Three-dimensional relief type virtual modeling device based on digital indoor and outdoor planning

Abstract

The invention relates to the technical field of digital modeling, in particular to a three-dimensional embossment type virtual modeling device based on digital indoor and outdoor planning, which utilizes a virtual unmanned aerial vehicle and a laser scanner to scan the inside and the outside of a building to form an embossment effect outside the building and an embossment effect of a floor, a wall body and a ceiling inside the building. The method comprises the following steps: modeling a whole building by utilizing three-dimensional laser scanning to obtain a three-dimensional effect picture of the building, generating a virtual embossment effect module to receive three-dimensional structure data, carrying out copying by utilizing a three-dimensional modeling technology, importing an embossment pattern, and finally constructing a virtual indoor and outdoor scene and rendering the height of the ground by utilizing layers with different heights through a multi-layer virtual terrain module. A multi-source data fusion acquisition technology improves the precision and integrity of a building model, provides a high-quality data basis for the generation of an embossment effect, and a layered adaptive embossment depth construction technology enables the embossment effect to better conform to the structural characteristics of a building, and enhances the stereoscopic impression and spatial expressive force of the model.

Description

Technical Field

[0001] The present invention relates to the technical field of digital modeling, and particularly to a three-dimensional relief virtual modeling device based on digital indoor and outdoor planning, which is mainly applied to fields such as architectural design, indoor and outdoor planning, virtual display, etc. Background Art

[0002] With the rapid development of digital technology, virtual modeling has been widely applied in the fields of architectural design and indoor and outdoor planning. Traditional three-dimensional modeling techniques mainly focus on the geometric expression of building structures and appearances, but there are obvious deficiencies in terms of detail performance, especially in the expressiveness of architectural relief effects, texture details, etc.

[0003] In the prior art, three-dimensional modeling usually uses a single data source for collection, resulting in insufficient data accuracy and integrity; during the model rendering process, the performance of different height levels is relatively single, lacking a sense of hierarchy; the degree of mechanization in the process of texture application is high, lacking artistry and personalization; at the same time, the prior art often processes the internal and external planning of buildings separately, resulting in poor overall coordination. In addition, the user interaction experience in the prior art is not good, lacking real-time feedback and optimization mechanisms.

[0004] Therefore, there is an urgent need for a virtual modeling device that can integrate internal and external building information, improve model fineness, enhance artistic expressiveness, and have a good interaction experience. Summary of the Invention

[0005] The purpose of the present invention is to provide a three-dimensional relief virtual modeling device based on digital indoor and outdoor planning, which realizes the construction of a fine three-dimensional model integrating the inside and outside of a building and relief-style artistic expression through technical means such as multi-source data collection and fusion, hierarchical adaptive relief depth construction, multi-scale texture rendering, intelligent relief style transfer, and interactive optimization adjustment.

[0006] The present invention proposes a three-dimensional relief virtual modeling device based on digital indoor and outdoor planning, including: A building feature collection module, which is used for: Scanning in a three-dimensional form of the building using a virtual drone to form an external relief effect of the building; Scanning the interior of the building using a laser scanner to form relief effects of the interior floor, walls, and ceiling of the building; Modeling the whole building using three-dimensional laser scanning to obtain a three-dimensional rendering of the building; A virtual relief effect generation module, connected to the building feature collection module, which is used for: Receiving the three-dimensional structure data collected by the building feature collection module; Using three-dimensional modeling technology to replicate the three-dimensional structure data; Render different data layers and import relief patterns; Generate a multi-layer virtual terrain module, connected to the virtual relief effect generation module, for: Based on the data layers rendered by the virtual relief effect generation module; Construct a virtual indoor and outdoor scene with layers of different heights; Render the ground height in the form of height.

[0007] Preferably, the device further includes: A terrain planning and simulation module, connected to the multi-layer virtual terrain generation module, for: Based on the three-dimensional virtual building constructed by the multi-layer virtual terrain generation module; Render the planning models for different height levels inside the building; Perform three-dimensional visualization.

[0008] Preferably, the device further includes: A planning scheme display module, connected to the terrain planning and simulation module, for: Based on the three-dimensional virtual building constructed by the terrain planning and simulation module; Plan a virtual modeling scheme; Combine with virtual simulation images to present the planning scheme when the user enters the planned virtual three-dimensional map; A planning scheme evaluation module, connected to the planning scheme display module, for: Based on the three-dimensional virtual building constructed by the planning scheme display module; According to the building area and the planned virtual three-dimensional map; Use deep learning technology to compare the planning scheme with the building area and give the planning result.

[0009] Preferably, the device further includes: An architectural facade design and modeling module, connected to the building feature collection module, for: Use three-dimensional modeling technology based on the three-dimensional structure data scanned by the building feature collection module; Construct a three-dimensional building drawing according to the architectural facade style; Compare the real facade photos of the same building at different times with the three-dimensional completed architectural facade; Combine with the architectural facade style data to obtain an architectural facade drawing similar to the real facade photo.

[0010] Preferably, the device further includes: An interior wall virtual processing module, connected to the building feature collection module, for: Replicate the floor and wallpaper patterns of the interior floor and ceiling of the building collected by the building feature collection module using 3D modeling technology; Based on the floor and wallpaper pattern data of the interior floor and ceiling of the building; Using deep learning technology, randomly generate floor and wallpaper patterns consistent with the actual ones; Visualize different pattern data.

[0011] Preferably, the device further includes: A building modeling data comparison module, connected to the building feature collection module, for: Using data to simulate building modeling and the scan data collected by the building feature collection module; Compare the building modeling image and the building scan image; Give an early warning prompt according to the change of building height.

[0012] Preferably, the device further includes: A user scan data comparison module, connected to the building modeling data comparison module, for: Use the obstacle model planned by the user's virtual hand or virtual modeling to scan the building modeling image; Give an alarm prompt according to the situation of the obstacle model planned by the user's virtual hand or virtual modeling.

[0013] Preferably, the device further includes: A user modeling behavior recording module, connected to the user scan data comparison module, for: Record the user's modeling behavior to form user behavior data; Store the user's modeling behavior data in the database.

[0014] Preferably, the process of constructing a three-dimensional virtual building modeling in the virtual relief effect generation module includes the following steps: S1: Import the modeling basic data. The building modeling basic data is to layer the real building modeling data according to different time nodes to form a 3D modeling diagram of different height layers; S2: Split and recombine the 3D modeling diagrams of different height layers to form modeling diagrams of different height floors; S3: Use the modeling data to render the split modeling diagrams to obtain modeling renderings of different height floors; S4: Use the modeling renderings of different height floors to perform 3D superposition according to different height layers to form a composite modeling diagram; S5: Use the spatial geometry algorithm to operate on the composite modeling diagram and output the modeling data; S6: Construct a 3D virtual building using the modeling data and perform real-time rendering on the building; S7: Analyze the spatial relationships of the 3D virtual building using the rendered modeling data and dynamically display the virtual building.

[0015] Preferably, the process of constructing the virtual three-dimensional model includes the following steps: Extract the modeling data from the virtual rendered building to obtain model data; Preprocess the model data to obtain a model mapping file; Extract data features from the model mapping file to obtain different modeling data; Perform hierarchical processing on the different modeling data to form modeling diagrams of different floors at different heights; Compare the real exterior facade photos of the same building at different times with the exterior facade of the three-dimensional built building, and combine with the building exterior facade style data to obtain a building exterior facade diagram similar to the real exterior facade photo; Label the building exterior facade diagram with the same style and classify the labels of the same style; Select the one with the highest similarity to the real exterior facade photo from the classified labels of the same style and map it to the corresponding exterior facade area to obtain the building exterior facade mapping result; Analyze the mapping result of the building exterior facade, and combine with the building exterior facade style data to obtain the building exterior facade mapping analysis result; Visualize the building exterior facade mapping analysis result to obtain the building modeling rendering effect diagram.

[0016] The present invention has the following beneficial effects: 1. Through the multi-source data fusion acquisition technology, the accuracy and integrity of the building model are improved, providing a high-quality data basis for the generation of the relief effect; 2. Adopting the hierarchical adaptive relief depth construction technology makes the relief effect more in line with the structural characteristics of the building itself, enhancing the three-dimensional sense and spatial expressiveness of the model; 3. Using the multi-scale texture rendering technology significantly improves the rendering efficiency and quality, making the relief texture present multi-level and multi-texture characteristics; 4. Through the relief style transfer technology assisted by deep learning, the intelligent matching and personalized customization of the relief style are realized, enriching the relief expression forms; 5. Combining with the real-time interactive relief effect adjustment mechanism improves the user participation and operation efficiency, and the system can continuously optimize the relief effect according to the user feedback. Description of the Drawings

[0017] Figure 1System architecture diagram of the three-dimensional relief virtual modeling device based on digital indoor and outdoor planning provided by the embodiments of the present invention; Figure 2 Schematic diagram of the working process of the building feature acquisition module of the present invention; Figure 3 Schematic diagram of the processing flow of the virtual relief effect generation module of the present invention; Figure 4 Schematic diagram of the hierarchical adaptive relief depth construction technology of the present invention; Figure 5 Schematic diagram of the multi-scale texture rendering processing flow of the present invention; Figure 6 Example of the three-dimensional rendering of indoor and outdoor buildings of the present invention; Figure 7 Example diagram of the relief effect of the building facade of the present invention; Figure 8 Schematic diagram of the user interactive adjustment interface of the present invention; Figure 9 Flowchart of the process of constructing a three-dimensional virtual building model of the present invention; Figure 10 Flowchart of the process of modeling a virtual three-dimensional map of the present invention. Detailed implementation manners

[0018] Please refer to the attached Figures 1-10 , and the following further details the specific implementation manners of the present invention in conjunction with the accompanying drawings.

[0019] As Figure 1 shown, the three-dimensional relief virtual modeling device based on digital indoor and outdoor planning proposed by the present invention includes functional modules such as a building feature acquisition module 1, a virtual relief effect generation module 2, a multi-layer virtual terrain generation module 3, a planned terrain simulation module 4, a planned scheme display module 5, a planned scheme evaluation module 6, a building facade design modeling module 7, an interior wall virtual processing module 8, a building modeling data comparison module 9, a user scan data comparison module 10, a user modeling behavior recording module 11, etc.

[0020] In an embodiment of the present invention, as Figures 1 to 3 shown, the building feature acquisition module 1 is used to: scan in a three-dimensional form of the building using a virtual drone to form an external relief effect of the building; scan the interior of the building using a laser scanner to form the relief effects of the interior floor, walls, and ceiling of the building; perform three-dimensional laser scanning on the whole building to obtain a three-dimensional rendering of the building.

[0021] Preferably, the building feature acquisition module 1 includes a three-dimensional laser scanner 101, a virtual drone 102, and a multi-angle camera 103, which are connected to the system central processing unit through a server. The three-dimensional laser scanner 101 is mainly responsible for collecting the internal details of the building, and can accurately capture the textures and geometric features of the walls, floors, and ceilings, with a scanning accuracy reaching the millimeter level. The virtual drone 102 is equipped with high-resolution imaging equipment and conducts a full-range scan of the building exterior on a preset flight path, usually covering an angle of 360 degrees to ensure blind area-free acquisition. The multi-angle camera 103 serves as an auxiliary device for collecting image data of special angles or detailed parts.

[0022] It should be noted that the building feature acquisition module 1 adopts a multi-source collaborative acquisition strategy to ensure that the positions captured by all devices for photographing are corresponding to each other and the marking points are consistent. For example, when the virtual drone 102 captures a certain point on the building facade, the system will record the three-dimensional coordinates of this point and establish a mapping relationship with the corresponding internal point collected by the laser scanner, thereby achieving the precise fusion of internal and external data.

[0023] The virtual relief effect generation module 2 is connected to the building feature acquisition module 1 and is used for: receiving the three-dimensional structure data collected by the building feature acquisition module 1; replicating the three-dimensional structure data using three-dimensional modeling technology; rendering different data layers and importing relief patterns. This module is the core innovative part of the present invention and includes a series of advanced data processing algorithms and rendering technologies.

[0024] In practical applications, the virtual relief effect generation module 2 first converts the color image into a grayscale image and then into three-dimensional point cloud data. The conversion process uses a color point cloud processing algorithm, which can be expressed as: , where: is the converted grayscale value at the coordinate point, with a value range of 0 - 255; is the red component value of the original image at the coordinate point, with a value range of 0 - 255; is the green component value of the original image at the coordinate point, with a value range of 0 - 255; is the blue component value of the original image at the coordinate point, with a value range of 0 - 255; 0.299, 0.587, and 0.114 are the standard weight coefficients for RGB to grayscale conversion, and these weight values take into account the sensitivity differences of the human eye to different colors. is the coordinate point is the coordinate point is the coordinate point is the coordinate point

[0025] This algorithm is particularly effective when dealing with the relief effect of ancient building facades. For example, for a Ming and Qing dynasty building with delicate brick carvings, through the above-mentioned grayscale conversion, the texture details of the brick carvings can be highlighted, making the subsequent generated relief effect clearer. Since the light reflection characteristics of different materials are different, the system will dynamically adjust the weight coefficients according to the type of building material. For example, for a stone building, the weights can be adjusted to 0.33, 0.56, and 0.11 to better retain the subtle changes in the stone texture.

[0026] Next, the virtual relief effect generation module 2 maps the point cloud data into a three-dimensional coordinate system to form standardized three-dimensional model data. This module uses a hierarchical processing technique to split and reorganize the building model according to different time nodes and height levels. In addition, this module is also responsible for implementing the rendering of the data layer and the import of relief patterns. By setting different rendering parameters, a relief effect with artistic expressiveness can be generated.

[0027] The multi-layer virtual terrain generation module 3 is connected to the virtual relief effect generation module 2 and is used for: based on the data layer rendered by the virtual relief effect generation module 2; constructing virtual indoor and outdoor scenes with different height layers; rendering the ground height in the form of height. This module realizes the three-dimensional expression of the terrain through the processing and combination of different height layers.

[0028] In specific implementation, the multi-layer virtual terrain generation module 3 receives the rendering data from the virtual relief effect generation module 2 and constructs a multi-level virtual scene according to the height information. For example, for a typical three-story building, this module can divide it into five height layers: basement, first floor, second floor, third floor, and roof. Each height layer has independent rendering parameters. This module uses the height map technology to express the terrain undulation and enhances the visual effect through color gradients and texture changes. For example, the height value can be mapped to the grayscale value range of 0 - 255, and for every 0.5-meter increase in height, the grayscale value increases by 10. This processing method makes the height difference more obvious visually and is convenient for users to understand the spatial relationship.

[0029] In another embodiment of the present invention, the device further includes a planned terrain simulation module 4, which is connected to the multi-layer virtual terrain generation module 3 and is used for: based on the three-dimensional virtual building constructed by the multi-layer virtual terrain generation module 3; rendering the planned models of different height levels inside the building; performing three-dimensional visualization.

[0030] After receiving the three-dimensional virtual building data, the planned terrain simulation module 4 first divides the building in terms of spatial levels, usually setting section layers at an interval of 0.5 meters in height. Then, appropriate rendering parameters are applied to each section layer, including lighting models, material properties, and texture details, etc. Finally, the rendering result is visualized through a three-dimensional graphics engine, and users can view the rendering effect through virtual reality devices or ordinary displays.

[0031] Preferably, the planned terrain simulation module 4 adopts physically based rendering (PBR) technology, taking into account physical properties such as the reflectivity, roughness, and metallicity of materials, to make the rendering effect more realistic. At the same time, this module supports multi-view observation and real-time roaming, and users can freely switch perspectives to deeply understand the internal structure of the building.

[0032] In another embodiment of the present invention, the device further includes a planned solution display module 5 and a planned solution evaluation module 6. The planned solution display module 5 is connected to the planned terrain simulation module 4 and is used for: constructing a three-dimensional virtual building based on the planned terrain simulation module 4; planning a virtual modeling solution; and presenting the planned solution when the user enters the planned virtual three-dimensional map in combination with the virtual simulation image.

[0033] The planned solution display module 5 provides an immersive planned solution display experience by integrating a variety of interaction technologies. This module can create a complete virtual scene including the building exterior, internal structure, and surrounding environment. After the user enters the virtual scene, they can see the real-time rendering effect of the planned solution, including the building facade, internal space layout, decorative details, etc. In addition, this module also supports adding dynamic elements, such as people flow, vehicles, light changes, etc., to enhance the realism and vividness of the scene.

[0034] The planned solution evaluation module 6 is connected to the planned solution display module 5 and is used for: constructing a three-dimensional virtual building based on the planned solution display module 5; according to the building area and the planned virtual three-dimensional map; using deep learning technology to compare the planned solution with the building area and give a planning result.

[0035] In a specific implementation, the planned solution evaluation module 6 uses a deep learning model based on a convolutional neural network (CNN) to evaluate the planned solution. The model structure includes an input layer, multiple convolutional layers, pooling layers, fully connected layers, and an output layer. The model input is the three-dimensional model data of the planned solution and the building area parameters, and the output is a planning score and optimization suggestions.

[0036] The evaluation process of the deep learning model can be expressed as: , where: Score is the evaluation score, with a value range of 0 - 100, representing the comprehensive quality score of the planned solution; is a score mapping function that maps the feature vector output by the CNN to the score range of 0 - 100; is a convolutional neural network model used to extract the spatial features, material features, and structural features of the planned solution; is the three-dimensional model data, including geometric information, material information, and spatial layout information; is the building area parameter, with the unit of square meters, which describes the overall area of the building and the area distribution of each functional area.

[0037] To make the evaluation results more interpretable, the evaluation criteria include multiple dimensions, and the specific scoring formula can be expressed as: , where: Score is the total score, and the value range is 0 - 100; is the space score, which evaluates the rationality of the space layout; is the aesthetics score, which evaluates the aesthetic value of the design; is the function score, which evaluates the practicality of the design; is the economic score, which evaluates the cost - benefit; , , , are the weight coefficients of each dimension, and satisfy , usually set as .

[0038] In practical applications, for example, when evaluating the planning scheme of a modern office building, the system may score based on the following specific indicators: space utilization rate (the ratio of the working area to the total area), rationality of the human flow line (the shortest average distance of the path), daylighting and ventilation effect (the natural light coverage rate), aesthetic value (the facade coordination index), etc. When the evaluation score is lower than the threshold (usually set at 70 points), the system will give specific improvement suggestions. For example, when the space utilization rate score is lower than 65 points, the system will suggest adjusting the internal layout and reducing the corridor area; when the daylighting score is lower than 60 points, the system will suggest increasing the window area or optimizing the window position.

[0039] In another embodiment of the present invention, the device further includes a building facade design modeling module 7, which is connected to the building feature acquisition module 1 and is used for: using three - dimensional modeling technology, based on the three - dimensional structure data scanned by the building feature acquisition module 1; constructing a three - dimensional building drawing according to the building facade style; comparing the real facade photos of the same building at different times with the three - dimensional built building facade; combining the building facade style data to obtain a building facade drawing similar to the real facade photo.

[0040] The building facade design modeling module 7 realizes the intelligent matching and style transfer functions of photos and models. This module first establishes a building facade style database, which contains building facade photos of different periods and different styles and their corresponding style tags. Then, through feature extraction algorithms, it analyzes the geometric features, material textures, and decorative elements of the building facade, and compares these features with the samples in the database.

[0041] The style matching process can be expressed as: , where: is the best matching style, representing the style category in the style library that best matches the current building; is the style library set, containing multiple predefined architectural styles (such as classical, modern, postmodern, industrial style, etc.); is the parameter that makes the following expression take the maximum value, here referring to selecting the style with the highest similarity; is the similarity calculation function, used to measure the similarity between two feature vectors; is the feature extraction function, which converts the image into a feature vector; is the input building facade image; is the sample in the style library.

[0042] When processing a historical building that combines multiple style elements, the system extracts features and performs matching in regions. For example, for a building with a Roman style on the ground floor and a Gothic style on the upper floor, the system extracts the features of the two parts respectively, matches them with the Roman style and Gothic style samples in the style library, and then generates a facade map of the combined style. The cosine similarity method is used for similarity calculation, defined as: , where: and are feature vectors; is the inner product of the vectors, calculated as ; and are the Euclidean norms of the vectors and respectively, calculated as and ; is the dimension of the feature vector, usually set to 256 or 512 in this system.

[0043] After finding the best matching style, the building facade design modeling module 7 applies the style features to the 3D model to generate a building facade map with a specific style. This process takes into account the structural features of the building to ensure that the style application does not affect the basic form of the building.

[0044] In another embodiment of the present invention, the device further includes an interior wall virtual processing module 8, connected to the building feature collection module 1, for: using 3D modeling technology to reproduce the floor and wallpaper patterns of the building interior floor and ceiling collected by the building feature collection module 1; based on the floor and wallpaper pattern data of the building interior floor and ceiling; using deep learning technology to randomly generate floor and wallpaper patterns consistent with the actual ones; visualizing different pattern data.

[0045] The interior wall virtual processing module 8 uses generative adversarial network (GAN) technology to achieve intelligent generation of texture patterns. This module contains two core components: a generator and a discriminator. The generator is responsible for generating texture patterns according to input conditions, and the discriminator is responsible for evaluating the similarity between the generated patterns and real patterns.

[0046] The training process of the generative adversarial network can be expressed as a min-max game problem: Where: represents minimizing and optimizing the parameters of the generator ; represents maximizing and optimizing the parameters of the discriminator ; is the objective function, representing the game objective between the generator and the discriminator; is the expectation operation, calculating the average value under the probability distribution; is the real data distribution, representing the distribution of real wallpaper or floor pattern samples; is the random noise distribution, usually using the standard normal distribution is the natural logarithm function; is the discrimination result of the discriminator for the input indicating is the probability of being a real sample, with a value range of is the sample generated by the generator according to the noise ; is the discrimination result of the discriminator for the generated sample.

[0047] In practical applications, when the system needs to generate wallpaper patterns that conform to the style of a specific period for the interior of a historical building, the generator will generate samples according to the style characteristics of that period, and the discriminator will evaluate the similarity between the samples and real historical wallpapers. For example, when generating wallpapers for the interior of a Victorian-era building, the system will learn the typical floral patterns and color characteristics of that period and then generate new patterns that conform to these characteristics.

[0048] The generator adopts the UNet architecture, which can retain the spatial information of the input image; the discriminator adopts the PatchGAN architecture, which focuses on the local consistency of textures. The training data includes a large number of real floor and wallpaper pattern samples, which are preprocessed and then input into the network. After training, the interior wall virtual processing module 8 can generate seamless texture patterns according to style descriptions or reference images and apply them to the virtual representation of the building interior.

[0049] In another embodiment of the present invention, the device further includes a building modeling data comparison module 9, which is connected to the building feature acquisition module 1 and is used for: simulating the building modeling and the scanning data acquired by the building feature acquisition module 1; comparing the building modeling image and the building scanning image; and giving a warning prompt according to the change of the building height.

[0050] The building modeling data comparison module 9 realizes the comparison and anomaly detection functions between the model and the actual scanning data. This module first unifies the simulated modeling data and the actual scanning data into the same coordinate system, and then achieves precise alignment through a registration algorithm. The registration process uses the Iterative Closest Point (ICP) algorithm, which can be expressed as: , Where: is the error function, representing the mean squared distance between two sets of point cloud data; is the rotation matrix, with a dimension of , describing the rotational transformation of the point cloud; is the translation vector, with a dimension of , describing the translational transformation of the point cloud; is the number of point pairs, representing the total number of point pairs participating in the calculation; represents the summation over all point pairs; is the -th point in the point cloud to be registered, which is a three-dimensional coordinate point; is the closest point in the target point cloud corresponding to , which is also a three-dimensional coordinate point, represents the square of the Euclidean distance between point and the transformed point .

[0051] The ICP algorithm iteratively optimizes the rotation matrix [[ID=4(0]] and the translation vector to minimize the error function . The iterative process of the algorithm includes the following steps: 1. For each point in the point cloud to be registered, find the closest point in the target point cloud; 2. Calculate the optimal rotation matrix and translation vector according to the current point pair; 3. Apply the transformation to update the position of the point cloud to be registered; 4. If the change of the error function is less than the threshold (usually set to 0.0001) or reaches the maximum number of iterations (usually set to 50 times), the algorithm terminates; otherwise, return to step 1.

[0052] In the ancient building protection project, this module can be used to monitor the deformation of the building. For example, for a century-old wooden structure building, the system will regularly scan the building structure and compare it with the historical model to detect problems such as structural settlement, tilt or cracking. After the registration is completed, the building modeling data comparison module 9 calculates the difference between the two sets of data and generates a difference heat map. When it is found that the building height change exceeds the preset threshold (usually set to 10 cm), the system will issue a warning prompt. The warning level is divided into three levels according to the degree of difference: minor (10 - 20 cm), moderate (20 - 30 cm) and severe (more than 30 cm), which are represented by yellow, orange and red in the heat map respectively.

[0053] In another embodiment of the present invention, the device further includes a user scan data comparison module 10, which is connected to the building modeling data comparison module 9 and is used for: scanning the building modeling image with the obstacle model of the user's virtual hand or virtual modeling plan; giving an alarm prompt according to the situation of the obstacle model of the user's virtual hand or virtual modeling plan.

[0054] The user scan data comparison module 10 enhances the interactivity and security of the system. This module captures the user's hand movements through a virtual reality device or receives the obstacle model created by the user, and performs real-time collision detection with the building model. The collision detection uses the bounding volume hierarchy (BVH) algorithm, which has high performance and accurate results.

[0055] The system will trigger an alarm when the following situations are detected: 1) the user's virtual hand penetrates the building entity structure; 2) the obstacle model interferes with the key structure of the building; 3) there is an unreasonable spatial layout in the planning scheme. The alarm prompt includes a sound warning, visual highlighting and text description to help the user quickly locate the problem area.

[0056] Preferably, the user scan data comparison module 10 also provides problem-solving suggestions, such as adjusting the position of the obstacle, modifying the planning scheme or optimizing the building structure, etc. These suggestions are based on the preset design specifications and best practice experiences, which help the user to improve the design scheme.

[0057] In another embodiment of the present invention, the device further includes a user modeling behavior recording module 11, which is connected to the user scan data comparison module 10 and is used for: recording the user's modeling behavior to form user behavior data; storing the user's modeling behavior data in the database.

[0058] The user modeling behavior recording module 11 records and analyzes the user's operations throughout the process, providing data support for system optimization and personalized recommendation. The content recorded by this module includes: 1) the user's perspective movement trajectory; 2) the interaction operation sequence; 3) the parameter adjustment history; 4) the planning modification record, etc. These data are stored in the behavior database after preprocessing to form a user operation portrait.

[0059] The processing of behavior data uses sequence pattern mining algorithms to identify common patterns and preference features in user operations. For example, the system can discover that users tend to adjust the exterior facade first and then design the internal space, or prefer a specific relief style. Based on these findings, the system can optimize the interface layout, adjust default parameters, or provide personalized operation suggestions to enhance the user experience.

[0060] In an important embodiment of the present invention, the process of constructing a three-dimensional virtual building model in the virtual relief effect generation module 2 includes the following steps: S1: Import the basic modeling data. The basic building modeling data is to perform hierarchical modeling on the real building modeling data according to different time nodes to form a 3D modeling diagram of different height layers; In this step, the system first imports the original data obtained from the building feature collection module 1. To reflect the change of the time dimension of the building, the system performs hierarchical processing on the data according to time nodes. For example, for a building with a history of a hundred years, the building data of the 1920s, 1950s, 1980s, and modern times can be imported respectively to form a time series. At the same time, the system divides the building into multiple layers according to height, usually with an interval of 3 meters or a standard floor height. This spatio-temporal two-dimensional hierarchical method lays the foundation for subsequent relief effect processing.

[0061] S2: Split and reorganize the 3D modeling diagrams of different height layers to form modeling diagrams of different height floors; This step realizes the refined processing of the model. The system first splits the 3D modeling diagram formed in S1 according to functional areas. For example, a residential building is split into areas such as living rooms, bedrooms, kitchens, and bathrooms. Then, the system reorganizes the split components according to the building structure and use to form a complete floor model. This split and reorganization method enables the system to apply different processing parameters to different areas, improving the fineness and realism of the model.

[0062] S3: Use the modeling data to perform rendering processing on the split modeling diagrams to obtain modeling renderings of different height floors; In the rendering process stage, the system assigns appropriate materials, textures, and lighting parameters to models in different areas. The rendering algorithm adopts physically based rendering (PBR) technology, considering the physical properties of materials, such as reflectivity, roughness, and metallicity, etc. The lighting model uses global illumination technology, supporting the calculation of direct and indirect lighting, making the rendering results more realistic. In addition, the system also applies effects such as ambient occlusion (AO), reflection, and shadow to enhance the depth and three-dimensional sense of the image.

[0063] S4: Use the modeling and rendering diagrams of different floors at different heights to perform a three-dimensional (3D) overlay according to different height layers to form a composite modeling diagram; This step combines the rendering diagrams of each floor obtained from the previous processing in a three-dimensional space. The system first determines the positions of each floor in the global coordinate system and then performs the overlay according to the actual structure of the building. During the overlay process, the system will handle the connection relationships between floors, such as stairways, elevator shafts, pipe shafts, etc. To ensure the consistency of the model, the system adopts a series of constraint conditions, such as wall alignment, opening matching, etc., to ensure the structural rationality of the overlaid model.

[0064] S5: Use a spatial geometry algorithm to operate on the composite modeling diagram and output modeling data; The spatial geometry algorithm is one of the core technologies of the present invention and is used to process complex three-dimensional model data. The algorithm includes multiple components such as mesh simplification, surface subdivision, and geometric feature extraction. Among them, the mesh simplification algorithm is used to reduce the model complexity and improve the rendering efficiency; the surface subdivision algorithm is used to enhance the model details and improve the visual quality; the geometric feature extraction algorithm is used to identify features such as edges, corners, and planes in the model, providing a basis for generating the relief effect.

[0065] The core algorithm of the spatial geometry processing can be expressed as: , where is the processed model, which is a three-dimensional mesh model containing vertex coordinates, face indices, and normal information, etc.; is the input model, which is also a three-dimensional mesh model; is the set of processing parameters, containing multiple parameters such as simplification rate, subdivision level, and feature threshold; SpatialProcess is the spatial processing function, containing a composite function of a series of geometric operations.

[0066] In the processing parameter the simplification rate is usually set to 0.5 - 0.8, indicating the proportion of the original mesh face number to be retained; the subdivision level is usually set to 1 - 3, indicating the number of additional subdivisions on the basis of the original model; the feature threshold is usually set to 0.01 - 0.05, indicating the sensitivity to identify geometric features, and the smaller the value, the more features are identified.

[0067] For the relief processing of historical buildings, the system will automatically adjust parameters according to the architectural style. For example, for Baroque-style buildings, the system will set a lower simplification rate (0.8) and a higher subdivision level (3) to retain rich decorative details; while for modernist buildings, a higher simplification rate (0.5) and a lower subdivision level (1) may be set to highlight the simple geometric forms.

[0068] S6: Construct a 3D virtual building using the modeling data and perform real-time rendering on the building; In this step, the system loads the modeling data obtained from the previous processing into a 3D graphics engine to construct a complete virtual building model. Real-time rendering is implemented using modern graphics APIs (such as OpenGL, DirectX, or Vulkan), supporting advanced rendering techniques such as deferred rendering, physically based material systems, and dynamic lighting. To balance visual quality and performance, the system uses the level of detail (LOD) technique to automatically adjust the model complexity according to the viewing distance.

[0069] S7: Analyze the spatial relationships of the 3D virtual building using the rendered modeling data and perform dynamic display of the virtual building.

[0070] The last step is spatial relationship analysis and dynamic display. The system first analyzes the spatial connection relationships inside the building, such as the passages between rooms, the positions of doors and windows, etc., to generate a spatial connectivity map. Then, based on this analysis result, the system supports multiple dynamic display methods, such as sectional display, exploded view, perspective rendering, etc. Users can select different display methods through the interactive interface to deeply understand the internal structure and spatial organization of the building.

[0071] In another important embodiment of the present invention, the process of constructing a virtual 3D model includes the following steps: Extract the modeling data from the virtual rendered building to obtain model data; The system first exports the geometric data and material data of the building model from the rendering engine. The geometric data includes vertex coordinates, patch indices, normal vectors, etc.; the material data includes colors, texture coordinates, material parameters, etc. The data extraction uses standard formats (such as FBX, COLLADA, or glTF) to ensure the integrity and compatibility of the data.

[0072] Perform data preprocessing on the model data to obtain a model mapping file; In the data preprocessing stage, the system cleans, repairs, and optimizes the extracted model data. During the cleaning process, redundant vertices, duplicate faces, and invalid materials are removed; during the repair process, common problems such as non-manifold geometry, flipped normals, and UV breaks are addressed; during the optimization process, adjacent vertices are merged, complex meshes are simplified, and textures are compressed, etc. After the processing is completed, the system generates a model mapping file, recording the corresponding relationships and processing parameters of each part of the model.

[0073] Data feature extraction is performed on the model mapping file to obtain different modeling data; Feature extraction is the process of converting the model mapping file into a high-level semantic representation. The system uses geometric feature extraction algorithms to identify features such as planes, surfaces, edges, and corners in the model, and calculates the attributes of these features, such as area, curvature, and direction, etc. At the same time, the system also extracts material features, such as color distribution, texture pattern, and reflection characteristics, etc. These features form a multi-dimensional feature vector for subsequent hierarchical processing.

[0074] The different modeling data are hierarchically processed to form modeling diagrams of different floor heights; The hierarchical processing is based on the features extracted previously, dividing the model into a meaningful hierarchical structure. The system first divides the model into different floors according to height information, and then further divides each floor according to functional areas. The division process uses the region growing algorithm, starting from the seed region and gradually expanding to adjacent regions until the entire model is covered.

[0075] Compare the real facade photos of the same building at different times with the three-dimensional built building facade, and combine the building facade style data to obtain a building facade diagram similar to the real facade photo; This step realizes the matching and style transfer between the photo and the model. The system first aligns the photo with the model through the feature point matching algorithm, and then extracts the texture and style features in the photo and applies them to the model surface. The style transfer uses a deep learning-based method, retaining the structural features of the building while transferring the style features of the photo.

[0076] Label the building facade diagram with the same style and classify the labels of the same style; The style label system is an important part of building style management. The system pre-defines a series of style categories, such as classical, modern, postmodern, industrial style, etc., and each category contains multiple sub-categories. The label assignment is based on the previous style analysis results, using a multi-label classification method, allowing a building to have multiple style features. The label information is stored in the meta-database for subsequent query and filtering.

[0077] Select the one with the highest similarity to the real facade photo from the classified labels of the same style and map it to the corresponding facade area to obtain the building facade mapping result; Similarity calculation uses a multi-feature fusion method, taking into account geometric similarity, texture similarity, and style similarity. The calculation formula is: , Among them: Similarity is the total similarity, the value range is [0,1], the larger the value, the higher the similarity; Geometric similarity measures the similarity between two samples in shape and structure; Texture similarity measures the similarity between two samples in texture and details; Style similarity measures the similarity between two samples in terms of style and aesthetic features; 、 and is the corresponding weight coefficient, and satisfies , usually set to ,These weights can be adjusted according to the specific application scenarios.

[0078] For example, for a building that blends traditional Chinese and modern Western elements, the system calculates geometric, texture, and style similarities with samples of different styles, then performs a weighted sum to arrive at an overall similarity. For the traditional Chinese elements, the system might choose to match them with traditional palace or garden architecture; for the modern elements, it might choose to match them with modernist architecture. The system selects the sample with the highest similarity for mapping, ensuring the generated result is consistent with the real photo.

[0079] Analyze the building facade mapping results and obtain the building facade mapping analysis results by combining the building facade style data; Mapping result analysis assesses mapping quality and consistency. The system checks the mapped area for boundary continuity, natural texture transitions, and overall stylistic harmony. If problems are found, the system flags the problem areas and provides remediation suggestions. For example, if texture discontinuities occur at boundaries, the system recommends applying a texture fusion algorithm; if local stylistic inconsistencies are detected, the system recommends adjusting the stylistic parameters for that area.

[0080] Visualize the building facade mapping analysis results to obtain the building modeling rendering effect diagram.

[0081] The last step is to visualize the analysis results and generate the final rendering effect diagram. The system supports multiple visualization methods, such as realistic rendering, stylized rendering, heat map rendering, etc. Realistic rendering pursues photo-realism; stylized rendering emphasizes artistic expressiveness; heat map rendering is used to highlight the distribution of specific attributes, such as similarity, consistency, etc. Users can select different visualization methods according to their needs to obtain the required visual information.

[0082] From the above detailed description, it can be seen that the three-dimensional relief virtual modeling device based on digital indoor and outdoor planning provided by the present invention realizes the construction of a fine three-dimensional model integrating the inside and outside of the building and the relief art performance through innovative technologies such as multi-source data acquisition and fusion, hierarchical adaptive relief depth construction, multi-scale texture rendering, intelligent relief style transfer, and interactive optimization and adjustment, providing new technical means and forms of expression for the field of architectural design and planning.

[0083] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A three-dimensional relief virtual modeling device based on digital indoor and outdoor planning, characterized in that, Including: A building feature acquisition module, used for: Using a virtual drone to scan in a three-dimensional form of the building to form an external relief effect of the building; Using a laser scanner to scan the interior of the building to form relief effects of the interior floor, walls, and ceiling of the building; Using three-dimensional laser scanning to model the entire building to obtain a three-dimensional rendering of the building; A virtual relief effect generation module, connected to the building feature acquisition module, used for: Receiving the three-dimensional structure data collected by the building feature acquisition module; Using three-dimensional modeling technology to replicate the three-dimensional structure data; Rendering different data layers and importing relief patterns; A multi-layer virtual terrain generation module, connected to the virtual relief effect generation module, used for: Based on the data layers rendered by the virtual relief effect generation module; Constructing virtual indoor and outdoor scenes with different height layers; Rendering the height of the ground in the form of height.

2. The three-dimensional relief virtual modeling device based on digital indoor and outdoor planning according to claim 1, characterized in that The device further includes: A planned terrain simulation module, connected to the multi-layer virtual terrain generation module, used for: Based on the three-dimensional virtual building constructed by the multi-layer virtual terrain generation module; Rendering the planned model for different height levels inside the building; Performing three-dimensional visualization.

3. The three-dimensional relief virtual modeling device based on digital indoor and outdoor planning according to claim 2 is characterized in that: The device further includes: A planned solution display module, connected to the planned terrain simulation module, used for: Based on the three-dimensional virtual building constructed by the planned terrain simulation module; Planning a virtual modeling solution; Combined with virtual simulation images, when the user enters the planned virtual three-dimensional map, presenting the planned solution; A planned solution evaluation module, connected to the planned solution display module, used for: Based on the three-dimensional virtual building constructed by the planned solution display module; According to the building area and the planned virtual three-dimensional map; Using deep learning technology to compare the planned solution with the building area and giving a planning result.

4. The three-dimensional relief virtual modeling device based on digital indoor and outdoor planning according to claim 3, characterized in that, The device further includes: An exterior building design modeling module, connected to the building feature acquisition module, used for: Using three-dimensional modeling technology, based on the three-dimensional structure data scanned by the building feature acquisition module; Constructing a three-dimensional building map according to the exterior building style; Comparing the real exterior building photos of the same building at different times with the three-dimensional constructed exterior building; Combined with the exterior building style data, obtaining an exterior building map similar to the real exterior building photo.

5. The three-dimensional relief virtual modeling device based on digital indoor and outdoor planning according to claim 4, characterized in that, The device further includes: An interior wall virtual processing module, connected to the building feature acquisition module, used for: Using three-dimensional modeling technology to replicate the floor and wallpaper patterns of the interior floor and ceiling collected by the building feature acquisition module; Based on the floor and wallpaper pattern data of the interior floor and ceiling of the building; Using deep learning technology to randomly generate floor and wallpaper patterns consistent with the actual ones; Visualizing different pattern data.

6. The three-dimensional relief virtual modeling device based on digital indoor and outdoor planning according to claim 5, characterized in that The device further includes: A building modeling data comparison module, connected to the building feature acquisition module, used for: Using data to simulate building modeling and the scanning data collected by the building feature acquisition module; Comparing the building modeling image with the building scanning image; Giving an early warning prompt according to the change situation of the building height.

7. The three-dimensional relief virtual modeling device based on digital indoor and outdoor planning according to claim 6, characterized in that, The device further includes: A user scan data comparison module, connected to the building modeling data comparison module, used for: Scan the building modeling image using the obstacle model planned by the user's virtual hand or virtual modeling. Give an alarm prompt according to the situation of the obstacle model planned by the user's virtual hand or virtual modeling.

8. The three-dimensional relief virtual modeling device based on digital indoor and outdoor planning according to claim 7, characterized in that, The device further includes: A user modeling behavior recording module, connected to the user scan data comparison module, for: Recording the user modeling behavior to form user behavior data; Storing the user modeling behavior data in a database.

9. The three-dimensional relief virtual modeling device based on digital indoor and outdoor planning according to claim 8, characterized in that, The process of constructing a three-dimensional virtual building modeling in the virtual relief effect generation module includes the following steps: S1: Import the basic modeling data. The basic building modeling data is a 3D modeling diagram of different height layers formed by hierarchical modeling of the real building modeling data according to different time nodes. S2: Split and reorganize the 3D modeling diagrams of different height layers to form modeling diagrams of different height floors. S3: Use the modeling data to perform rendering processing on the split modeling diagrams to obtain modeling rendering diagrams of different height floors. S4: Use the modeling rendering diagrams of different height floors to perform 3D stacking according to different height layers to form a composite modeling diagram. S5: Use a spatial geometry algorithm to perform operations on the composite modeling diagram and output modeling data. S6: Use the modeling data to construct a 3D virtual building and perform real-time rendering on the building. S7: Use the rendered modeling data to perform spatial relationship analysis on the 3D virtual building and dynamically display the virtual building.

10. The three-dimensional relief virtual modeling device based on digital indoor and outdoor planning according to claim 9, wherein The process of constructing a virtual three-dimensional diagram modeling includes the following steps: Extract the modeling data from the virtual rendered building to obtain model data. Perform data preprocessing on the model data to obtain a model mapping file. Extract data features from the model mapping file to obtain different modeling data. Perform hierarchical processing on the different modeling data to form modeling diagrams of different height floors. Compare the real facade photos of the same building at different times with the 3D built building facade, and combine the building facade style data to obtain a building facade diagram similar to the real facade photo. Label the building facade diagram with the same style and classify the labels of the same style. Select the label with the highest similarity to the real facade photo from the classified labels of the same style and map it to the corresponding facade area to obtain the building facade mapping result. Perform mapping result analysis on the building facade mapping result and combine the building facade style data to obtain the building facade mapping analysis result. Visualize the building facade mapping analysis result to obtain the building modeling rendering effect diagram.

Images