Terrain digital twin scene surveying and mapping system based on three-dimensional modeling

Through the terrain digital twin scene surveying and mapping system based on three-dimensional modeling, combined with intelligent topography feature analysis and adaptive modeling strategy selection, the problem of difficulty in taking into account the accuracy and efficiency of traditional topography surveying and mapping methods is solved, and efficient and accurate terrain modeling and Improved visual effects.

CN119006735AActive Publication Date: 2024-11-22NANTONG DADI SURVEYING & MAPPING CO LTD

Patent Information

Application Number
CN202411494381.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-24
Publication Date
2024-11-22
Estimated Expiration
2044-10-24

AI Technical Summary

Technical Problem

Traditional terrain mapping methods have problems such as low accuracy, long time and high cost. Especially when dealing with complex terrain, modeling accuracy and efficiency are difficult to take into account, which affects processing speed and rendering performance.

Method used

Adopt the terrain digital twin scene surveying and mapping system based on three-dimensional modeling, including intelligent terrain feature analysis module, adaptive modeling strategy selection module, self-learning and optimization module, and terrain digital twin scene construction and display module, and analyze terrain features through deep learning algorithms. , select appropriate modeling strategies, balance model accuracy and data volume, and continuously improve modeling accuracy and efficiency through self-learning optimization module.

Benefits of technology

It significantly improves the accuracy and efficiency of terrain surveying, improves visual effects and user experience, and achieves a balance of visual experience while maintaining the overall model efficiency, enhancing the application value and user satisfaction of the system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119006735A_ABST
    Figure CN119006735A_ABST
Patent Text Reader

Abstract

The invention discloses a topographic digital twinborn scene surveying and mapping system based on three-dimensional modeling, and relates to the technical field of surveying and mapping science, and the system comprises the following components: an intelligent topographic feature analysis module, a self-adaptive modeling strategy selection module, a self-learning and optimization module and a topographic digital twinborn scene construction and display module. Through the intelligent topographic feature analysis module and the self-adaptive modeling strategy selection module, topographic features can be automatically analyzed, a proper modeling strategy can be matched, a grid simplification algorithm is adopted in a flat area, unnecessary details are effectively removed, key features are kept, the model precision is ensured, and the modeling efficiency is improved. A multi-resolution curved surface modeling technology is adopted in a complex terrain, the model precision is flexibly adjusted, so that the surveying and mapping precision and the modeling efficiency are remarkably improved, in addition, the system is further integrated with a self-learning and optimizing module, modeling parameters and strategies can be continuously optimized, and the surveying and mapping precision and the modeling efficiency are further improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of surveying and mapping science and technology, and in particular to a terrain digital twin scene mapping system based on three-dimensional modeling. Background Art

[0002] With the rapid development of geographic information technology and computer graphics, terrain surveying and modeling technology are playing an increasingly important role in urban planning, disaster emergency response, and military simulation. Traditional terrain surveying and mapping methods often rely on manual measurement and two-dimensional maps, which have the disadvantages of low accuracy, long time consumption, and high cost. They are difficult to meet the needs of modern complex terrain surveying and mapping. In recent years, three-dimensional modeling technology has gradually attracted attention because it can express terrain features more realistically and intuitively.

[0003] Traditional adaptive modeling strategy selection modules often have the following shortcomings when dealing with complex terrain. When faced with complex and changeable terrain, the amount of high-precision modeling data is huge, and it is often difficult to balance modeling accuracy and efficiency, and the model has poor adaptability, which affects the processing speed and rendering performance. Therefore, it is particularly important to develop a terrain digital twin scene mapping system based on three-dimensional modeling. Summary of the invention

[0004] In order to make up for the shortcomings of the prior art, the purpose of the present invention is to provide a terrain digital twin scene mapping system based on three-dimensional modeling. The system flexibly matches the modeling strategy according to the complexity and characteristics of different terrains through an adaptive modeling strategy selection module, balances the model accuracy and data volume, and at the same time, has a built-in self-learning and optimization module to continuously learn and optimize modeling parameters and strategies to improve the accuracy and efficiency of modeling.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a terrain digital twin scene mapping system based on three-dimensional modeling, the system comprising an intelligent terrain feature analysis module, an adaptive modeling strategy selection module, a self-learning and optimization module and a terrain digital twin scene construction and display module; The intelligent terrain feature analysis module receives surveying and mapping data from multiple data sources such as drones, satellites, and ground surveying and mapping equipment, and uses deep learning algorithms to automatically analyze the terrain's undulation, landform type, and vegetation coverage density characteristics. Through feature extraction and classification, it provides basic data support for subsequent modeling strategy selection. The adaptive modeling strategy selection module receives the terrain undulation, landform type, and vegetation coverage density feature data from the intelligent terrain feature analysis module, analyzes the data, determines the complexity and feature classification of the current processed terrain, and preliminarily matches the modeling strategy according to the terrain feature classification; For flat areas, select the mesh simplification algorithm, which can identify and remove unnecessary details in the model while keeping key features to maintain model accuracy. Set algorithm parameters to balance model accuracy and data volume. ,in, represents the comprehensive evaluation value of the simplified model, A measure representing a key characteristic, Represents the total metric value of the model in its initial state, Represents the algorithm parameters, when When the value increases, it means that the amount of data is more likely to be reduced. The impact of this item increases, and more non-critical parts are removed. When the value decreases, more emphasis is placed on model accuracy. The increase in the impact of this term will better preserve key features, by adjusting The value of can balance the model accuracy and data volume. For complex terrain in mountainous areas, multi-resolution surface modeling technology is selected to apply models of different accuracy in different areas to efficiently express complex terrain and determine multi-level resolution levels. ,in, Represents the comprehensive resolution value of the complex terrain model, A measure representing an area of ​​important geomorphic features, Represents the total metric value of the entire complex terrain area, , , when the area with important geomorphic features accounts for a large proportion, that is Close to 1, at this time Closer , indicating that the entire model is presented with higher accuracy. When the proportion of non-important areas is large, Close to 0, Closer Indicates that the entire model is presented with lower precision. and The proportion and and The value of can flexibly control the accuracy of complex terrain models in different areas to achieve the purpose of efficiently expressing complex terrain. In non-critical areas, lower-precision models are used. Combined with high-precision texture mapping technology, realistic surface materials are added to the model to enhance the visual effect. Detail level technology is applied to dynamically adjust the model complexity according to the observation distance to improve rendering efficiency. According to the detailed characteristics of the specific terrain and modeling requirements, further adjust the parameters of the modeling strategy, including but not limited to adjusting the mesh resolution, texture detail level, and surface fitting algorithm, so that the final model meets the accuracy requirements and achieves the best performance; Execute the selected modeling strategy to generate a three-dimensional terrain model, verify the generated model to check whether it accurately reflects the terrain characteristics, including but not limited to terrain undulations, landform types, and vegetation cover, evaluate the data volume, processing speed, and rendering performance of the model, collect feedback information during the modeling process, and optimize and improve the modeling strategy selection process and parameter settings based on the feedback results; The self-learning and optimization module has a built-in machine learning engine that can continuously learn and optimize modeling parameters and strategies. With the continuous accumulation of surveying and mapping data, it can automatically adjust the modeling algorithm to improve the accuracy and efficiency of modeling, and continuously optimize the model quality through user feedback and model evaluation mechanism; The terrain digital twin scene construction and display module: based on the output of the above module, a highly realistic terrain digital twin scene is constructed. The scene supports multi-perspective, multi-scale browsing and interactive operations, facilitates users to conduct in-depth geographic analysis and decision support, provides rich visualization tools and interfaces, and expands the breadth and depth of application scenarios.

[0006] Furthermore, the terrain digital twin construction process of the intelligent terrain feature analysis module starts with data reception and preprocessing, covering real-time or regular collection of surveying and mapping data from multiple data sources, and verification and preprocessing, including but not limited to image correction and point cloud filtering, to ensure uniform data quality. Subsequently, feature extraction is performed to obtain detailed terrain features through terrain undulation, landform recognition and vegetation cover density analysis.

[0007] Furthermore, in the feature classification and fusion stage of the intelligent terrain feature analysis module, the extracted features are refined into standardized sets and multi-dimensional data are integrated to form a comprehensive data model. This data model supports the subsequent modeling strategy selection. The entire process focuses on feedback and optimization, and continuously iterates to improve the intelligence level of the system, providing a solid foundation for the construction of terrain digital twins.

[0008] Furthermore, the adaptive modeling strategy selection module uses a lower precision model in non-critical areas, combined with high-precision texture mapping technology, to add realistic surface materials to the model and enhance the visual effect: in, Represents the comprehensive value of the visual effect of the non-critical area model. Indicates the effect value of high-precision texture mapping. represents the metric value of the non-critical area, represents the total metric value for the entire terrain area, Represents the visual effect value of the low-precision model itself. When the proportion of non-critical areas is larger, Close to 1, the overall visual effect value Closer to the visual effect value of the low-precision model But it will also be affected by the high-precision texture mapping effect. When the non-critical area accounts for a smaller proportion, Close to 0, Closer to the effect value of high-precision texture mapping , by adjusting and The proportion and and The value of can achieve different degrees of visual effect balance in non-critical areas, which not only ensures the efficiency of the low-precision model, but also improves the visual experience to a certain extent through high-precision texture mapping.

[0009] Furthermore, the adaptive modeling strategy selection module applies the level of detail technology to dynamically adjust the model complexity according to the observation distance to improve the rendering efficiency: ,in, represents the complexity of the model, Indicates the observation distance, represents the maximum observation distance, is an adjustable parameter that controls the effect of observation distance on complexity. is a small positive number, indicating that when the observation distance is very close to the maximum observation distance, the minimum complexity of the model is avoided to avoid the situation where the complexity is zero. When it is close to zero, that is, the observer is very close to the model, tends to infinity, the model complexity Mainly by The model shows higher complexity as the observation distance increases. Increase, gradually decreases, and Gradually increases, the model complexity gradually decreases, and by adjusting the parameters , the speed of change of model complexity at different observation distances can be balanced according to actual needs.

[0010] Furthermore, the self-learning and optimization module system continuously collects surveying and mapping data from drones, satellites and ground surveying and mapping equipment, and performs data cleaning, format unification and coordinate correction preprocessing to ensure data quality. Subsequently, the modeling parameters and strategies are initialized based on existing data and past experience, and learning goals and optimization indicators are set. By executing modeling tasks and evaluating model quality, the system collects user feedback and converts it into quantitative indicators to drive the machine learning engine to automatically adjust modeling parameters and strategies.

[0011] Furthermore, the above process of the self-learning and optimization module constitutes an iterative optimization loop. By repeatedly modeling, evaluating and collecting feedback, the model is gradually optimized. The optimized model is deployed to actual applications and updated regularly as needed. At the same time, the system continuously monitors the model performance, collects new data and user feedback, so that the machine learning engine can automatically adapt to changes in terrain characteristics and user needs.

[0012] Furthermore, the terrain digital twin scene construction and display module integrates optimized three-dimensional terrain data, uses three-dimensional modeling technology to build a basic framework, and designs interactive functions to enhance user experience. At the same time, it develops visualization tools to enrich the scene expressiveness and expands application scenarios to meet different needs. After comprehensive testing, the module will be deployed and provide continuous maintenance.

[0013] Compared with the existing technology, a terrain digital twin scene mapping system based on 3D modeling has the following beneficial effects: 1. The present invention can automatically analyze terrain features and match appropriate modeling strategies through an intelligent terrain feature analysis module and an adaptive modeling strategy selection module. In flat areas, a grid simplification algorithm is used to effectively remove unnecessary details while maintaining key features, thereby ensuring model accuracy. In complex terrains, multi-resolution surface modeling technology is used to flexibly adjust model accuracy, thereby significantly improving surveying and mapping accuracy and modeling efficiency. In addition, the system also integrates a self-learning and optimization module, which can continuously optimize modeling parameters and strategies to further improve surveying and mapping accuracy and modeling efficiency.

[0014] 2. The present invention adopts lower-precision models in non-critical areas and combines high-precision texture mapping technology to add realistic surface materials to the model, thereby significantly improving the visual effect of the terrain digital twin scene, adjusting the accuracy and texture details of the models in different areas, and achieving a balance between maintaining the overall model efficiency and improving the visual experience. This strategy not only ensures the authenticity of the terrain digital twin scene, but also significantly improves the user's visual experience during use, thereby enhancing the application value of the system and user satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0016] Figure 1 An operation flow chart of a terrain digital twin scene mapping system based on three-dimensional modeling. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Embodiment 1 Receive mapping data from multiple data sources such as drones, satellites, and ground mapping equipment in real time or regularly. These data may include high-resolution images, point cloud data, and terrain elevation models. Check the integrity and format correctness of the received data to ensure data quality. Remove or mark outliers and noise data. Perform geometric correction, radiation correction, and atmospheric correction on image data to improve image quality. Filter point cloud data to remove non-ground points and generate accurate terrain surface models. Unify data formats and coordinate systems to ensure compatibility between different data sources. Use DEM data to analyze the undulation of terrain by calculating slope, aspect, and elevation variation parameters. Use spatial analysis methods and image classification technology to identify different landform types by combining spectral information and texture information in remote sensing images. Use deep learning models for image segmentation to improve classification accuracy. Analyze vegetation indices in remote sensing images to evaluate vegetation coverage density and health status. Combine spectral features and texture features to distinguish different vegetation types and their densities using classification algorithms. The extracted terrain undulations, landform types, and vegetation coverage features are further classified to form a standardized feature set. Clustering algorithms or decision tree machine learning methods are used to refine the classification of features. Features from different data sources and different dimensions are integrated to form a comprehensive data model that fully describes the terrain features. Considering the correlation and redundancy between features, feature selection and dimensionality reduction are performed to improve the efficiency and accuracy of data processing. The processed feature data and comprehensive data model are stored in the database for subsequent query and use. A data index and query mechanism is established to improve data access efficiency. Based on the results of terrain feature analysis, basic data support is provided for subsequent modeling work. According to the needs of specific application scenarios, appropriate modeling algorithms and strategies are selected, feedback data during the modeling process is collected, and the performance and accuracy of the terrain feature analysis module are evaluated. Based on the evaluation results, the data processing process, feature extraction algorithm, and classification model are optimized and improved to continuously improve the intelligence level of the system.

[0019] Receive terrain undulation, landform type, and vegetation coverage density feature data from the intelligent terrain feature analysis module, analyze these data, determine the complexity and feature classification of the current terrain, and preliminarily match the modeling strategy based on the terrain feature classification; For flat areas, select the mesh simplification algorithm, which can identify and remove unnecessary details in the model while keeping key features to maintain model accuracy. Set algorithm parameters to balance model accuracy and data volume. ,in, represents the comprehensive evaluation value of the simplified model, A measure representing a key characteristic, Represents the total metric value of the model in its initial state, Represents the algorithm parameters, when When the value increases, it means that the amount of data is more likely to be reduced. The impact of this item increases, and more non-critical parts are removed. When the value decreases, more emphasis is placed on model accuracy. The increase in the impact of this term will better preserve key features, by adjusting The value of can balance the model accuracy and data volume. For complex terrain in mountainous areas, multi-resolution surface modeling technology is selected to apply models of different accuracy in different areas to efficiently express complex terrain and determine multi-level resolution levels. ,in, Represents the comprehensive resolution value of the complex terrain model, A measure representing an area of ​​important geomorphic features, Represents the total metric value of the entire complex terrain area, , , when the area with important geomorphic features accounts for a large proportion, that is Close to 1, at this time Closer , indicating that the entire model is presented with higher accuracy. When the proportion of non-important areas is large, Close to 0, Closer Indicates that the entire model is presented with lower precision. and The proportion and and The value of can flexibly control the accuracy of complex terrain models in different areas, so as to achieve the purpose of efficiently expressing complex terrain. In non-critical areas, lower-precision models are used. Combined with high-precision texture mapping technology, realistic surface materials are added to the model to enhance the visual effect. in, Represents the comprehensive value of the visual effect of the non-critical area model. Indicates the effect value of high-precision texture mapping. represents the metric value of the non-critical area, represents the total metric value for the entire terrain area, Represents the visual effect value of the low-precision model itself. When the proportion of non-critical areas is larger, Close to 1, the overall visual effect value Closer to the visual effect value of the low-precision model But it will also be affected by the high-precision texture mapping effect. When the non-critical area accounts for a smaller proportion, Close to 0, Closer to the effect value of high-precision texture mapping , by adjusting and The proportion and and The value of can achieve different degrees of visual effect balance in non-critical areas, which not only ensures the efficiency of low-precision models, but also improves a certain visual experience through high-precision texture mapping. The detail level technology is applied to dynamically adjust the model complexity according to the observation distance to improve rendering efficiency. ,in, represents the complexity of the model, Indicates the observation distance, represents the maximum observation distance, is an adjustable parameter that controls the effect of observation distance on complexity. is a small positive number, indicating that when the observation distance is very close to the maximum observation distance, the minimum complexity of the model is avoided to avoid the situation where the complexity is zero. When it is close to zero, that is, the observer is very close to the model, tends to infinity, the model complexity Mainly by The model shows higher complexity as the observation distance increases. Increase, gradually decreases, and Gradually increases, the model complexity gradually decreases, and by adjusting the parameters , the speed of change of model complexity at different observation distances can be balanced according to actual needs; According to the detailed characteristics of the specific terrain and modeling requirements, further adjust the parameters of the modeling strategy, including but not limited to adjusting the mesh resolution, texture detail level, and surface fitting algorithm, so that the final model meets the accuracy requirements and achieves the best performance; Execute the selected modeling strategy, generate a three-dimensional terrain model, verify the generated model, check whether it accurately reflects the terrain characteristics, including but not limited to terrain undulations, landform types, and vegetation cover, evaluate the model's data volume, processing speed, and rendering performance, collect feedback information during the modeling process, and optimize and improve the modeling strategy selection process and parameter settings based on the feedback results.

[0020] Continuously receive new surveying and mapping data from external data sources, pre-process the new data, including data cleaning, format unification, and coordinate correction to ensure data quality, initialize modeling parameters and strategies based on currently available data and previous modeling experience, set the learning goals and optimization indicators of the machine learning engine, use the current modeling parameters and strategies to perform modeling tasks, generate a three-dimensional terrain model, evaluate the model quality through a preset model evaluation mechanism, collect user feedback on model quality, and convert user feedback into quantifiable evaluation indicators for subsequent learning and optimization. The machine learning engine automatically adjusts modeling parameters and strategies based on new surveying and mapping data, model evaluation results, and user feedback, and uses reinforcement learning, supervised learning, or unsupervised learning methods to continuously optimize the modeling algorithm to improve the accuracy and efficiency of modeling; In each iteration, the modeling parameters and strategies are gradually optimized according to new data and feedback until the set learning goals or optimization indicators are achieved. The optimized model is deployed to the actual application scenario for users to use. The model is updated regularly or as needed to ensure that the model always maintains a high level of accuracy and efficiency. The performance of the model in actual applications is monitored, and new data and user feedback are collected. The machine learning engine continues to learn and automatically adjusts the model according to new data and feedback to adapt to the ever-changing terrain characteristics and user needs.

[0021] Receive optimized 3D terrain model data and related attribute information from the self-learning and optimization module, further integrate the data to ensure seamless connection and consistency between different data sources, perform necessary data preprocessing to improve subsequent processing efficiency, and build the basic framework of the terrain digital twin scene using 3D modeling software or engine based on the optimized 3D terrain data, introduce multi-view and multi-scale support, and achieve terrain display of different precisions through layered refinement technology to meet different analysis needs, and add necessary geographical feature elements such as roads, buildings, and water bodies to enrich the scene content; Design and implement user interaction interfaces, including but not limited to basic operations such as perspective switching, zooming, and panning. Introduce advanced interactive functions such as terrain profile analysis, terrain height query, and region selection to meet the needs of in-depth geographic analysis, ensure the smoothness and response speed of interactive operations, and improve user experience. Develop rich visualization tools to enhance the expressiveness and readability of the scene. Optimize the rendering of the scene, including material optimization, lighting calculation optimization, and shadow processing to improve the realism and rendering efficiency of the scene. Introduce dynamic weather systems and day and night alternating environmental effects to make the scene more vivid and realistic. According to user needs, expand the application scenarios of terrain digital twin scenes, provide API interfaces or SDKs so that other systems or applications can easily integrate and use terrain digital twin scenes, support customized development, personalize and optimize scenes according to the needs of specific projects, conduct comprehensive tests on the constructed terrain digital twin scenes, including functional testing, performance testing, and compatibility testing, collect user feedback, understand the problems and suggestions encountered by users during use, so as to carry out subsequent optimization and improvement, deploy the terrain digital twin scenes to the target platform or server for user use, and regularly update and maintain the scene content, including data updates, functional optimization, and security reinforcement, to ensure the stable operation and sustainable development of the scene.

[0022] Embodiment 2 Using drones for high-altitude photography to obtain high-resolution images, combined with satellite remote sensing data covering a large area, and precise measurements from ground surveying and mapping equipment, we comprehensively collect multivariate surveying and mapping data of urban terrain, and perform strict verification, image correction, and point cloud filtering preprocessing on the collected data to ensure data accuracy and consistency. Through data cleaning, we remove noise and erroneous data to lay a solid foundation for subsequent analysis. Using deep learning algorithms, we can accurately identify and classify various features of urban terrain. Through feature extraction and classification, we can refine the terrain features obtained through analysis into standardized sets, and integrate multidimensional data to form a comprehensive urban terrain integrated data model. This model provides strong data support for the selection of subsequent modeling strategies.

[0023] In flat urban areas, the mesh simplification algorithm is automatically selected for modeling. ,in, represents the comprehensive evaluation value of the simplified model, The measurement value of key features, the number and area of ​​key features of important buildings and road intersections in flat urban areas, Indicates the total measurement value of the model in the initial state, the total number of vertices and total area of ​​the flat urban area model at the beginning, Represents the algorithm parameters, when When the value increases, it means that the amount of data is more likely to be reduced. The impact of this item increases, and more non-critical parts are removed. When the value decreases, more emphasis is placed on model accuracy. The increase in the impact of this term will better preserve key features, by adjusting The value of can balance the model accuracy and data volume. By optimizing the algorithm parameters, the system can accurately identify and retain key features, while removing unnecessary details to reduce the amount of data and increase the model processing speed. For complex terrain areas such as mountains and rivers, multi-resolution surface modeling technology is used. ,in, Represents the comprehensive resolution value of the complex terrain model, A measure representing an area of ​​important geomorphic features, Represents the total metric value of the entire complex terrain area, , , when the area with important geomorphic features accounts for a large proportion, that is Close to 1, at this time Closer , indicating that the entire model is presented with higher accuracy. When the proportion of non-important areas is large, Close to 0, Closer Indicates that the entire model is presented with lower precision. and The proportion and and The value of can flexibly control the accuracy of complex terrain models in different areas, and achieve the purpose of efficiently expressing complex terrain. This technology can dynamically adjust the model accuracy according to the complexity of the terrain, ensuring that important landform feature areas are presented with high accuracy, while using lower-precision models in non-critical areas. Combined with high-precision texture mapping technology, the overall visual effect is improved. in, Represents the comprehensive value of the visual effect of the non-critical area model. Indicates the effect value of high-precision texture mapping. represents the metric value of the non-critical area, represents the total metric value for the entire terrain area, Represents the visual effect value of the low-precision model itself. When the proportion of non-critical areas is larger, Close to 1, the overall visual effect value Closer to the visual effect value of the low-precision model But it will also be affected by the high-precision texture mapping effect. When the non-critical area accounts for a smaller proportion, Close to 0, Closer to the effect value of high-precision texture mapping , by adjusting and The proportion and and The value of can achieve different degrees of visual effect balance in non-critical areas, which not only ensures the efficiency of low-precision models, but also improves a certain visual experience through high-precision texture mapping. In order to further improve rendering efficiency, the system also applies detail level technology. ,in, represents the complexity of the model, Indicates the observation distance, represents the maximum observation distance, is an adjustable parameter that controls the effect of observation distance on complexity. is a small positive number, indicating that when the observation distance is very close to the maximum observation distance, the minimum complexity of the model is avoided to avoid the situation where the complexity is zero. When it is close to zero, that is, the observer is very close to the model, tends to infinity, the model complexity Mainly by The model shows higher complexity as the observation distance increases. Increase, gradually decreases, and Gradually increases, the model complexity gradually decreases, and by adjusting the parameters , it can balance the speed of change of model complexity at different observation distances according to actual needs. This technology can automatically adjust the complexity of the model according to the distance of the observer, presenting a high-precision model when observed at a close distance, and automatically reducing the model complexity at a long distance, thereby improving system performance while ensuring visual effects.

[0024] The system has a built-in powerful machine learning engine that can continuously learn and optimize modeling parameters and strategies. With the continuous accumulation of surveying and mapping data and the collection of user feedback, the machine learning engine can automatically adjust the modeling algorithm to improve the accuracy and efficiency of modeling, and build an iterative optimization cycle mechanism. By continuously repeating the modeling, evaluation and feedback collection process, the system can gradually optimize the model quality. The optimized model will be deployed in actual applications and updated regularly as needed to ensure that it always maintains the best state.

[0025] Based on the optimized 3D terrain data, a highly realistic urban terrain digital twin scene is constructed using 3D modeling technology. This scene not only contains rich terrain information, but also integrates a variety of interactive functions. In order to enhance the user experience and expand the application scenarios, the system has also developed a wealth of visualization tools and interfaces. These tools include but are not limited to terrain analysis charts, heat maps, and 3D animations, which can help users understand terrain information more intuitively and make scientific decisions. After comprehensive testing, the terrain digital twin scene construction and display module will be deployed to the urban planning department's server, and continuous technical support and maintenance services will be provided to ensure the stable operation of the system and to be upgraded and optimized at any time according to user needs.

[0026] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A terrain digital twin scene mapping system based on 3D modeling, It is characterized in that The system includes an intelligent terrain feature analysis module, an adaptive modeling strategy selection module, a self-learning and optimization module, and a terrain digital twin scene construction and display module; The intelligent terrain feature analysis module receives surveying and mapping data from multiple data sources such as drones, satellites, and ground surveying and mapping equipment, and uses deep learning algorithms to automatically analyze the terrain's undulation, landform type, and vegetation coverage density characteristics. Through feature extraction and classification, it provides basic data support for subsequent modeling strategy selection. The adaptive modeling strategy selection module receives the terrain undulation, landform type, and vegetation coverage density feature data from the intelligent terrain feature analysis module, analyzes the data, determines the complexity and feature classification of the current processed terrain, and preliminarily matches the modeling strategy according to the terrain feature classification; For flat areas, select the mesh simplification algorithm, which can identify and remove unnecessary details in the model while keeping key features to maintain model accuracy. Set algorithm parameters to balance model accuracy and data volume. ,in, represents the comprehensive evaluation value of the simplified model, A measure representing a key characteristic, Represents the total metric value of the model in its initial state, Represents the algorithm parameters, when When the value increases, it means that the amount of data is more likely to be reduced. The impact of this item increases, and more non-critical parts are removed. When the value decreases, more emphasis is placed on model accuracy. The increase in the impact of this term will better preserve key features, by adjusting The value of can balance the model accuracy and data volume. For complex terrain in mountainous areas, multi-resolution surface modeling technology is selected to apply models of different accuracy in different areas to efficiently express complex terrain and determine multi-level resolution levels. ,in, Represents the comprehensive resolution value of the complex terrain model, A measure representing an area of ​​important geomorphic features, Represents the total metric value of the entire complex terrain area, , , when the area with important geomorphic features accounts for a large proportion, that is Close to 1, at this time Closer , indicating that the entire model is presented with higher accuracy. When the proportion of non-important areas is large, Close to 0, Closer Indicates that the entire model is presented with lower precision. and The proportion and and The value of can flexibly control the accuracy of complex terrain models in different areas to achieve the purpose of efficiently expressing complex terrain. In non-critical areas, lower-precision models are used. Combined with high-precision texture mapping technology, realistic surface materials are added to the model to enhance the visual effect. Detail level technology is applied to dynamically adjust the model complexity according to the observation distance to improve rendering efficiency. According to the detailed characteristics of the specific terrain and modeling requirements, further adjust the parameters of the modeling strategy, including but not limited to adjusting the mesh resolution, texture detail level, and surface fitting algorithm, so that the final model meets the accuracy requirements and achieves the best performance; Execute the selected modeling strategy to generate a three-dimensional terrain model, verify the generated model to check whether it accurately reflects the terrain characteristics, including but not limited to terrain undulations, landform types, and vegetation cover, evaluate the data volume, processing speed, and rendering performance of the model, collect feedback information during the modeling process, and optimize and improve the modeling strategy selection process and parameter settings based on the feedback results; The self-learning and optimization module has a built-in machine learning engine that can continuously learn and optimize modeling parameters and strategies. With the continuous accumulation of surveying and mapping data, it can automatically adjust the modeling algorithm to improve the accuracy and efficiency of modeling, and continuously optimize the model quality through user feedback and model evaluation mechanism; The terrain digital twin scene construction and display module: based on the output of the above module, a highly realistic terrain digital twin scene is constructed. The scene supports multi-perspective, multi-scale browsing and interactive operations, facilitates users to conduct in-depth geographic analysis and decision support, provides rich visualization tools and interfaces, and expands the breadth and depth of application scenarios.

2. According to the terrain digital twin scene mapping system based on three-dimensional modeling as described in claim 1, It is characterized in that The terrain digital twin construction process of the intelligent terrain feature analysis module starts with data reception and preprocessing, which covers real-time or regular collection of surveying and mapping data from multiple data sources, and verification and preprocessing, including but not limited to image correction and point cloud filtering. Subsequently, feature extraction is performed to obtain detailed terrain features through terrain undulation, landform recognition and vegetation cover density analysis.

3. According to the terrain digital twin scene mapping system based on three-dimensional modeling as described in claim 2, It is characterized in that The feature classification and fusion stage of the intelligent terrain feature analysis module refines the extracted features into standardized sets and fuses multi-dimensional data to form a comprehensive data model. This data model supports the subsequent modeling strategy selection. The entire process focuses on feedback and optimization, and continuously iterates to improve the intelligence level of the system.

4. According to the terrain digital twin scene mapping system based on three-dimensional modeling as described in claim 1, It is characterized in that The adaptive modeling strategy selects modules and uses lower-precision models in non-critical areas, combined with high-precision texture mapping technology to add realistic surface materials to the model and improve visual effects: in, Represents the comprehensive value of the visual effect of the non-critical area model. Indicates the effect value of high-precision texture mapping. represents the metric value of the non-critical area, represents the total metric value for the entire terrain area, Represents the visual effect value of the low-precision model itself. When the proportion of non-critical areas is larger, Close to 1, the overall visual effect value Closer to the visual effect value of the low-precision model But it will also be affected by the high-precision texture mapping effect. When the non-critical area accounts for a smaller proportion, Close to 0, Closer to the effect value of high-precision texture mapping , by adjusting and The proportion and and The value of can achieve different degrees of visual effect balance in non-critical areas, which not only ensures the efficiency of the low-precision model, but also improves the visual experience to a certain extent through high-precision texture mapping.

5. According to the terrain digital twin scene mapping system based on three-dimensional modeling as described in claim 1, It is characterized in that The adaptive modeling strategy selection module applies the level of detail technology to dynamically adjust the model complexity according to the observation distance to improve rendering efficiency: ,in, represents the complexity of the model, Indicates the observation distance, represents the maximum observation distance, is an adjustable parameter that controls the effect of observation distance on complexity. is a small positive number, indicating that when the observation distance is very close to the maximum observation distance, the minimum complexity of the model is avoided to avoid the situation where the complexity is zero. When it is close to zero, that is, the observer is very close to the model, tends to infinity, the model complexity Mainly by The model shows higher complexity as the observation distance increases. Increase, gradually decreases, and Gradually increases, the model complexity gradually decreases, and by adjusting the parameters , the speed of change of model complexity at different observation distances can be balanced according to actual needs.

6. According to the terrain digital twin scene mapping system based on three-dimensional modeling as described in claim 1, It is characterized in that The self-learning and optimization module system continuously collects surveying and mapping data from drones, satellites and ground surveying and mapping equipment, and performs data cleaning, format unification and coordinate correction preprocessing to ensure data quality. Subsequently, the modeling parameters and strategies are initialized based on existing data and past experience, and learning goals and optimization indicators are set. By executing modeling tasks and evaluating model quality, the system collects user feedback and converts it into quantitative indicators to drive the machine learning engine to automatically adjust modeling parameters and strategies.

7. According to the terrain digital twin scene mapping system based on three-dimensional modeling as described in claim 6, It is characterized in that The above process of the self-learning and optimization module constitutes an iterative optimization loop. Through repeated modeling, evaluation and feedback collection, the model is gradually optimized. The optimized model is deployed to actual applications and updated regularly as needed. At the same time, the system continuously monitors the model performance, collects new data and user feedback, so that the machine learning engine can automatically adapt to changes in terrain characteristics and user needs.

8. According to the terrain digital twin scene mapping system based on three-dimensional modeling as described in claim 1, It is characterized in that The terrain digital twin scene construction and display module integrates optimized three-dimensional terrain data, uses three-dimensional modeling technology to build a basic framework, and designs interactive functions to enhance user experience. At the same time, it develops visualization tools to enrich the scene expressiveness and expands application scenarios to meet different needs. After comprehensive testing, the module will be deployed and provided with continuous maintenance.

Citation Information

Patent Citations

  • Digital mapping method and system based on live-action 3D model

    CN107356230A

  • Digital twinborn three-dimensional scene visual angle operation method and device

    CN113204826A

  • Engineering topographic surveying and mapping system and surveying and mapping method based on three-dimensional constructed model

    CN118393516A

  • Digital twinborn simulation optimization system for intelligent power grid operation and maintenance scene

    CN118821374A

  • ARTIFICIAL INTELLIGENCE (AI) SYSTEM USING HEIGHT SEED INITIALIZATION FOR EXTRACTION OF DIGITAL ELEVATION MODELS (DEMs) AND RELATED METHODS

    US20220092850A1

Cited By

  • Adaptive TIN digital twin terrain grid density adjustment method and device based on machine learning

    CN119229060A

  • Automatic classification design model for surveying and mapping and geographic information achievements

    CN119293135A

  • Three-dimensional simulation management method and system for intelligent stereoscopic warehouse

    CN119475825A

  • A three-dimensional simulation management method and system for intelligent stereoscopic warehouse

    CN119475825B

  • Surface quality detection method and device for high-precision metal component

    CN119936031A