Three-dimensional modeling method and system for high-altitude curtain wall

Abstract

The invention relates to the technical field of building information modeling (BIM), in particular to a three-dimensional modeling method and system of a high-altitude curtain wall, multispectral data, structured light scanning data and thermal imaging data are collected through an unmanned aerial vehicle, the data are subjected to multidimensional analysis and fusion, three-dimensional modeling is performed through a multi-modal feature fusion algorithm, and the three-dimensional modeling method and system of the high-altitude curtain wall are obtained. And carrying out global and local geometric correction, material feature optimization and thermal feature correction on the model by utilizing a dynamic feedback and recursive optimization mechanism. And finally, generating high-precision three-dimensional model data through a multi-channel fusion algorithm. The method effectively solves the problems of insufficient geometric precision, limited material processing and neglect of thermal characteristics in the prior art, remarkably improves the precision and reality of the model, and is suitable for the fields of building design, structural analysis, energy consumption evaluation and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of building information modeling (BIM), and in particular to a three-dimensional modeling method and system for high-altitude curtain walls. Background Art

[0002] As the exterior wall of modern high-rise buildings, high-altitude curtain walls not only play an important role in aesthetics, but are also an important part of the building structure. 3D modeling of high-altitude curtain walls is of great significance in many aspects such as architectural design, structural analysis, maintenance and management. Through accurate 3D modeling, designers can simulate the light and shadow effects, structural strength, and thermodynamic behavior of curtain walls under different climatic conditions. In addition, 3D models also play a vital role in the construction, cleaning and maintenance of buildings, and can detect potential structural problems in advance and formulate effective solutions.

[0003] At present, the 3D modeling method of high-altitude curtain walls (Chinese invention patent, publication number: CN117893681A, name: 3D modeling and image transmission method of high-altitude curtain walls) mainly relies on drones to collect images and remotely transmit them to image processing systems for modeling. The existing technology uses image acquisition, remote transmission, contour modeling, local feature scanning, 3D wireframe model generation and other means. Although this method can automatically generate 3D models, it has the following major defects:

[0004] Due to the resolution and angle limitations of the image acquisition device, the geometric accuracy of the model is easily affected, especially in the complex surface and connection point areas, where the error is large;

[0005] The existing technology mainly focuses on texture rendering when processing materials, but lacks fine processing of details such as the optical properties, reflection and refraction of materials, resulting in the final model not being realistic enough in visual performance;

[0006] Existing technologies hardly consider the thermal characteristics of curtain walls, which makes the model unable to simulate the behavior of curtain walls under different temperature conditions, limiting the application of the model in energy consumption analysis and thermodynamic simulation;

[0007] The existing 3D modeling process is mostly linear processing, and there is a lack of coordinated optimization between the steps, which leads to inconsistencies in the geometry, material and thermal characteristics of the final model. Summary of the invention

[0008] In view of the many problems existing in the above-mentioned prior art, the present invention provides a three-dimensional modeling method and system for high-altitude curtain walls. The present invention collects multimodal data (including geometric, material and thermal imaging data) through drones, and uses a multimodal feature fusion algorithm to deeply fuse and optimize these data. Through dynamic feedback and recursive optimization mechanisms, a high-precision three-dimensional model is generated. The model is not only accurate in geometry, but also can truly reproduce the material characteristics and thermal response of the curtain wall, and is suitable for application scenarios such as architectural design, structural analysis and energy consumption assessment.

[0009] A three-dimensional modeling method for a high-altitude curtain wall comprises the following steps:

[0010] The multi-spectral data, structured light scanning data and thermal imaging data of the high-altitude curtain wall are collected by drones, and the collected data are synchronously processed in time and space, and pre-processed fusion data is generated through preliminary fusion;

[0011] Performing multi-dimensional cross-analysis on the pre-processed fusion data, extracting geometric structure data, material feature data and thermal feature data in sequence, and generating fusion feature data through hierarchical feature fusion and dynamic feedback optimization;

[0012] Based on the fused feature data, three-dimensional modeling is performed by a multimodal feature fusion algorithm, the geometric structure data, the material feature data and the thermal feature data are deeply fused, and at the same time, optimization model data is generated by using real-time feedback and a recursive optimization mechanism;

[0013] Performing global geometric correction, local geometric refinement correction, material feature optimization and thermal feature correction on the optimization model data, and generating final corrected optimization model data through multi-dimensional collaborative optimization;

[0014] Based on the final corrected and optimized model data, geometric structure rendering, material feature rendering and thermal feature rendering are performed, and the final three-dimensional model data is generated through a multi-channel fusion algorithm.

[0015] Preferably, the multidimensional cross-analysis comprises the following steps:

[0016] The geometric structure data in the preprocessed fusion data is preliminarily analyzed to generate preliminary geometric structure data. The material feature data and thermal feature data are combined to optimize the boundary accuracy of the geometric structure data through multi-dimensional cross analysis to generate geometric feature point data. The preliminary geometric structure data is further optimized using the geometric feature point data to generate optimized geometric structure data.

[0017] Preferably, the extraction of material feature data comprises the following steps:

[0018] The material information in the preprocessed fusion data is subjected to light reflection analysis, refraction analysis and surface texture analysis to generate material optical feature data, and the material optical feature data is coupled with the optimized geometric structure data to further optimize the material feature data and generate optimized material feature data.

[0019] Preferably, the extraction of thermal characteristic data comprises the following steps:

[0020] The thermal gradient analysis is performed on the thermal imaging information in the preprocessed fusion data. The thermal feature model data is generated through the thermal feature modeling algorithm in combination with the optimized geometric structure data and the optimized material optical feature data. The thermal feature model data is then corrected and optimized in real time using a dynamic feedback mechanism to generate optimized thermal feature data.

[0021] Preferably, the multimodal feature fusion algorithm comprises the following steps:

[0022] The optimized geometric structure data, the optimized material feature data and the optimized thermal feature data are weightedly fused, where:

[0023] The weight coefficient W1 is used to indicate the importance of geometric structure data; the weight coefficient W2 is used to indicate the importance of material feature data; the weight coefficient W3 is used to indicate the importance of thermal feature data;

[0024] The weight coefficients are dynamically adjusted through a recursive optimization mechanism to generate optimized model data.

[0025] Preferably, the specific expression of the weighted fusion calculation is:

[0026] M final =W1×G+W2×M+W3×T

[0027] Among them, M final represents the optimized model data; G represents the optimized geometric structure data; M represents the optimized material feature data; T represents the optimized thermal feature data; W1, W2 and W3 represent the weight coefficients of the geometric structure data, material feature data and thermal feature data respectively.

[0028] Preferably, the global geometric correction comprises the following steps:

[0029] Based on the optimized model data, the macro-geometric structure of the three-dimensional model is globally optimized, and the macro-geometric errors are detected and eliminated using the global optimization algorithm. The global correction geometric model data is generated by adjusting the spatial position, scale and angle of the overall shape of the model.

[0030] Preferably, the local geometry refinement correction comprises the following steps:

[0031] The detail areas in the global correction geometry model data are locally optimized, the local geometry errors are corrected by the detail enhancement algorithm, and the key areas are refined by the micro-adjustment algorithm to generate the final geometry correction model data.

[0032] Preferably, the multi-channel fusion algorithm comprises the following steps:

[0033] The final geometric correction model data, optimized material feature data and optimized thermal feature data are rendered independently respectively. The geometric structure is rendered through the geometric channel, the material features are rendered through the material channel, and the thermal effect is rendered through the thermal feature channel. The rendering results of each channel are subjected to multi-channel fusion processing to generate the final 3D model data.

[0034] A system for implementing the three-dimensional modeling method of the high-altitude curtain wall, comprising:

[0035] The data acquisition module is used to collect multi-spectral data, structured light scanning data and thermal imaging data of high-altitude curtain walls through drones, and synchronously process the collected data in time and space to generate pre-processed fusion data;

[0036] The data analysis and feature extraction module is used to perform multi-dimensional cross-analysis on the pre-processed fusion data, extract geometric structure data, material feature data and thermal feature data in turn, and generate fusion feature data through hierarchical feature fusion and dynamic feedback optimization;

[0037] The 3D modeling and feature fusion module is used to perform 3D modeling based on fused feature data through a multimodal feature fusion algorithm, deeply fuse geometric structure data, material feature data, and thermal feature data, and generate optimized model data using real-time feedback and recursive optimization mechanisms;

[0038] The correction and optimization module is used to perform global geometry correction, local geometry refinement correction, material feature optimization and thermal feature correction on the optimization model data, and generate the final correction optimization model data through multi-dimensional collaborative optimization;

[0039] The rendering and output module is used to optimize the model data based on the final correction, perform geometric structure rendering, material feature rendering and thermal feature rendering, and generate the final three-dimensional model data through a multi-channel fusion algorithm.

[0040] Compared with the prior art, the advantages and beneficial effects of the present invention are:

[0041] The present invention realizes the deep fusion of geometric structure, material features and thermal features through a multi-modal feature fusion algorithm, and generates a more accurate and realistic three-dimensional model;

[0042] The present invention realizes real-time correction and optimization of the model through dynamic feedback and recursive optimization mechanism, ensuring the accuracy of the model in different application scenarios;

[0043] The present invention realizes comprehensive coordinated optimization of geometry, material and thermal characteristics through multi-dimensional collaborative optimization technology, so that the model achieves optimal effects in all dimensions. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a schematic diagram of the process of the present invention;

[0045] Figure 2 This is a schematic diagram of the data collection process of the drone in the present invention;

[0046] Figure 3 Schematic diagram of the multimodal feature fusion algorithm in the present invention;

[0047] Figure 4 It is a schematic diagram of the multi-channel fusion algorithm in the present invention;

[0048] Figure 5 It is a structural block diagram of the system of the present invention. DETAILED DESCRIPTION

[0049] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is obvious that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0050] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0051] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0052] like Figure 1 As shown, a 3D modeling method for a high-altitude curtain wall comprises the following steps:

[0053] like Figure 2As shown, the multispectral data, structured light scanning data and thermal imaging data of the high-altitude curtain wall are collected by drones, the collected data are synchronously processed in time and space, and pre-processed fusion data are generated through preliminary fusion;

[0054] The multispectral camera carried by the drone is used to collect multispectral data of high-altitude curtain walls. Multispectral data covers multiple spectral ranges such as visible light and non-visible light (such as near-infrared light, ultraviolet light), which can reveal the reflection, absorption and transmission characteristics of curtain wall materials in different bands. These spectral characteristics play an important role in identifying and distinguishing the different components of curtain wall materials, especially in detecting potential subtle differences and defects on the surface of curtain walls. For example, by analyzing the near-infrared spectrum, the uniformity of the curtain wall surface coating and potential degradation areas can be identified, which provides the necessary data support for subsequent material feature extraction and thermal feature analysis.

[0055] Structured light scanning technology is implemented through structured light scanning equipment carried by drones. The principle of this technology is to project a light beam of known pattern, such as stripes or grids, onto the surface of the curtain wall, and capture the deformation of the light beam on the surface of the object through a camera. Using these deformed patterns, high-precision three-dimensional geometric shape data can be reconstructed to capture the complex geometric structure, surface bumps, edge features and other details of the curtain wall. This is crucial for generating accurate geometric models because it ensures the accuracy of the model in shape, size and spatial position. For example, in implementation, structured light scanning can detect tiny deformations and damaged areas on the surface of the curtain wall, and this information is crucial for subsequent global geometric correction and local refinement.

[0056] Thermal imaging technology uses a thermal imaging camera mounted on a drone to collect thermal imaging data and capture the temperature distribution on the surface of the curtain wall. The principle of thermal imaging is based on the different intensities of infrared radiation emitted by objects at different temperatures. Thermal imaging cameras can convert this radiation difference into a visual temperature distribution map. By analyzing these temperature data, thermal anomaly areas in the curtain wall structure, such as degradation of the insulation layer or thermal bridge effects, can be detected. These thermal feature information is very helpful in evaluating the thermodynamic properties and potential structural problems of the curtain wall. In practical applications, thermal imaging technology can identify material fatigue or cracks caused by temperature gradients, so as to make targeted corrections and optimizations during the modeling process.

[0057] After the above data is collected, it is necessary to perform synchronous processing in time and space. The principle of this step is that since multispectral data, structured light scanning data, and thermal imaging data are collected at different times and viewing angles, these data must be aligned and corrected to ensure their consistency in the spatial coordinate system. This synchronous processing can be achieved through image registration technology and sensor fusion algorithms to ensure that all data matches the actual geometric position of the curtain wall, thereby avoiding model deviations or errors caused by data misalignment. In effect, synchronous processing in time and space can significantly improve the fusion quality of multi-source data and ensure the accuracy of subsequent modeling processes.

[0058] Through the preliminary fusion step, the synchronously processed multispectral data, structured light scanning data and thermal imaging data are preliminarily fused to generate preprocessed fused data. This fusion process involves weighted synthesis and feature extraction of information from different data sources to generate a preliminary multidimensional data set. In principle, preliminary fusion combines key information of geometry, material properties and thermal features to form a more complete description, which can play a fundamental role in subsequent multidimensional cross-analysis and feature extraction. In effect, the preliminary fusion data removes redundant information while retaining the important features of each data source, providing the necessary data foundation for subsequent high-precision modeling.

[0059] Performing multi-dimensional cross-analysis on the pre-processed fusion data, extracting geometric structure data, material feature data and thermal feature data in sequence, and generating fusion feature data through hierarchical feature fusion and dynamic feedback optimization;

[0060] Multidimensional cross-analysis is the process of cross-checking and optimizing data sources from different sensors. Its core lies in further improving the accuracy and reliability of data by analyzing the relationship between geometric structure, material characteristics and thermal characteristics in preprocessed fusion data. In the preprocessed fusion data, geometric structure data, material feature data and thermal feature data respectively reflect the spatial form, material properties and temperature distribution of the curtain wall, and this information is interrelated in physical space. Through multidimensional cross-analysis, one feature data (such as geometric data) can be used to correct and optimize other feature data (such as material and thermal characteristics), thereby eliminating deviations caused by incomplete or inaccurate data. For example, in the implementation process, the precise boundaries of the geometric structure data can be used to correct the errors caused by uneven reflected light in the material feature data, thereby improving the accuracy of the material features.

[0061] After multi-dimensional cross-analysis, geometric structure data, material feature data and thermal feature data are extracted in sequence. The purpose of extracting geometric structure data is to obtain the precise three-dimensional shape of the curtain wall, including key geometric information such as contours, edges and surface details. This information is the basis of three-dimensional modeling, ensuring that the model is consistent with the actual curtain wall in space. The extraction of material feature data focuses on the optical properties and texture characteristics of the curtain wall surface, such as reflectivity, transparency and surface roughness, which directly affect the visual performance and light and shadow effects of the model. The extraction of thermal feature data is used to capture the temperature distribution and thermal effects on the curtain wall surface, helping to identify the thermal response of the material and potential structural problems. In practical applications, the sequential extraction of these feature data ensures that each type of data can be accurately extracted under the most suitable analysis conditions, thereby improving the overall authenticity of the model.

[0062] After the feature data is extracted, the final fused feature data is generated through hierarchical feature fusion and dynamic feedback optimization. Hierarchical feature fusion refers to the fusion of geometric structure, material features and thermal features in a hierarchical manner according to the importance and priority of the feature data. Usually, the geometric structure data is first established as the skeleton of the model as the basic layer, and then the material feature data and thermal feature data are superimposed on the geometric structure in turn to form a multi-dimensional comprehensive feature model. Through this hierarchical fusion method, it can be ensured that each feature data can maintain its unique attributes during the fusion process, while forming a close synergistic relationship with other data.

[0063] Dynamic feedback optimization plays a crucial role in this process. In principle, dynamic feedback optimization monitors the intermediate data generated during the fusion process in real time and adjusts the fusion process in a timely manner based on this data. For example, when fusing material feature data, if material changes caused by thermal effects are detected, this information will be immediately fed back to the fusion algorithm to adjust the weights and fusion strategies of the material feature data to ensure that the final fusion feature data can truly reflect the performance of the curtain wall under different conditions. Through this dynamic feedback mechanism, the system can adaptively optimize the fusion strategy and significantly improve the accuracy and consistency of the final model.

[0064] The fusion feature data generated in this step not only contains the complete geometric information, material properties and thermal effect characteristics of the curtain wall, but also eliminates the contradictions and inconsistencies between various data sources through multi-dimensional cross-analysis and dynamic feedback optimization, forming a highly accurate and coordinated 3D model basic data. This provides a solid foundation for subsequent modeling and rendering, ensuring that the final 3D model reaches the optimal level in terms of visual effects and physical consistency.

[0065] Preferably, the multidimensional cross-analysis comprises the following steps:

[0066] The geometric structure data in the preprocessed fusion data is preliminarily analyzed to generate preliminary geometric structure data. The material feature data and thermal feature data are combined to optimize the boundary accuracy of the geometric structure data through multi-dimensional cross analysis to generate geometric feature point data. The preliminary geometric structure data is further optimized using the geometric feature point data to generate optimized geometric structure data.

[0067] The geometric structure data in the preprocessed fusion data is initially parsed. The principle of the preliminary parsing is to identify the basic geometric shape and main structural features of the curtain wall through the parsing algorithm. The preliminary geometric structure data generated in this step contains the general outline and key geometric elements of the curtain wall, such as panel edges, connection points, and surface curvature. However, due to the noise, occlusion or resolution limitations that may exist during data acquisition and preliminary processing, the preliminary geometric structure data often has certain errors or incompleteness, so further optimization and correction are required.

[0068] The material feature data and thermal feature data are combined with the preliminary geometric structure data, and the boundary accuracy of the geometric structure data is optimized through multi-dimensional cross analysis. The principle of multi-dimensional cross analysis is to use the inherent correlation between different feature data to further correct the errors in the geometric structure data. For example, the material feature data can provide information about the surface texture and material distribution of the curtain wall. By analyzing this information, the boundary areas in the geometric structure data can be identified and these areas can be refined and corrected. Similarly, the thermal feature data can reveal the impact of the material on the geometric shape during thermal expansion or contraction by reflecting the temperature distribution on the curtain wall surface, and then correct the boundaries of the geometric structure data. For example, in implementation, if the thermal feature data shows that a part of the curtain wall has a slight deformation due to thermal expansion and contraction, this information can be used to correct the geometric data, making the model more accurate under thermodynamic conditions.

[0069] The geometric feature point data generated by multi-dimensional cross analysis is the key to optimizing geometric structure data. These geometric feature point data represent the important nodes and boundaries in the curtain wall geometry. These nodes and boundaries play a skeleton role in the model and determine the overall shape and detailed expression of the model. For example, feature points may include panel connection points, edge intersections, and key turning points of surfaces. During the model construction process, these geometric feature points will serve as the basis for further optimization to ensure the accuracy of the model in these key areas.

[0070] The preliminary geometric structure data is further optimized using the geometric feature point data to generate optimized geometric structure data. The principle of the optimization process is to adjust and refine the entire geometric model based on the feature points through interpolation, curve fitting, or feature point-based mesh refinement technology. This optimization step can effectively eliminate discontinuities and errors in the preliminary geometric structure data, making the boundaries of the geometric model clearer and the shape more accurate. For example, in a specific implementation, the interpolation algorithm can be used to smooth the surface of the geometric structure to make the edge transition more natural, while the mesh refinement technology can add more mesh nodes around the feature points to improve the resolution and accuracy of the model.

[0071] The optimized geometric structure data generated through the above steps has a high degree of accuracy and consistency, laying a solid foundation for subsequent 3D modeling and rendering. This process not only ensures the morphological accuracy of the geometric model, but also enables the model to reflect the actual performance of the curtain wall under different environmental conditions by combining material and thermal characteristic data, thereby enhancing the application value of the model. In the fields of building structure analysis, thermodynamic research, and visual simulation, the optimized geometric structure data can provide a more reliable reference, making subsequent analysis and decision-making more accurate.

[0072] Preferably, the extraction of material feature data comprises the following steps:

[0073] The material information in the preprocessed fusion data is subjected to light reflection analysis, refraction analysis and surface texture analysis to generate material optical feature data, and the material optical feature data is coupled with the optimized geometric structure data to further optimize the material feature data and generate optimized material feature data.

[0074] Perform light reflection analysis on the material information in the preprocessed fusion data. The principle of light reflection analysis is to determine the optical properties of the material, such as reflectivity, glossiness, and specular reflection ratio, by calculating and measuring the reflection characteristics of the curtain wall surface to different lights. These optical properties directly affect the visual effect of the model. For example, materials with high reflectivity will show more obvious highlights and reflection effects under light, while materials with low reflectivity will show matte or diffuse reflection. In practical applications, by performing light reflection analysis on the preprocessed fusion data, the reflection characteristics of the curtain wall material under different lighting conditions can be accurately extracted. For example, the reflection characteristics of the glass curtain wall used on the facade of the building can accurately reflect the changes in ambient light, thereby showing realistic reflection effects in the three-dimensional model.

[0075] Refraction analysis is the study of the propagation path of light inside a material to determine the refractive index and light transmittance of the material. The principle of refraction analysis is based on the phenomenon that when light enters a medium of different density, the propagation direction will be deflected. For transparent or translucent materials such as glass or polymers used in high-altitude curtain walls, the refractive property is an important factor in determining their visual performance. Through refraction analysis, the path of light inside these materials can be accurately calculated to generate the refractive index data of the material. In the implementation process, refraction analysis can reveal the light transmission effect and internal refraction phenomenon of the glass curtain wall, so that the model can truly reproduce the transmission image of light after passing through the glass, thereby enhancing the realism and visual impact of the model.

[0076] Surface texture analysis is to analyze the microstructure of the curtain wall surface to obtain the surface roughness, texture direction and detail characteristics of the material. The principle of surface texture analysis is to reconstruct the texture pattern and details of the material by scanning and data processing the microstructure of the material surface. Different surface textures will significantly affect the visual performance of the material. For example, a rough surface will produce more diffuse reflection, while a smooth surface will produce stronger specular reflection. In specific implementation, through surface texture analysis, the texture pattern data of the curtain wall surface can be generated and applied to the model to ensure the realistic reproduction of surface details. For example, the subtle brushed texture of a metal curtain wall can be accurately simulated through this process, enhancing the texture and visual effect of the model.

[0077] Through the above light reflection analysis, refraction analysis and surface texture analysis, material optical feature data are generated. These data contain the complete optical properties of the material and reflect the performance of the curtain wall under different lighting and viewing angles. However, in order to ensure the perfect match between the material feature data and the geometric model, the material optical feature data must be coupled with the optimized geometric structure data. The principle of coupling calculation is that by closely combining the material characteristics with the geometric shape, the material properties can be reflected in an accurate geometric coordinate system. This process can eliminate material performance deviations caused by geometric errors or misalignment. For example, by coupling the refractive index data with the geometric model of the glass curtain wall, it can be ensured that the light will show accurate refraction effects when passing through glass of different thicknesses, thereby enhancing the authenticity of the model.

[0078] Through coupled calculation and further optimization, optimized material feature data is generated. This optimization process ensures a high degree of consistency between the material feature data and the geometric structure data, and eliminates any possible discontinuity or error by adjusting the spatial distribution of the material features. In effect, the optimized material feature data not only improves the visual effect of the model, but also ensures the accuracy of the model in physical simulation. For example, in thermodynamic analysis, accurate material feature data can accurately reflect the thermal response of the material, thereby improving the reliability of the analysis.

[0079] Preferably, the extraction of thermal characteristic data comprises the following steps:

[0080] The thermal gradient analysis is performed on the thermal imaging information in the preprocessed fusion data. The thermal feature model data is generated through the thermal feature modeling algorithm in combination with the optimized geometric structure data and the optimized material optical feature data. The thermal feature model data is then corrected and optimized in real time using a dynamic feedback mechanism to generate optimized thermal feature data.

[0081] Thermal gradient analysis is a preliminary step in thermal feature data extraction. The principle of thermal gradient analysis is to calculate the temperature gradient of different areas on the curtain wall surface based on the temperature distribution data in the thermal imaging information. The temperature gradient refers to the rate of change of temperature in space, which can reveal the thermal conductivity characteristics and thermal stress distribution of the curtain wall in different areas. Through thermal gradient analysis, it is possible to identify thermal anomaly areas on the curtain wall surface, such as thermal bridges, insulation defects or material fatigue points. During implementation, thermal gradient analysis can accurately locate these abnormal areas and provide key input data for subsequent thermal feature modeling. For example, on the curtain wall of a high-altitude building in winter, thermal imaging may show a thermal bridge effect caused by the temperature difference between indoor and outdoor. Through thermal gradient analysis, these areas can be accurately located, so that the real thermal response of the curtain wall can be reflected in the model.

[0082] Thermal feature modeling is performed by combining the optimized geometric structure data and the optimized material optical feature data. The principle of thermal feature modeling is to generate a thermal feature model that can accurately reflect the characteristics of heat conduction, thermal radiation and thermal convection by closely combining thermal imaging data with geometric structure and material characteristics. The geometric structure data provides the spatial shape and structural information of the curtain wall, which is crucial to determining how heat propagates on the surface and inside the curtain wall. The material optical feature data reflects the absorptivity, reflectivity and transmittance of the material, which affect the absorption and dissipation of heat. In the thermal feature modeling process, these data are integrated into a unified physical model that can simulate the thermal behavior of the curtain wall under different environmental conditions. For example, through thermal feature modeling, the temperature change on the curtain wall surface under sunlight can be simulated, and the possible thermal stress concentration area can be predicted, which is of great significance for the safety assessment and energy consumption optimization of the building.

[0083] After the thermal characteristic model data is generated, the dynamic feedback mechanism is used to perform real-time correction and optimization on the thermal characteristic model data. The principle of the dynamic feedback mechanism is to monitor and analyze the performance of the thermal characteristic model under different simulation conditions in real time, and adjust the model parameters in time to ensure the accuracy and consistency of the model. This mechanism can dynamically correct the thermal characteristic model according to new input data or deviations in the model operation to avoid errors in the model due to initial settings or incomplete data. For example, if it is found during the modeling process that the thermal conductivity coefficient of a certain material does not match the actual one, the dynamic feedback mechanism can automatically adjust the parameter to correct the overall thermal characteristic performance of the model. In specific implementation, this mechanism can significantly improve the robustness and accuracy of the thermal characteristic model, ensuring that the generated thermal characteristic data can truly reflect the thermodynamic behavior of the curtain wall.

[0084] After correction and optimization by the dynamic feedback mechanism, optimized thermal characteristic data is generated. The optimized thermal characteristic data not only includes the accurate temperature distribution and heat conduction path of the curtain wall under different temperature conditions, but also integrates the influence of geometric structure and material characteristics to form a highly accurate thermal characteristic model. In effect, these data can be used to simulate the thermal stress distribution of the curtain wall under different climatic conditions, predict possible thermal fatigue areas, and optimize the thermal insulation performance and energy efficiency of the building. For example, in tropical climate conditions, the optimized thermal characteristic data can be used to analyze the risk of thermal deformation of the curtain wall under long-term high temperature exposure, thereby providing an important reference for building design and maintenance.

[0085] Based on the fused feature data, three-dimensional modeling is performed by a multimodal feature fusion algorithm, the geometric structure data, the material feature data and the thermal feature data are deeply fused, and at the same time, optimization model data is generated by using real-time feedback and a recursive optimization mechanism;

[0086] like Figure 3 As shown, preferably, the multimodal feature fusion algorithm includes the following steps:

[0087] The optimized geometric structure data, the optimized material feature data and the optimized thermal feature data are weightedly fused, where:

[0088] The weight coefficient W1 is used to indicate the importance of geometric structure data; the weight coefficient W2 is used to indicate the importance of material feature data; the weight coefficient W3 is used to indicate the importance of thermal feature data;

[0089] The weight coefficients are dynamically adjusted through a recursive optimization mechanism to generate optimized model data.

[0090] The multimodal feature fusion algorithm is a core algorithm that deeply fuses feature information from different data sources, such as geometric structure data, material feature data, and thermal feature data. The principle is that although each feature data is important in describing different aspects of the curtain wall, it is difficult to fully reflect the overall state and performance of the curtain wall with only a single data source. Through multimodal feature fusion, these different data sources can be integrated in a unified model to form a more comprehensive and accurate three-dimensional model. For example, geometric structure data can describe the shape and size of the curtain wall, but if it is not combined with material feature data, it is difficult to truly reproduce the surface optical effect of the curtain wall; similarly, thermal feature data can provide thermodynamic information, but without the support of geometric structure data, its spatial positioning may not be accurate enough. Therefore, multimodal feature fusion is a necessary process for three-dimensional modeling of high-altitude curtain walls. It ensures the integrity and consistency of the model by coordinating the relationship between different data sources.

[0091] In a specific implementation, the multimodal feature fusion algorithm includes weighted fusion of optimized geometric structure data, optimized material feature data, and optimized thermal feature data. The principle of weighted fusion is that, considering that different feature data have different degrees of influence on model accuracy, a weight coefficient is assigned to each feature data to reflect its importance in the overall model. These weight coefficients are W1, W2, and W3, respectively, where W1 is used to indicate the importance of geometric structure data, W2 is used to indicate the importance of material feature data, and W3 is used to indicate the importance of thermal feature data. The setting of the weight coefficient can be adjusted according to the specific needs of modeling. For example, in an application that emphasizes geometric accuracy more, the weight of W1 can be increased, while in a scene that pays more attention to material performance or thermal effects, the weight of W2 or W3 can be higher.

[0092] It is worth noting that these weight coefficients are not static, but are dynamically adjusted through a recursive optimization mechanism. The principle of the recursive optimization mechanism is to gradually adjust each weight coefficient through a repeated iterative process so that the final generated model can achieve the optimal balance in multiple dimensions. In each iteration, the algorithm will compare the difference between the expected effect and the actual result based on the performance of the current model, and then correct the weight coefficient. For example, in the initial iteration of the model, if it is found that the simulation accuracy of the thermal feature is not enough, the algorithm can increase the weight of W3, so that the thermal feature data occupies a more important position in the subsequent fusion. As the number of iterations increases, various aspects of the model will tend to balance, and the optimized model data finally generated will have a high degree of accuracy and consistency.

[0093] Through this recursive optimization mechanism, the optimized model data generated in the end not only achieves the best state in terms of geometry, material performance and thermal response, but also can adapt to the needs of different application scenarios. For example, in the architectural design process, the optimized model data can be used to simulate the effects of sunlight exposure in different time periods, thereby helping designers optimize the appearance and energy efficiency of the building; in structural safety assessments, these data can be used to predict the performance of curtain walls under extreme temperature conditions, thereby providing a scientific basis for maintenance decisions.

[0094] In summary, 3D modeling through multimodal feature fusion algorithm and optimized model data generated through weighted fusion and recursive optimization mechanism play a vital role in the 3D modeling process of high-altitude curtain walls. It can not only integrate geometric, material and thermal feature data to generate a comprehensive and accurate 3D model, but also ensure the efficiency and reliability of the model in practical applications through dynamic adjustment and optimization. The application of this method provides strong technical support for fields such as architectural design, structural analysis and thermodynamic research, ensuring that the model achieves optimal performance in all aspects.

[0095] Preferably, the specific expression of the weighted fusion calculation is:

[0096] M final =W1×G+W2×M+W3×T

[0097] Among them, M final represents the optimized model data; G represents the optimized geometric structure data; M represents the optimized material feature data; T represents the optimized thermal feature data; W1, W2 and W3 represent the weight coefficients of the geometric structure data, material feature data and thermal feature data respectively.

[0098] M in the formula final It represents the final optimized model data, which is the result of weighted fusion of geometric structure data (G), material feature data (M) and thermal feature data (T). In the 3D modeling process, geometry, material and thermal features represent the spatial form, surface optical properties and thermodynamic behavior of the curtain wall respectively. These three types of data together determine the accuracy and authenticity of the model.

[0099] G stands for optimized geometric data. The geometric data provides basic structural information such as the three-dimensional shape, boundary lines, and surface curvature of the curtain wall, which is the basis for modeling. Without accurate geometric data, the spatial layout and form of the model will be deviated.

[0100] M represents the optimized material feature data. The material feature data includes the optical properties of the curtain wall surface, such as reflectivity, refractive index, surface roughness, etc. These properties determine the visual performance of the model. The accuracy of the material data directly affects the realism and visual effects of the model, such as the light and shadow effects under sunlight.

[0101] T is the optimized thermal characteristic data. The thermal characteristic data reflects the thermal response of the curtain wall material at different temperatures, such as thermal expansion coefficient, thermal conductivity and temperature distribution. These characteristics are crucial for simulating the thermodynamic behavior of the curtain wall, especially under extreme environmental conditions.

[0102] The core idea of ​​weighted fusion is to assign appropriate weights to these three types of data according to their importance in different modeling tasks. W1, W2, and W3 in the formula correspond to the weight coefficients of geometric structure, material characteristics, and thermal characteristics data, respectively. These weight coefficients are set to balance the influence of different feature data according to specific needs when generating the final model. For example, in a modeling task that focuses on visual effects, the weight of material feature data may be higher; while in a scene that focuses on thermodynamic analysis, the weight of thermal feature data may be increased.

[0103] In actual implementation, weighted fusion calculation is a dynamic process, and the values ​​of W1, W2 and W3 are adjusted through a recursive optimization mechanism so that the model can gradually approach the optimal state. The principle of the recursive optimization mechanism is to gradually adjust the weight coefficients through repeated iterations and error correction to ensure that the final model achieves the expected accuracy and performance in all aspects. For example, during the initial modeling, a higher geometric data weight may be set to ensure the basic structural accuracy of the model; then, by correcting the material performance and thermal response of the model, the weights of W2 and W3 are gradually increased until the final generated model reaches a balance in terms of geometry, material and thermodynamic performance.

[0104] In effect, weighted fusion calculation ensures that the integration of multimodal data is not only accurate at the physical level, but also realistic at the visual and functional levels. For example, in the sunlight simulation of high-altitude curtain walls, the optimized model data M final It can accurately reproduce the light and shadow effects of curtain walls in different time periods and under different lighting conditions, while taking into account the impact of thermal expansion on geometric shapes and the optical properties of materials, thus providing a more accurate reference for architectural design.

[0105] In summary, weighted fusion calculation generates highly optimized 3D model data M by integrating geometric structure data, material feature data and thermal feature data. finalThis process ensures the accuracy and practicality of the model in different application scenarios by reasonably setting and dynamically adjusting the weight coefficients. This method can not only improve the physical accuracy of the 3D model, but also enhance the comprehensive performance of the model in architectural design, structural analysis and thermodynamic simulation, providing a powerful technical support platform for 3D modeling of high-altitude curtain walls.

[0106] Performing global geometric correction, local geometric refinement correction, material feature optimization and thermal feature correction on the optimization model data, and generating final corrected optimization model data through multi-dimensional collaborative optimization;

[0107] Global geometry correction is a macro adjustment of the entire 3D model to ensure the consistency of the overall structure of the model with the actual curtain wall. The principle is to analyze and optimize the global geometric errors that may exist in the model data, such as overall shape deformation, scale inconsistency or posture deviation, and use the global optimization algorithm to correct these errors. For example, due to the limitation of the sensor viewing angle during data acquisition, the overall shape of the model may be slightly distorted. Global geometry correction can correct these deviations to the precise shape that conforms to the actual curtain wall structure by adjusting the coordinate system and shape parameters of the model. In effect, global geometry correction ensures that the 3D model is consistent with the actual shape of the curtain wall as a whole, so that the model can accurately reflect the actual situation in large-scale building structures.

[0108] After completing the global geometry correction, perform local geometry refinement correction. The principle of local geometry refinement correction is to make more precise geometric adjustments to key areas or details in the model, especially to correct minor errors in complex structures, connection points or curved areas. Through refinement correction, the resolution and detail performance of the model can be further improved to ensure that these key areas will not cause functional or visual defects due to geometric inaccuracies in subsequent applications. For example, in the connection point area of ​​the curtain wall, there may be problems with uneven edges or misaligned connections due to the resolution limitations of the initial data. Through local refinement correction, these subtle geometric errors can be corrected to generate a more accurate and smooth 3D model. In effect, this process makes the key details of the model more realistic, which is particularly important in high-resolution visualization or fine structure analysis.

[0109] Material feature optimization is a further adjustment to the material performance of the model. The principle of material feature optimization is to readjust the mapping method of material feature data according to the model form after global and local geometric correction to ensure the consistency and correctness of material properties on the entire model surface. For example, during the geometric correction process, the model surface may cause the material texture to stretch or compress due to shape adjustment. Material feature optimization can correct these inconsistencies by re-adjusting the texture mapping parameters so that the material texture can still be naturally distributed on the adjusted geometric form. In addition, material feature optimization can also include further adjustments to the optical properties of the material to cope with changes in lighting conditions or the needs of specific scenes. For example, in practical applications, the reflectivity and roughness parameters can be adjusted to make the model more realistic under different lighting conditions. In effect, material feature optimization ensures that the visual performance of the three-dimensional model is consistent with the material effect of the real curtain wall, so that the model not only has accurate geometric form, but also exhibits optical properties consistent with the actual material.

[0110] Thermal feature correction is an adjustment to the thermodynamic performance of the model. Its principle is to ensure that the model can accurately reflect the thermal response of the curtain wall under different temperature conditions by correcting the thermal feature data in the model. Thermal feature correction includes corrections to the heat conduction path, thermal expansion effect, and temperature distribution. In particular, after geometric and material adjustments, the thermal feature data may need to be recalibrated to ensure its accuracy on the new geometric structure and material mapping. For example, during the geometric correction process, if some parts of the model undergo slight deformation, this may affect the heat conduction path in the area, and thus affect the performance of the thermal effect. Through thermal feature correction, the thermodynamic parameters of these areas can be recalculated to ensure the accuracy of the model in thermal analysis. In effect, thermal feature correction ensures the accuracy of the model when simulating thermodynamic behavior, making the model more practical and reliable in applications such as energy consumption analysis and thermal stress assessment.

[0111] The entire process combines the correction and optimization of the three dimensions of geometry, material and thermal characteristics through multi-dimensional collaborative optimization to generate the final corrected and optimized model data. The principle of multi-dimensional collaborative optimization is to perform global optimization in a systematic way by simultaneously considering the mutual influence and synergy between different feature data, ensuring that all aspects of the model remain consistent during the correction and optimization process. For example, geometric correction may affect the distribution of material texture, and material optimization may affect the performance of thermal characteristics. Multi-dimensional collaborative optimization takes these factors into consideration to ensure that each optimization step can be carried out in a coordinated and consistent manner. The model data finally generated is both accurate and unified, avoiding conflicts or inconsistencies between the various optimization processes.

[0112] The final corrected and optimized model data has extremely high accuracy and consistency, and can truly reflect the geometric form, material characteristics and thermodynamic behavior of high-altitude curtain walls under different environmental conditions. This makes the model widely applicable in many fields such as architectural design, structural analysis, and energy consumption assessment. It not only improves the accuracy of modeling, but also enhances the actual application effect of the model, providing strong technical support for the construction industry.

[0113] Preferably, the global geometric correction comprises the following steps:

[0114] Based on the optimized model data, the macro-geometric structure of the three-dimensional model is globally optimized, and the macro-geometric errors are detected and eliminated using the global optimization algorithm. The global correction geometric model data is generated by adjusting the spatial position, scale and angle of the overall shape of the model.

[0115] The core of global geometry correction is to globally optimize the macroscopic geometry of the 3D model based on the optimized model data. The macroscopic geometry refers to the overall shape of the model, including the model's spatial position, scale, angle and other properties. During data acquisition and preliminary modeling, the model may deviate from the macroscopic view due to sensor errors, viewing angle limitations or environmental factors. For example, the overall shape of the model may be slightly distorted, or the spatial position of the model may not be exactly the same as the actual position. If these macroscopic geometric errors are not corrected, the model will become inaccurate in subsequent applications, affecting the reliability of the overall modeling.

[0116] In order to eliminate these macro-geometric errors, global geometry correction uses a global optimization algorithm for comprehensive analysis and adjustment. The principle of the global optimization algorithm is to identify the deviations in the model through mathematical analysis of the overall morphology of the model, and eliminate these deviations by adjusting the model's spatial position, scale, angle and other parameters. This process usually involves repositioning the model's overall coordinate system and correcting the model's scale and rotation. For example, if the model has a slight rotation deviation during the initial modeling process, the global optimization algorithm can adjust the model's rotation angle by calculating the model's overall angular offset to make it consistent with the actual curtain wall's posture. Similarly, if the model is stretched or compressed in a certain direction, the algorithm can adjust the model's scale factor to restore it to the correct size.

[0117] In actual implementation, global geometry correction usually includes several key steps: first, by scanning and analyzing the optimized model data as a whole, the macro-geometric errors in the model are identified; then, based on these errors, the spatial position, scale and angle of the model are adjusted; finally, the globally corrected geometry model data is generated. These steps ensure that the model is completely consistent with the actual building structure in terms of overall morphology. For example, when processing a curtain wall model of a high-rise building, global geometry correction can ensure that the overall height, width and depth of the model are consistent with the actual building, and the position and posture of the model in three-dimensional space also match the actual structure.

[0118] Global geometry correction significantly improves the overall accuracy of the 3D model, making the model not only accurate in details but also highly consistent with the actual situation in overall form. This is crucial for applications such as structural analysis, architectural design and visualization of high-altitude curtain walls. For example, in architectural design, a 3D model that has undergone global geometry correction can be used to accurately simulate the overall appearance and structure of a building, helping designers evaluate the visual effects and spatial layout of a building from different perspectives; in structural analysis, the corrected model can be used to accurately calculate the stress and stability of the building, ensuring the safety of the building in actual use.

[0119] Preferably, the local geometry refinement correction comprises the following steps:

[0120] The detail areas in the global correction geometry model data are locally optimized, the local geometry errors are corrected by the detail enhancement algorithm, and the key areas are refined by the micro-adjustment algorithm to generate the final geometry correction model data.

[0121] The starting point of local geometry refinement correction is to identify and optimize the detail areas of the global correction geometry model data. Although global geometry correction has eliminated the macroscopic errors of the model, there may still be small geometric errors in some key detail areas, such as curtain wall connection points, edge processing, complex surfaces and decorative details. These errors are usually caused by the resolution limitation of the initial data acquisition or the neglect of details during the global optimization process. However, in 3D modeling, these detail areas often determine the final quality and accuracy of the model, so local refinement processing is required.

[0122] In order to correct these local geometric errors, the detail enhancement algorithm is first applied. The principle of the detail enhancement algorithm is to capture more geometric features and correct the errors in these areas by increasing the resolution of the detail areas in the model. For example, at the edge of the curtain wall, the detail enhancement algorithm can increase the point density of the edge through interpolation or subdivision algorithms, so that the edge lines are smoother and more precise. In addition, for the curved surface area, the algorithm can further optimize the curvature and smoothness of the surface, making the surface transition more natural and avoiding the uneven surface caused by insufficient resolution. In the implementation process, the detail enhancement algorithm usually includes multi-level subdivision and curve fitting technology to ensure that each detail area is fully optimized.

[0123] Based on the detail enhancement algorithm, the micro-adjustment algorithm is used to refine key areas. The principle of the micro-adjustment algorithm is to accurately fine-tune the most critical areas of the model to ensure that the geometric characteristics of these areas are exactly consistent with the actual physical structure. Critical areas usually include joints, fixing points, corners and any parts with complex geometric shapes in the curtain wall. In these areas, the micro-adjustment algorithm corrects any minor errors caused by the data collection or modeling process by accurately fitting curves and surfaces to the local geometry. For example, at the connection points of the curtain wall, the micro-adjustment algorithm can ensure the precise position and shape of each connection point, so that the entire structure is highly consistent both visually and structurally.

[0124] This series of local optimization and micro-adjustment ultimately generated the final geometric correction model data. These data are not only highly consistent with the actual curtain wall structure in general, but also achieve extremely high accuracy in details, ensuring that the model can truly reflect the complex structure and fine features of the curtain wall.

[0125] The local geometry refinement correction significantly improves the detail performance of the 3D model, making the model not only accurate in the overall structure, but also achieving high resolution and high precision in every subtle area. For example, in the visualization of the building, the model after refinement correction can clearly show every detail of the curtain wall, from the connection points to the decorative elements, without omission; in the structural analysis, the refinement correction ensures that the geometric data of the model can accurately reflect the actual structure, thereby improving the reliability of the analysis.

[0126] Based on the final corrected and optimized model data, geometric structure rendering, material feature rendering and thermal feature rendering are performed, and the final three-dimensional model data is generated through a multi-channel fusion algorithm.

[0127] like Figure 4 As shown, preferably, the multi-channel fusion algorithm includes the following steps:

[0128] The final geometric correction model data, optimized material feature data and optimized thermal feature data are rendered independently respectively. The geometric structure is rendered through the geometric channel, the material features are rendered through the material channel, and the thermal effect is rendered through the thermal feature channel. The rendering results of each channel are subjected to multi-channel fusion processing to generate the final 3D model data.

[0129] Geometric rendering is the most basic rendering step in 3D modeling. The principle of geometric rendering is to optimize the model data according to the final correction and accurately visualize the 3D geometric shape of the curtain wall. This includes the rendering of the vertices, edges and faces of the model, as well as the presentation of the spatial layout and shape of the model. The key to geometric rendering is to accurately reproduce the macroscopic form and detailed structure of the curtain wall, so that the final model can faithfully reflect the geometric characteristics of the actual building. For example, in implementation, geometric rendering will use lighting models and shadow algorithms to realistically reproduce the appearance of the building curtain wall, ensuring that the model has consistent shape performance and light and shadow effects under different lighting conditions. In effect, geometric rendering lays the foundation for the overall appearance of the model, ensuring the accuracy and visual integrity of the model in three-dimensional space.

[0130] Next is material feature rendering, which is responsible for mapping the optical properties of the material to the geometric structure. The principle of material feature rendering is to accurately render the reflection, refraction, texture and color of the curtain wall surface based on the optimized material feature data. The purpose of material feature rendering is to make the model not only realistic in form, but also highly realistic in visual effect. For example, in the building curtain wall, material feature rendering can accurately express the transparency and reflection characteristics of glass, the glossiness of metal, and the surface texture and texture of other materials. In the implementation process, material feature rendering is usually achieved through advanced rendering technology such as PBR (physical based rendering), which can simulate optical phenomena in the real world, making the model visually closer to the actual physical material. In effect, material feature rendering gives the model a sense of visual reality, making the model not only accurate in geometry, but also highly realistic in material expression.

[0131] Thermal feature rendering is a special and important part of 3D modeling. It simulates the thermal response of curtain wall materials under different temperature conditions. The principle of thermal feature rendering is to use optimized thermal feature data to present the performance of materials when heated or cooled, such as temperature distribution, thermal radiation and thermal conductivity effects. Thermal feature rendering is essential for building energy consumption analysis and thermodynamic simulation. For example, in practical applications, thermal feature rendering can show the temperature changes of curtain walls under direct sunlight, or the thermal radiation effect during cooling at night. Through the visualization of these thermal features, designers and engineers can intuitively understand the thermal behavior of buildings under different environmental conditions, so as to make corresponding design or maintenance decisions. In effect, thermal feature rendering makes the 3D model not only accurate in static state, but also can dynamically show the real response of the building under temperature changes, which enhances the application value of the model in energy consumption analysis and thermodynamic research.

[0132] The multi-channel fusion algorithm is a key step in integrating the rendering results of geometric structure, material features and thermal features into a unified 3D model. The principle is that after independently rendering each channel (i.e. geometry, material, thermal features), the rendering results of these channels are superimposed into a final model using a fusion algorithm. This multi-channel fusion ensures the consistency of the model in all dimensions, so that there are no contradictions or inconsistencies between the geometric shape, material performance and thermal response. For example, the shadow effect generated in the geometric structure rendering needs to be coordinated with the glossiness and reflectivity in the material feature rendering, and the thermal radiation effect in the thermal feature rendering also needs to be considered. The multi-channel fusion algorithm balances and integrates these different dimensions of information to ensure that the final generated model has a high degree of consistency and accuracy in both visual performance and physical simulation.

[0133] In actual implementation, the multi-channel fusion algorithm usually includes the following steps: first, the geometric structure, material characteristics and thermal characteristics are rendered independently; then, the rendering results of each channel are pre-processed to ensure their consistency in spatial position and data format; finally, these pre-processed rendering results are combined through weighted superposition or other fusion techniques to generate the final 3D model data. For example, in the visualization process of architectural design, the fused 3D model can not only accurately display the shape and material of the curtain wall, but also simulate its thermal response in real time, providing a comprehensive reference for design and decision-making.

[0134] Through the multi-channel fusion of geometric structure rendering, material feature rendering and thermal feature rendering, the 3D modeling of high-altitude curtain walls can not only truly reproduce the appearance of the building in form and vision, but also simulate the dynamic performance of the building under different environmental conditions. The multi-channel fusion algorithm ensures the comprehensiveness and consistency of the final model data, making the model have a wide range of application value in the fields of architectural design, structural analysis, thermodynamics research, etc., providing comprehensive technical support for the construction industry.

[0135] like Figure 5 As shown, a system for implementing the three-dimensional modeling method of the high-altitude curtain wall comprises:

[0136] The data acquisition module is used to collect multispectral data, structured light scanning data and thermal imaging data of high-altitude curtain walls through drones, and synchronously process the collected data in time and space to generate pre-processed fusion data; this module collects multispectral data, structured light scanning data and thermal imaging data of high-altitude curtain walls through drones. Multispectral data is used to capture the optical properties of curtain walls, structured light scanning data is used to obtain the geometric shape of curtain walls, and thermal imaging data is used to record the temperature distribution on the surface of curtain walls. In order to ensure the temporal and spatial consistency of data, the collected data is synchronously processed to generate pre-processed fusion data. The key to this step is to accurately collect and synchronously process multi-source data to ensure the accuracy and consistency of subsequent modeling.

[0137] The data analysis and feature extraction module is used to perform multi-dimensional cross-analysis on the pre-processed fusion data, extract the geometric structure data, material feature data and thermal feature data in turn, and generate fusion feature data through hierarchical feature fusion and dynamic feedback optimization; This module performs multi-dimensional cross-analysis on the pre-processed fusion data to extract the geometric structure data, material feature data and thermal feature data of the curtain wall. Through hierarchical feature fusion and dynamic feedback optimization, fusion feature data is generated. The core here is multi-dimensional data analysis. Through layer-by-layer analysis, key features of different dimensions are extracted, and data is optimized on this basis to provide high-quality input for subsequent modeling processes.

[0138] The 3D modeling and feature fusion module is used to perform 3D modeling based on fused feature data through a multimodal feature fusion algorithm, deeply fuse geometric structure data, material feature data, and thermal feature data, and generate optimized model data using real-time feedback and recursive optimization mechanisms; based on fused feature data, this module performs 3D modeling through a multimodal feature fusion algorithm, deeply fuses geometric structure data, material feature data, and thermal feature data. Using real-time feedback and recursive optimization mechanisms, the model data is continuously corrected and optimized, and a highly accurate optimized model is finally generated. This process ensures the high consistency and accuracy of the 3D model in terms of geometry, material performance, and thermal effects.

[0139] The correction and optimization module is used to perform global geometric correction, local geometric refinement correction, material feature optimization and thermal feature correction on the optimized model data, and generate the final corrected optimized model data through multi-dimensional collaborative optimization; This module performs global geometric correction, local geometric refinement correction, material feature optimization and thermal feature correction on the optimized model data. Through multi-dimensional collaborative optimization, the final corrected optimized model data is generated. The key here is to make fine adjustments to the model to ensure that it is not only accurate in overall form, but also highly consistent with the actual curtain wall structure in details.

[0140] The rendering and output module is used to perform geometric structure rendering, material feature rendering and thermal feature rendering based on the final correction and optimization model data, and generate the final 3D model data through a multi-channel fusion algorithm. Based on the final correction and optimization model data, this module is responsible for geometric structure rendering, material feature rendering and thermal feature rendering. Through a multi-channel fusion algorithm, these rendering results are integrated to generate the final 3D model data. This process not only ensures the visual authenticity of the model, but also enables the model to accurately simulate the optical and thermodynamic properties of the actual curtain wall. The final output 3D model has a high degree of visual and functional consistency.

[0141] In practical applications, after extracting the construction drawing information, the extracted information is classified and summarized, and a preliminary three-dimensional model is generated through a modeling tool. This model is supported by multi-user collaboration and cloud platform sharing services to support the overall modeling needs of the project. The modeling tool first generates a wireframe model, then converts the wireframe model into a controlled model, and assigns facade information to the controlled model. Furthermore, the facade information is generated into a unit plate type and stored in a database, and each unit plate is assigned a corresponding building address. After the aluminum profile mold drawing is transferred to the database, it is optimized and finally a parameterized three-dimensional machining model is generated for precise processing by the factory. After that, through a series of modular processing of the system of the present invention, the multimodal data of the high-altitude curtain wall is gradually collected, parsed, modeled, corrected and rendered, and finally a highly accurate three-dimensional model with real physical properties is generated. This model can be applied to architectural design, structural analysis and thermodynamic research to provide comprehensive curtain wall performance.

[0142] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware.

[0143] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A three-dimensional modeling method for a high-altitude curtain wall, characterized in that: The following steps are involved: The multi-spectral data, structured light scanning data and thermal imaging data of the high-altitude curtain wall are collected by drones, and the collected data are synchronously processed in time and space, and pre-processed fusion data is generated through preliminary fusion; Performing multi-dimensional cross-analysis on the pre-processed fusion data, extracting geometric structure data, material feature data and thermal feature data in sequence, and generating fusion feature data through hierarchical feature fusion and dynamic feedback optimization; Based on the fused feature data, three-dimensional modeling is performed by a multimodal feature fusion algorithm, the geometric structure data, the material feature data and the thermal feature data are deeply fused, and at the same time, optimization model data is generated by using real-time feedback and a recursive optimization mechanism; Performing global geometric correction, local geometric refinement correction, material feature optimization and thermal feature correction on the optimization model data, and generating final corrected optimization model data through multi-dimensional collaborative optimization; Based on the final corrected and optimized model data, geometric structure rendering, material feature rendering and thermal feature rendering are performed, and the final three-dimensional model data is generated through a multi-channel fusion algorithm.

2. The high-altitude curtain wall three-dimensional modeling method according to claim 1 is characterized in that: The multidimensional cross-analysis comprises the following steps: The geometric structure data in the preprocessed fusion data is preliminarily analyzed to generate preliminary geometric structure data. The material feature data and thermal feature data are combined to optimize the boundary accuracy of the geometric structure data through multi-dimensional cross analysis to generate geometric feature point data. The preliminary geometric structure data is further optimized using the geometric feature point data to generate optimized geometric structure data.

3. The high-altitude curtain wall three-dimensional modeling method according to claim 1 is characterized in that: The extraction of material feature data comprises the following steps: The material information in the preprocessed fusion data is subjected to light reflection analysis, refraction analysis and surface texture analysis to generate material optical feature data, and the material optical feature data is coupled with the optimized geometric structure data to further optimize the material feature data and generate optimized material feature data.

4. The high-altitude curtain wall three-dimensional modeling method according to claim 3 is characterized in that: The extraction of thermal characteristic data comprises the following steps: The thermal gradient analysis is performed on the thermal imaging information in the preprocessed fusion data. The thermal feature model data is generated through the thermal feature modeling algorithm in combination with the optimized geometric structure data and the optimized material optical feature data. The thermal feature model data is then corrected and optimized in real time using a dynamic feedback mechanism to generate optimized thermal feature data.

5. The high-altitude curtain wall three-dimensional modeling method according to claim 1 is characterized in that: The multimodal feature fusion algorithm comprises the following steps: The optimized geometric structure data, the optimized material feature data and the optimized thermal feature data are weightedly fused, where: The weight coefficient W1 is used to indicate the importance of geometric structure data; the weight coefficient W2 is used to indicate the importance of material feature data; the weight coefficient W3 is used to indicate the importance of thermal feature data; The weight coefficients are dynamically adjusted through a recursive optimization mechanism to generate optimized model data.

6. The high-altitude curtain wall three-dimensional modeling method according to claim 5, characterized in that: The specific expression of the weighted fusion calculation is: M final =W1×G+W2×M+W3×T Among them, M final represents the optimized model data; G represents the optimized geometric structure data; M represents the optimized material feature data; T represents the optimized thermal feature data; W1, W2 and W3 represent the weight coefficients of the geometric structure data, material feature data and thermal feature data respectively.

7. The high-altitude curtain wall three-dimensional modeling method according to claim 1 is characterized in that: The global geometric correction comprises the following steps: Based on the optimized model data, the macro-geometric structure of the three-dimensional model is globally optimized, and the macro-geometric errors are detected and eliminated using the global optimization algorithm. The global correction geometric model data is generated by adjusting the spatial position, scale and angle of the overall shape of the model.

8. The high-altitude curtain wall three-dimensional modeling method according to claim 7, characterized in that: The local geometry refinement correction comprises the following steps: The detail areas in the global correction geometry model data are locally optimized, the local geometry errors are corrected by the detail enhancement algorithm, and the key areas are refined by the micro-adjustment algorithm to generate the final geometry correction model data.

9. The high-altitude curtain wall three-dimensional modeling method according to claim 1, characterized in that: The multi-channel fusion algorithm comprises the following steps: The final geometric correction model data, optimized material feature data and optimized thermal feature data are rendered independently respectively. The geometric structure is rendered through the geometric channel, the material features are rendered through the material channel, and the thermal effect is rendered through the thermal feature channel. The rendering results of each channel are subjected to multi-channel fusion processing to generate the final 3D model data.

10. A system for implementing the three-dimensional modeling method of a high-altitude curtain wall according to any one of claims 1 to 9, characterized in that: include: The data acquisition module is used to collect multi-spectral data, structured light scanning data and thermal imaging data of high-altitude curtain walls through drones, and synchronously process the collected data in time and space to generate pre-processed fusion data; The data analysis and feature extraction module is used to perform multi-dimensional cross-analysis on the pre-processed fusion data, extract geometric structure data, material feature data and thermal feature data in turn, and generate fusion feature data through hierarchical feature fusion and dynamic feedback optimization; The 3D modeling and feature fusion module is used to perform 3D modeling based on fused feature data through a multimodal feature fusion algorithm, deeply fuse geometric structure data, material feature data, and thermal feature data, and generate optimized model data using real-time feedback and recursive optimization mechanisms; The correction and optimization module is used to perform global geometry correction, local geometry refinement correction, material feature optimization and thermal feature correction on the optimization model data, and generate the final correction optimization model data through multi-dimensional collaborative optimization; The rendering and output module is used to optimize the model data based on the final correction, perform geometric structure rendering, material feature rendering and thermal feature rendering, and generate the final three-dimensional model data through a multi-channel fusion algorithm.

Images