereof will be omitted. Some of the blocks shown in the accompanying drawings represent functional entities that do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0030] The present invention provides a digital visual optimization method for a product based on artificial intelligence. Figure 1The figure shows a flow chart of a method for optimizing digital visuals based on artificial intelligence for a product according to an embodiment of the present invention. The method can be applied to electronic devices such as personal computers and servers. The method can be performed by a device that can be implemented by software and / or hardware. The method can specifically include the following steps S101 to S102:

[0031] In step S101, the real-time operation data and model data of the target object are obtained, the geometric mesh information of the model data is read using a 3D engine, and the number of vertices and triangular meshes in the three-dimensional mesh information of the model data is counted; only the coordinate set of the vertex is extracted and saved from the three-dimensional mesh information, and according to the geometric and topological characteristics, the coordinates of the vertices of the non-shrinkable edges are retained, and other coordinates are deleted, so that the shape information of the model data is separated separately; the original triangular mesh is deleted, and the deleted triangular mesh is evaluated based on curvature calculation and quadratic error matrix method to ensure that the shape change of the model data after deleting the triangular mesh is within a preset range; based on the triangle barycenter algorithm of the barycenter coordinates, the coordinate average of the three vertices of the triangular mesh is calculated, the coordinate average is used as the new vertex, and the new triangular mesh is reconstructed according to the shortest distance principle of the adjacent multiple vertices, and the operation is repeated until the reduction value of the number of vertices in the three-dimensional mesh information reaches a threshold.

[0032] The target object is one of a mobile device and a fixed device, and can be any device, system or building. After the device is digitally visualized, its large size will affect the operation of the entire system and needs to be optimized. Geometric optimization is used here. Geometric optimization is an important step in the model optimization process. It refers to deleting or modifying the mesh of the model to reduce the volume of the model while maintaining the original shape. In order to import the model into a virtual engine, a neutral file format such as IFC, FBX or OB can be used. Models created in 3D engines using BIM (such as Unreal Engine, Unity, etc.) use surface modeling methods. These methods represent 3D entities through polygonal meshes composed of triangles and vertices. Unlike BIM solid methods such as boundary representation and constructed solid geometry, the imported model file format is not affected by the model representation method in the 3D engine. After importing, the model is represented by a polygonal mesh, regardless of the format of the transferred file.

[0033] Among them, the proposed triangle centroid algorithm based on barycentric coordinates is a mesh reconstruction technology that uses the triangle centroid formula based on barycentric coordinates to merge three triangles into one triangle by calculating the average value of the x, y, and z coordinates of the triangle vertices. Figure 2As shown, let the vertices of the triangle be P1, P2, and P3, and C be the center of mass of the triangle. The coordinates of the center of mass of C (s1:s2:s3) are expressed as follows:

[0034]

[0035] The midpoint of an edge is represented by:

[0036]

[0037] Through point and the straight line L between P3 123 (t1) is given by:

[0038]

[0039] Line L123(t1) contains one of the medians of triangle P1P2P3. By periodically looping, we can derive the equations for the three medians and ultimately determine the center of mass C. Specifically:

[0040]

[0041] Among them, t1,t2,t3∈R.

[0042] Figure 2 The centroid C of the triangle in the above formula is the intersection of the three lines, which is determined by solving the formula, where L 123 (t1) = L 231 (t2) = L 312 (t3), t1=t2=t3=2 / 3, and the center of mass C is:

[0043]

[0044] Since the coordinates of vertices P1, P2, and P3 are known when the model is imported, the coordinates of the center of mass can be obtained.

[0045] In step S102, a corresponding digital twin virtual model is established based on the operating characteristics of the target object and the model data, and the operating data of the target object is read based on the sensor, and the operating data is connected to and synchronously driven to make the digital twin virtual model consistent with the actual operating status of the target object in real time.

[0046] To do this, sensors (such as temperature sensors, vibration sensors, and flow sensors) are needed to collect various operational data from the target object. Sensor data can comprehensively reflect the target object's real-time status, such as its motion trajectory, operating status, and fault diagnosis. The data in the digital twin virtual model includes not only the target object's geometric structure information, but also its behavioral characteristics, operating parameters, and operating status. For example, a digital twin model of a building or equipment includes not only the building's three-dimensional form but also data from temperature and humidity sensors, energy consumption, and equipment operating status. In industrial production lines, targets (such as machines or production equipment) collect their operating parameters in real time through sensors. By accessing this real-time data, the digital twin model dynamically adjusts various indicators in the virtual model to ensure that the model fully matches the actual status of the equipment. This allows users or operators to visualize the actual status of the target object in a virtual environment without relying directly on on-site data from the physical equipment, achieving data visualization.

[0047] In step S103, a machine learning model for performance prediction is constructed and trained based on the historically collected operating data. The machine learning model is selected from support vector regression, artificial neural network, or a combination of the two. The trained machine learning model is used to predict the future state of the target object, and the prediction accuracy is evaluated through error analysis and statistical indicators.

[0048] Among them, based on the historically collected target object operation data, these data can be used to build and train a machine learning model for performance prediction. The model used can be selected from support vector regression (SVR), artificial neural network (ANN), or a combination of the two to adapt to different types of targets and prediction tasks. Support vector regression can effectively handle nonlinear relationships and is suitable for operation data with more complex trends; while artificial neural networks can adaptively learn deep-level features in the data through the calculation of multiple layers of neurons, and have strong modeling capabilities for complex relationships between multiple variables. By using historical operation data as a training set, the model will learn the relationship between the performance indicators of the target object (such as temperature, load, operating efficiency, etc.) and the operating parameters, and accurately predict the future state of the target object, usually including remaining service life, probability of failure, performance degradation trend, etc. The trained model will provide a scientific basis for equipment maintenance, adjustment and optimization by predicting future operating status. At the same time, in order to ensure the accuracy of the prediction, the error analysis method is used to evaluate the prediction results. Commonly used statistical indicators include root mean square error (RMSE), mean absolute error (MAE) and R-squared (R 2 ), etc., ultimately achieving high-precision prediction and effective monitoring of the target object's operating status, providing strong support for intelligent operation and maintenance, preventive maintenance and system optimization.

[0049] In step S104, the operating status and prediction results of the digital twin virtual model are displayed in real time through a three-dimensional visualization interface, and a user interactive operation function is provided to monitor and manage the target object, and the operating parameters of the target object are optimized according to the prediction results.

[0050] The 3D visualization interface displays the operating status and prediction results of the digital twin virtual model in real time, allowing users to intuitively view the current state of the target object, historical trends, and predicted future state. This interface not only provides a dynamic display of the target object's operating status but also uses integrated data visualization tools to reflect key operating parameters such as temperature, pressure, and load in real time. Furthermore, the interface features user interaction, allowing operators to interact with the virtual model via a mouse, keyboard, or touchscreen. Based on the model's operating status and prediction results, users can adjust the target object's operating parameters in real time, such as changing the equipment's workload, adjusting temperature control system settings, or optimizing production processes. This interactive feature enables users to make quick decisions based on real-time feedback, improving management efficiency and operational flexibility while ensuring optimal performance of the target object. Based on the prediction results, the system can also automatically recommend optimization measures to help users make maintenance decisions or avoid potential failures, thereby achieving intelligent monitoring and management.

[0051] In one embodiment, the digital twin virtual model runs on a visualization platform, which also includes roads and buildings.

[0052] Specifically, when creating a road, vector centerline data is obtained. This data includes the road's width, speed limit, and channel attributes. After obtaining this centerline data, each road centerline is offset by a fixed distance. Spatial buffering is performed on all centerline polylines of each road to create a polygon around each road centerline. Adjacent polygons are fused to form the road boundary region. This boundary region is rasterized to convert the vector data into raster data. The raster data is then tiled into multiple small blocks, renamed, and input into the visualization platform. Each road centerline is offset by a fixed distance. Spatial buffering is performed on all centerline polylines of each road category. This operation aims to create a polygon representing the road width around each road centerline, thereby generating a virtual "road boundary region" for each road. After the generated buffer polygons are created, a fusion operation is performed to merge adjacent buffers into a single road region. If the road network is disconnected, multiple fused regions will be generated; however, typically, multiple roads form a closed region, i.e., a complete road boundary.

[0053] Specifically, when creating a building, vector data of a footprint containing the building's outline and point cloud data of the building collected using optical radar are obtained. The point cloud data includes the building's height information and three-dimensional structure. The point cloud data is clipped to the vector data. The average building height is determined by calculating the median value of the Z-axis height in the vector data of the footprint. Based on the average height, the footprint of the building is stretched along the Z-axis to the average height, thereby converting the two-dimensional outline data of the building into three-dimensional mesh data. Point cloud data typically contains a large amount of three-dimensional coordinate data, which can accurately describe the shape of the ground and the building. The point cloud data is clipped to the building's outline (i.e., the building's footprint). This ensures that only point cloud data related to the building is processed, and computing resources are not wasted on data outside the building. In the point cloud data above the building's footprint, the average building height is determined by calculating the median value of the z dimension (height). The z value here represents the building's height. In this way, a typical height value representing the top of the building can be obtained. Using the calculated average height, the building's footprint is "extruded" along the z-axis to this height. This process converts the building's 2D outline data into a 3D mesh, generating a basic building model mesh. The height of this mesh is calculated based on the point cloud data. To make the generated building mesh compatible with the coordinate system in Unreal Engine, the mesh needs to be translated. This is because Unreal Engine typically uses a different coordinate system, requiring the building mesh to be transformed to align with other generated virtual data. After completing these steps, the generated building mesh is imported into the 3D engine, which processes the mesh data for rendering and further interactive operations such as scene display and environmental analysis.

[0054] Through the above steps, combined with the equipment model, the entire industrial park can be digitally visualized.

[0055] Based on the same idea, Figure 2 As shown, a digital visual optimization system based on artificial intelligence is provided, comprising:

[0056] The model structure optimization module 201 is used to obtain real-time operation data and model data of the target object, use a 3D engine to read the geometric mesh information of the model data, and count the number of vertices and triangular meshes in the three-dimensional mesh information of the model data; extract and save only the coordinate set of the vertex from the three-dimensional mesh information, and retain the coordinates of the vertices of the non-shrinkable edges based on geometric and topological characteristics, and delete other coordinates so that the shape information of the model data is separated separately; delete the original triangular mesh, and evaluate the deleted triangular mesh based on curvature calculation and quadratic error matrix method to ensure that the shape change of the model data after deleting the triangular mesh is within a preset range; calculate the coordinate average of the three vertices of the triangular mesh based on the triangle barycentric coordinate algorithm, use the coordinate average as the new vertex, and reconstruct the new triangular mesh according to the shortest distance principle of multiple adjacent vertices, and repeat the operation until the reduction value of the number of vertices in the three-dimensional mesh information reaches a threshold;

[0057] A model generation module 202 is configured to establish a corresponding digital twin virtual model based on the operating characteristics of the target object and the model data, read the operating data of the target object based on sensors, connect the operating data to the digital twin virtual model, and synchronously drive the digital twin virtual model so that the digital twin virtual model is consistent with the actual operating status of the target object in real time;

[0058] A prediction module 203 is configured to construct and train a machine learning model for performance prediction based on the historically collected operating data, wherein the machine learning model is selected from support vector regression, artificial neural network, or a combination thereof, and to use the trained machine learning model to predict the future state of the target object, and to evaluate the prediction accuracy through error analysis and statistical indicators;

[0059] The visualization display module 204 is used to display the operating status and prediction results of the digital twin virtual model in real time through a three-dimensional visualization interface, and provide user interactive operation functions to monitor and manage the target object, and optimize the operating parameters of the target object according to the prediction results.

[0060] This system extracts and retains key information from the target object's 3D mesh, removes redundant data, and uses curvature calculations and the quadratic error matrix method to evaluate the shape changes after model simplification. This ensures that the model's visualization and accuracy are maintained while reducing the number of vertices. This optimization method significantly reduces the amount of computation required, allowing the digital twin virtual model to maintain efficient rendering speeds even when processing large amounts of data.

[0061] By combining the object's real-time operational data with optimized 3D mesh data, the system synchronizes the object's digital twin model with its actual operational state in real time, presenting it through optimized visualization. This process ensures high consistency between the digital twin model and the actual state, and displays the object's dynamic changes through real-time graphical rendering, providing users with clear, intuitive, and real-time feedback.

[0062] This system displays the target's operating status and prediction results in real time through a 3D visualization interface, making the data presentation more intuitive and enhancing the user experience. Users can interact with the digital twin model through the interface, adjusting operating parameters and optimizing the target's status based on the prediction results. This allows users to efficiently manage and monitor the target, improving the system's operability and decision-making capabilities.

[0063] The system uses historical operational data and machine learning models (support vector regression and artificial neural networks) to predict the future state of an object. Combined with error analysis and statistical evaluation, the system provides highly accurate predictions of future operational states and optimizes the object's operation based on these predictions. The optimized visualization intuitively illustrates the discrepancies between predictions and actual conditions, helping users quickly identify potential issues and make adjustments.

[0064] The specific details of each module / unit in the above system have been described in detail in the implementation method part. For undisclosed details, please refer to the implementation method part, and thus will not be repeated here.

[0065] Through the description of the above embodiments, it is easy for those skilled in the art to understand that the example embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution according to the embodiment of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes a number of instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the exemplary embodiment of the present invention.

[0066] Furthermore, the figures above are merely illustrative of the processes included in the method according to exemplary embodiments of the present invention and are not intended to be limiting. It is readily understood that the processes illustrated in the figures above do not indicate or limit the temporal order of these processes. Furthermore, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0067] It should be noted that, although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to an exemplary embodiment of the present invention, the features and functions of two or more modules or units described above can be concretized in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided into multiple modules or units to be concretized.

[0068] Those skilled in the art will readily identify other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The description and embodiments are to be considered as exemplary only, with the true scope and spirit of the invention being indicated by the claims.

[0069] It should be understood that the present invention is not limited to the exact construction described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.

Claims

1. A digital visual optimization method based on artificial intelligence, characterized in that: The method comprises: Acquire real-time operation data and model data of the target object, use a 3D engine to read the geometric mesh information of the model data, and count the number of vertices and triangular meshes in the three-dimensional mesh information of the model data; extract and save only the coordinate set of the vertex from the three-dimensional mesh information, and retain the coordinates of the vertices of the non-shrinkable edges according to the geometric and topological characteristics, and delete other coordinates so that the shape information of the model data is separated separately; delete the original triangular mesh, and evaluate the deleted triangular mesh based on curvature calculation and quadratic error matrix method to ensure that the shape change of the model data after deleting the triangular mesh is within a preset range; calculate the coordinate average of the three vertices of the triangular mesh based on the triangle barycentric coordinate algorithm, use the coordinate average as the new vertex, and reconstruct the new triangular mesh according to the shortest distance principle of the adjacent multiple vertices, and repeat the operation until the reduction value of the number of vertices in the three-dimensional mesh information reaches a threshold; Establishing a corresponding digital twin virtual model according to the operating characteristics of the target object and the model data, and reading the operating data of the target object based on a sensor, connecting the operating data to and synchronously driving the digital twin virtual model, so that the digital twin virtual model is consistent with the actual operating status of the target object in real time; Building and training a machine learning model for performance prediction based on the historically collected operating data, wherein the machine learning model is selected from support vector regression, artificial neural network, or a combination of the two; using the trained machine learning model to predict the future state of the target object, and evaluating the prediction accuracy through error analysis and statistical indicators; The operating status and prediction results of the digital twin virtual model are displayed in real time through a three-dimensional visualization interface, and user interactive operation functions are provided to monitor and manage the target object, as well as optimize the operating parameters of the target object according to the prediction results.

2. The digital visual optimization method based on artificial intelligence according to claim 1, characterized in that: The digital twin virtual model runs on a visualization platform, which also includes roads and buildings.

3. The digital visual optimization method based on artificial intelligence according to claim 2, characterized in that: When the road is created, the centerline data of the road in vector format is obtained, and the centerline data includes the width information, speed limit value, and channel attributes of the road. After obtaining the centerline data, the centerline of each road is offset by a fixed distance, and spatial buffering is performed on all centerline polylines of each road to create a polygon around the centerline of each road. Adjacent polygons are merged to form the boundary area of ​​the road, and the boundary area is rasterized to convert the vector data into raster data. After the raster data is tiled into multiple small blocks, it is renamed and input into the visualization platform.

4. The digital visual optimization method based on artificial intelligence according to claim 2, characterized in that: When the building is created, vector data of the coverage area containing the outline of the building and point cloud data of the building collected based on the optical radar are obtained, and the point cloud data includes the height information and three-dimensional structure of the building. The point cloud data is clipped into the vector data, and the average height of the building is determined by calculating the median value of the Z-axis height in the vector data of the coverage area. Based on the average height, the coverage area of ​​the building is stretched along the Z-axis direction to the average height, so that the two-dimensional outline data of the building is converted into three-dimensional grid data.

5. The digital visual optimization method based on artificial intelligence according to claim 1, characterized in that: The target object is one of a mobile device and a fixed device.

6. A digital visual optimization system based on artificial intelligence, characterized in that: include: A model structure optimization module is used to obtain real-time operation data and model data of the target object, read geometric mesh information of the model data using a 3D engine, and count the number of vertices and triangle meshes in the three-dimensional mesh information of the model data; Only the coordinate sets of the vertices are extracted and saved from the three-dimensional mesh information, and based on geometric and topological characteristics, the coordinates of the vertices of the non-shrinkable edges are retained, and other coordinates are deleted, so that the shape information of the model data is separated separately; the original triangular mesh is deleted, and the deleted triangular mesh is evaluated based on curvature calculation and quadratic error matrix method to ensure that the shape change of the model data after deleting the triangular mesh is within a preset range; based on the triangle barycentric algorithm of barycentric coordinates, the coordinate average of the three vertices of the triangular mesh is calculated, the coordinate average is used as the new vertex, and the new triangular mesh is reconstructed according to the shortest distance principle of the adjacent multiple vertices, and the operation is repeated until the reduction value of the number of the vertices in the three-dimensional mesh information reaches a threshold value; A model generation module, the model generation module is used to establish a corresponding digital twin virtual model based on the operating characteristics of the target object and the model data, and read the operating data of the target object based on a sensor, connect the operating data to the digital twin virtual model and synchronously drive the digital twin virtual model, so that the digital twin virtual model is consistent with the actual operating status of the target object in real time; a prediction module, the prediction module being configured to construct and train a machine learning model for performance prediction based on the historically collected operating data, the machine learning model being selected from support vector regression, artificial neural network, or a combination thereof, predict the future state of the target object using the trained machine learning model, and evaluate the prediction accuracy through error analysis and statistical indicators; A visualization display module is used to display the operating status and prediction results of the digital twin virtual model in real time through a three-dimensional visualization interface, and provide user interactive operation functions to monitor and manage the target object, and optimize the operating parameters of the target object according to the prediction results.

Images