A method for automatically constructing a three-dimensional model of a building using building vector surfaces

Abstract

This invention relates to the field of 3D geographic information data production technology, specifically to a method for automatically constructing 3D building models using building vector surfaces. The method includes: preprocessing the building base surface in a topographic map to extract and form standardized building base vector surface files; automatically extracting the top and bottom elevations of each building base vector surface based on point cloud data; manually checking and correcting elevation anomalies through a 3D visualization environment; automatically stretching and merging the 2D base into a 3D white model based on the elevation information; and finally, performing intelligent texture matching and mapping on the model surface. This method effectively solves the problems of low efficiency, insufficient standardization, and difficulty in quality control associated with traditional manual modeling methods by combining fully automated processing with controlled manual intervention in key stages.

Description

Technical Field

[0001] This invention relates to the field of three-dimensional geographic information data production technology, specifically to a method for automatically constructing three-dimensional models of buildings using building vector surfaces. Background Technology

[0002] Urban 3D models are the core data foundation for constructing digital twin cities and a realistic 3D China, playing an irreplaceable role in numerous fields such as land spatial planning, intelligent traffic management, ecological environmental protection, and public safety emergency response. As a major component of urban 3D models, architectural 3D models are divided into multiple Levels of Detail (LODs) based on their geometric details, texture realism, and application scenarios. LOD 1.3 models require accurate planar outlines, flat roof structures, and typified exterior textures. While they do not pursue the detailed components like window frames and balconies of LOD 3 models, they emphasize visual rationality and mass production capabilities within a large-scale macro-scene, making them an important data type required for spatial statistical analysis and planning decisions at the national and provincial levels.

[0003] Currently, the mainstream method for producing LOD1.3 level 3D building models still heavily relies on manual operation. Production personnel need to manually outline the 2D base contours of each building in professional 3D modeling software, using high-precision oblique photogrammetry 3D models, LiDAR point clouds, or stereo aerial imagery as references. They then manually measure and input the building height based on experience or from reference data, construct the geometry by extruding the contours, and finally manually find and paste textures. The entire process is cumbersome, time-consuming, and costly. More importantly, the subjectivity of manual operation leads to differences in geometric accuracy, height consistency, and texture style between models produced by different operators, and even by the same operator at different times. This makes it difficult to meet the requirements of standardized and large-scale construction, and has become a bottleneck restricting the efficiency and quality of Real-Scene 3D China construction.

[0004] Therefore, the industry urgently needs a technical solution that can automate and streamline the production of LOD1.3 level 3D building models. This method should fully utilize existing and readily available standard topographic map data (providing accurate 2D base contours) and point cloud data (providing reliable 3D elevation information), automatically completing the entire process from contour extraction, elevation calculation, model construction to texture mapping through programmed algorithms. Only a small amount of necessary manual intervention should be introduced at key quality control points, thereby achieving a leap in production efficiency, a significant reduction in production costs, and a fundamental guarantee of model quality and standardization. Summary of the Invention

[0005] The purpose of this invention is to provide a method for automatically constructing three-dimensional building models using architectural vector surfaces, thereby solving the problems of low production efficiency, insufficient model standardization, and excessive reliance on manual experience for quality control in existing technologies.

[0006] To achieve the above objectives, a method for automatically constructing a 3D model of a building using architectural vector surfaces is provided, comprising the following steps:

[0007] S1. Preprocessing of building foundation surface: Extract the specified layer containing residential features from the topographic map data, and filter out the planar geometric shapes of built houses, houses in buildings, toilets, stilt houses, corridors, sheds, pavilions and colonnades based on the feature name attributes of the residential features, form the building foundation vector surface and save it as a building foundation vector surface file;

[0008] S2. Obtain the top elevation of the building: Read the building base vector surface file and corresponding point cloud data, extract the point cloud data within the range of each building base vector surface in the building base vector surface file, and count and use the most frequent elevation value as the top elevation H. T Write it into the building base vector surface file;

[0009] S3. Obtain the bottom elevation of the building: Extract the point cloud data within the outer buffer of the vector plane of each building base, and calculate and use the minimum elevation value as the bottom elevation H. B Write it into the building base vector surface file;

[0010] S4. Manual interactive inspection of building elevation: The point cloud data is constructed into a 3D surface model for display, and the building base vector surface is determined according to H. T and H B The extrusion is displayed as a three-dimensional overlay of patches; its elevation range dH is indicated. T or dH B Building base vector surfaces exceeding a preset threshold are manually checked for their 3D patch fit with the 3D surface model, and non-fitting building base vector surfaces are manually corrected. T or H B value;

[0011] S5. Construct a 3D white model: Based on H T With H B The difference will stretch the building base vector plane into a three-dimensional block, and adjust the bottom of the three-dimensional block to H. B Elevation is used to merge different 3D blocks belonging to the same building into a single 3D white model of the building by performing Boolean operations.

[0012] S6. Perform model texture mapping: Divide the 3D white model of the building into multiple planes and classify them. The planes include the elevation, top, and bottom. Calculate the number of rows and columns of texture maps required for each plane. Then, automatically match the required type of texture map according to the attributes of the 3D white model of the building. Perform texture mapping according to the number of rows and columns and the type of texture map, and reassemble to generate a textured 3D model.

[0013] Furthermore, in step S1, the specified layer is the residential JMD layer defined in the topographic map specification, and is filtered based on the attribute value whose attribute field name is the feature name.

[0014] Furthermore, in step S2, all point clouds that fall completely within the boundary of the building base vector plane are extracted, and the maximum and minimum elevation difference of the point clouds within this range is recorded simultaneously as the top elevation variation range dH. T Write the building base vector surface file; in step S3, set the buffer distance according to the point cloud density and terrain, and simultaneously record the difference between the maximum and minimum elevations of the point cloud within the buffer as the bottom elevation variation range dH. B Write to the building base vector surface file.

[0015] Furthermore, in step S4, the three-dimensional surface model is in OSGB format, and the preset threshold is set based on point cloud accuracy and regional building experience.

[0016] Furthermore, in step S5, the Boolean operation is a three-dimensional Boolean union operation, which merges the three-dimensional blocks and translates the bottom surface of the merged three-dimensional block to its H. B The horizontal plane represented by elevation.

[0017] Furthermore, in step S6, when classifying planes according to the direction of the normal, planes with vertically upward normals are classified as the top surface of the building, planes with vertically downward normals are classified as the bottom surface of the building, and the remaining planes are classified as the facades of the building.

[0018] Furthermore, in step S6, when automatically matching texture maps, the texture map that matches the attribute recorded in the building base vector surface file is called from the preset texture library.

[0019] Principles and advantages:

[0020] 1. This solution automates spatial analysis and statistical calculations of building base vector surfaces and point cloud data from topographic maps, enabling the batch and streamlined construction of 3D models from 2D bases. The method first utilizes the inherent layers and attribute information of the topographic map to automatically filter and extract building base outlines that conform to the specifications, replacing the traditional manual method of drawing each outline individually, significantly improving the efficiency and accuracy of initial data preparation.

[0021] 2. This solution performs elevation statistical analysis on the point cloud within the building's base area and its surrounding buffer zone. This method can automatically and objectively calculate the top and bottom elevations of each building. This process not only significantly reduces the workload and subjective errors of manual interpretation, but also introduces statistical results (such as the elevation variation range dH) to automatically and objectively calculate the top and bottom elevations of each building. T / dH B The system's anomaly detection mechanism intelligently identifies complex situations that may require manual review. This hybrid approach, which prioritizes automated processing and supplements it with targeted manual intervention, effectively optimizes the quality control process and reduces the costs and uncertainties associated with manual inspection while ensuring the overall accuracy of the model.

[0022] 3. In the model building phase, this solution automatically stretches the 2D base into a 3D block based on a uniformly calculated height value. It then intelligently merges spatially connected blocks using 3D Boolean operations, ensuring the geometric integrity and topological correctness of the final 3D white model. The model generated by this method has a standardized structure and consistent data standards, meeting the high consistency requirements of large-scale modeling projects.

[0023] 4. In the texture mapping stage, this solution automatically classifies the top, facade, and bottom surfaces by analyzing the geometric features of the model surface. It then intelligently matches texture library resources based on the inherent attribute information of the building foundation, achieving automated and standardized mapping of exterior textures. This completely changes the traditional inefficient workflow of manually selecting, aligning, and pasting textures, improving mapping efficiency while ensuring the professionalism and consistency of the model's visual effects.

[0024] 5. By constructing a fully integrated, automated processing flow and setting controlled human intervention points in key stages, this invention successfully improves the production efficiency of 3D building models to meet the requirements of large-scale city-level modeling while ensuring the geometric accuracy and standardization of the models. It provides an efficient, reliable, and quality-controllable technical solution for various standardized 3D modeling projects. Attached Figure Description

[0025] Figure 1 This is a logic block diagram of a method for automatically constructing a three-dimensional model of a building using architectural vector surfaces, according to an embodiment of the present invention. Detailed Implementation

[0026] The following detailed description illustrates the specific implementation method:

[0027] Example

[0028] A method for automatically constructing 3D building models using architectural vector surfaces, basically as follows: Figure 1 As shown, it includes the following steps:

[0029] S1. Building Base Surface Preprocessing: Extract the designated layer containing residential features from the standard topographic map data. Based on the feature name attributes recorded by the geometric figures in the designated layer, filter out specific types of planar geometric figures that conform to the definition of building base. The planar geometric figures include completed buildings, buildings under buildings, toilets, stilt houses, corridors, sheds, pavilions, and colonnades. All the selected planar geometric figures are formed into building base vector surfaces and saved as building base vector surface files. Finally, output as an independent building base vector surface file. In step S1, the designated layer is the residential JMD layer defined in the topographic map specification, and the filtering is based on the attribute values ​​whose attribute field name is the feature name.

[0030] JMD layers contain three types of geometric shapes: points, lines, and polygons. All geometric shapes have a "Feature Name" attribute that indicates the name of the feature they represent. Based on this attribute, geometric shapes can contain various types of features, including but not limited to artificial hills, toilets, steps, pavilions, completed buildings, buildings under construction, suspended corridors, building pillars, colonnades, sheds, eaves, cantilevered corridors, and balconies. According to relevant building codes, only completed buildings, buildings under construction, toilets, suspended buildings, corridors, sheds, pavilions, and colonnades can constitute the base surface of a building.

[0031] S2. Obtain the top elevation of the building: Read the building base vector surface file and the point cloud data of the corresponding area. Perform spatial analysis on each building base vector surface in the building base vector surface file, extract all point cloud data that fall completely within its boundary range, and count the elevation values ​​of this part of the point cloud data. The highest frequency elevation value in the statistical results is determined as the top elevation H of the building. T The maximum and minimum elevation differences of the point cloud within this range are recorded synchronously as the top elevation variation range dH. T ; and H T and dH T This attribute is written into the building base vector surface file;

[0032] S3. Obtain the bottom elevation of the building: For each building base vector plane, establish a buffer zone at a certain distance around it, extract all point cloud data falling within this buffer zone, and calculate the elevation values ​​of this portion of point cloud data. The minimum elevation value in the statistical results is determined as the bottom elevation H of the building. B The difference between the maximum and minimum elevations of the point cloud within the synchronous recording buffer is used as the bottom elevation variation range dH. B ; and H B and dH B This attribute is written into the building base vector surface file;

[0033] S4. Manual interactive inspection of building elevation: The point cloud data is constructed into a visualized 3D surface model and loaded for display. Simultaneously, the building base vector surface is adjusted according to its corresponding top elevation H. T and bottom elevation H B Extrude each element into a 3D patch and overlay them for display; automatically identify the elevation range dH. T or dH B For building units exceeding a preset threshold, the operator is notified to compare them in a 3D visualization environment. The operator's judgment on whether the 3D patch conforms to the top and bottom terrain of the building as reflected in the 3D surface model is obtained. If the conclusion is no conformity, manual intervention is performed on the non-conforming building base vector surface to correct it and collect its H... T or H B Value; In step S4, the three-dimensional surface model is in OSGB format, and the preset threshold is set based on point cloud accuracy and regional building experience.

[0034] S5. Construct a 3D white model: Read the building base vector surface file that has undergone elevation checking and correction, and construct the model based on its recorded top elevation H. T With bottom elevation H B The difference is used to calculate the building height. Each building base vector plane is then stretched vertically to the calculated building height to generate a 3D block. The bottom planes of all 3D blocks are then uniformly adjusted to their corresponding H-shape. B Elevation location; Finally, perform a 3D Boolean union operation on multiple spatially connected 3D blocks belonging to the same entity (the same building) (merge the 3D blocks through the 3D Boolean union operation, and translate the bottom surface of the merged 3D block to its H position). B (The horizontal plane represented by the elevation) eliminates the overlapping parts in the merging process and merges them into a single, complete 3D white model of the building;

[0035] S6. Perform model texture mapping: Read the 3D white model of the building and decompose it into several independent planes, including elevations, top surfaces, and bottom surfaces. Calculate the normal direction of each plane and classify it as the bottom, top, or elevation of the building based on the normal direction. Calculate the width and height of each plane. Based on the dimensions corresponding to the texture, use integer division to calculate the number of rows and columns required for each plane. Then, automatically match the required type of texture map based on the attributes of the 3D white model of the building (when automatically matching texture maps, call the texture map that matches the attribute recorded in the building base vector plane file from the preset texture library). Perform texture mapping based on the number of rows and columns and the type of the texture map. Finally, recombine all the textured planes to generate a 3D model with a complete appearance texture. Output the 3D graphics and save them as a new data file as the architectural 3D model data result.

[0036] The technical principle of this invention is as follows: Irrelevant data in the topographic map is separated to extract planar geometric shapes belonging only to the building base; spatial analysis is performed on these geometric shapes and point clouds, using different analysis and statistical methods to extract the top and bottom elevations corresponding to each planar geometric shape; three-dimensional visualization data is generated from the building base surface and point cloud, and manual interactive inspection is conducted, with manual comparison of graphics showing abnormal indicators in the statistical values, and adjustments made to erroneous elevation values; the two-dimensional building base surface is stretched into a three-dimensional model, and the models are merged; the three-dimensional model is analyzed to calculate the texture information at different locations on the top and facade surfaces, completing the automatic texture mapping process.

[0037] The above descriptions are merely embodiments of the present invention. Commonly known structures and characteristics are not described in detail here. Those skilled in the art are aware of all common technical knowledge in the field prior to the application date or priority date, are aware of all existing technologies in that field, and have the ability to apply conventional experimental methods prior to that date. Those skilled in the art can, based on the guidance provided in this application, improve and implement this solution in combination with their own capabilities. Some typical known structures or methods should not be obstacles for those skilled in the art to implement this application. It should be noted that those skilled in the art can make several modifications and improvements without departing from the structure of the present invention. These should also be considered within the scope of protection of the present invention, and will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A method for automatically constructing a three-dimensional model of a building using architectural vector surfaces, characterized in that, Includes the following steps: S1. Preprocessing of building foundation surface: Extract the specified layer containing residential features from the topographic map data, and filter out the planar geometric shapes of built houses, houses in buildings, toilets, stilt houses, corridors, sheds, pavilions and colonnades based on the feature name attributes of the residential features, form the building foundation vector surface and save it as a building foundation vector surface file; S2, acquiring the top elevation of the building: reading the building base vector plane file and corresponding point cloud data, extracting the point cloud data within each building base vector plane range in the building base vector plane file, counting and taking the highest frequency elevation value as the top elevation H T , and writing into the building base vector plane file; S3, obtain the bottom elevation of the building: extract the point cloud data within the buffer zone of the outer periphery of each building base vector plane, count and take the minimum elevation value as the bottom elevation H B , write into the building base vector plane file; S4, Artificial interactive inspection of building elevation: the point cloud data is constructed into a three-dimensional surface model for display, and the building base vector surface is displayed according to H T and H B The stretch is displayed as a three-dimensional face sheet superposition; identify its elevation range dH T or dH B The building base vector surface exceeding the preset threshold value is provided for artificial inspection of the adhesion of its three-dimensional face sheet and the three-dimensional surface model, and the building base vector surface not adhering is artificially corrected in H T or H B Value; S5. Construct a 3D white model: Based on H T With H B The difference will stretch the building base vector plane into a three-dimensional block, and adjust the bottom of the three-dimensional block to H. B Elevation is used to merge different 3D blocks belonging to the same building into a single 3D white model of the building by performing Boolean operations. S6. Perform model texture mapping: Divide the 3D white model of the building into multiple planes and classify them. The planes include the elevation, top, and bottom. Calculate the number of rows and columns of texture maps required for each plane. Then, automatically match the required type of texture map according to the attributes of the 3D white model of the building. Perform texture mapping according to the number of rows and columns and the type of texture map, and reassemble to generate a textured 3D model.

2. The method for automatically constructing a three-dimensional building model using architectural vector surfaces according to claim 1, characterized in that: In step S1, the specified layer is the residential JMD layer defined in the topographic map specification, and is filtered based on the attribute value whose attribute field name is the feature name.

3. The method for automatically constructing a three-dimensional building model using architectural vector surfaces according to claim 2, characterized in that: In step S2, all point clouds that fall completely within the boundary of the building base vector plane are extracted, and the difference between the maximum and minimum elevations of the point clouds within this range is recorded as the top elevation variation range dH. T Write the building base vector surface file; in step S3, set the buffer distance according to the point cloud density and terrain, and simultaneously record the difference between the maximum and minimum elevations of the point cloud within the buffer as the bottom elevation variation range dH. B Write to the building base vector surface file.

4. The method for automatically constructing a three-dimensional model of a building using architectural vector surfaces according to claim 3, characterized in that: In step S4, the three-dimensional surface model is in OSGB format, and the preset threshold is set based on point cloud accuracy and regional building experience.

5. The method for automatically constructing a three-dimensional model of a building using architectural vector surfaces according to claim 4, characterized in that: In step S5, the Boolean operation is a three-dimensional Boolean union operation, which merges the three-dimensional blocks and translates the bottom surface of the merged three-dimensional block to its H. B The horizontal plane represented by elevation.

6. The method for automatically constructing a three-dimensional model of a building using architectural vector surfaces according to claim 5, characterized in that: In step S6, when classifying planes according to the direction of the normal, planes with vertically upward normals are classified as the top surface of the building, planes with vertically downward normals are classified as the bottom surface of the building, and the remaining planes are classified as the facades of the building.

7. The method for automatically constructing a three-dimensional model of a building using architectural vector surfaces according to claim 6, characterized in that: In step S6, when automatically matching texture maps, the texture map that matches the attribute recorded in the building base vector surface file is called from the preset texture library.

Images