A modeling and construction method and system based on BIM building model

Abstract

The present invention discloses a modeling and construction method and system based on BIM building model, comprising: obtaining building space information corresponding to a building to be constructed from multiple data sources respectively, so as to construct a three-dimensional space model corresponding to the building to be constructed; laying out each room in the three-dimensional space model according to building planning data, and generating a building facade image according to the target room layout, so as to add the building facade image to the three-dimensional space model, and construct a three-dimensional layout model; parsing the three-dimensional layout model, so as to obtain a plurality of components corresponding to the building facade image and component information of each of the components; extracting a plurality of semantic areas corresponding to the building facade image, so as to add the components to the three-dimensional layout model according to the semantic areas and the component information, and obtain a target three-dimensional model corresponding to the building to be constructed, so as to improve the accuracy of modeling.

Description

Technical Field

[0001] The present invention relates to the technical field of scene modeling, and in particular to a modeling construction method and system based on a BIM building model. Background Art

[0002] At present, the application of BIM technology in the field of engineering construction is developing rapidly. BIM technology can share and visualize parameters and information, thus providing assistance for engineering design, construction and operation and maintenance. In recent years, BIM technology has also gradually been applied in the fields of building renovation and construction. The established collaborative work platform for construction and operation and maintenance provides intuitive information management for the entire construction cycle.

[0003] However, traditional BIM modeling methods often rely on a single method in the collection of building data, such as manual measurement or simple instruments and equipment, which leads to insufficient data integrity and accuracy, making it difficult to form accurate basic building data for modeling. Secondly, when modeling the specific details of a building, it often relies on manual interpretation of architectural design documents, standard text materials, and architectural drawings, and then manually adding details to the model, which not only leads to extremely low modeling efficiency, but also is prone to errors, resulting in mismatches between the details of the model and the key information of the design. Summary of the invention

[0004] In order to solve the above technical problems, the present invention discloses a modeling and construction method and system based on BIM building model, which are used to improve the accuracy of modeling.

[0005] In order to achieve the above object, the present invention discloses a modeling and construction method based on a BIM building model, comprising:

[0006] Acquire architectural space information corresponding to the building to be constructed from multiple data sources respectively, and perform decision-level fusion on the architectural space information through a preset data fusion algorithm to construct a three-dimensional space model corresponding to the building to be constructed;

[0007] Acquire architectural planning data corresponding to the building to be constructed, and parse the architectural planning data through a pre-trained layout constraint model to obtain identification nodes corresponding to each room in the building to be constructed and constraints between the identification nodes;

[0008] According to the identified nodes and the constraint factors, obtaining several room layouts corresponding to the building to be constructed;

[0009] Extracting architectural style information corresponding to the building to be constructed from the architectural planning data, so as to determine a target room layout corresponding to the building to be constructed from the plurality of room layouts according to the architectural style information;

[0010] Generating a building facade image corresponding to the building to be constructed according to the target room layout, and adding the building facade image to the three-dimensional space model to construct a three-dimensional layout model corresponding to the building to be constructed;

[0011] Parsing the three-dimensional layout model, and acquiring a plurality of components corresponding to the building facade image and component information of each of the components from the building planning data;

[0012] A plurality of semantic regions corresponding to the building facade image are extracted, and the components are added to the three-dimensional layout model according to the semantic regions and the component information, so as to obtain a target three-dimensional model corresponding to the building to be constructed.

[0013] The present invention discloses a modeling and construction method based on BIM building model. First, basic building space information corresponding to the building to be constructed is obtained from multiple data sources to solve the technical problem of single data collection in the prior art. Further, the obtained multi-source spatial data is fused at the decision level to form unified and comprehensive building space data for modeling, thereby improving the accuracy of the three-dimensional spatial model after modeling. Secondly, the layout constraints between the rooms corresponding to the building to be constructed are obtained from the architectural planning data corresponding to the building to be constructed, so as to layout the rooms in the three-dimensional spatial model, so that the three-dimensional layout model finally established fits the actual functional requirements of the user and improves the practicality of modeling. Among them, when laying out each of the rooms, the corresponding architectural style information is extracted from the architectural planning data to determine the final room layout and ensure the accuracy of the room layout. Finally, on the three-dimensional layout model combining the room layout and the spatial data, the corresponding components are matched for each building facade in the three-dimensional layout model from the architectural planning data, and then the components are added to the model to obtain an accurate target three-dimensional model.

[0014] As a preferred example, the step of respectively acquiring the building space information corresponding to the building to be constructed from multiple data sources includes:

[0015] The ground area of ​​the building to be constructed is surveyed and mapped by traversing the planned walking path through a preset mobile measurement system, and mobile measurement system data with position marks is obtained; wherein the mobile measurement system data includes ground geometry data and ground texture data;

[0016] Scanning the internal space structure and space details of the building to be constructed by a preset three-dimensional laser to obtain three-dimensional laser data with position marks; wherein the three-dimensional laser data includes internal space structure data and internal space detail data;

[0017] Obtain a satellite remote sensing image of the building to be constructed, and perform image processing on the satellite remote sensing image to obtain remote sensing image data of the building to be constructed; wherein the remote sensing image data includes appearance data, contour data and surrounding environment information.

[0018] The present invention uses a mobile measurement system to obtain the ground data of the building, uses a three-dimensional laser to obtain the spatial structure data and spatial detail data inside the building, and uses satellite remote sensing to obtain the external overall data of the building, so as to improve the accuracy of modeling the three-dimensional spatial model corresponding to the building to be constructed by comprehensively collecting the basic spatial data of the building.

[0019] As a preferred example, the decision-level fusion of the building space information by a preset data fusion algorithm to construct a three-dimensional space model corresponding to the building to be constructed includes:

[0020] Performing correction and denoising processing on the mobile measurement system data, the three-dimensional laser data and the remote sensing image data to obtain first mobile measurement system data, first three-dimensional laser data and first remote sensing image data;

[0021] Extracting a ground point cloud corresponding to the first mobile measurement system data, a space point cloud corresponding to the first three-dimensional laser data, and a building point cloud corresponding to the first remote sensing image data;

[0022] An iterative closest point algorithm is used to perform one-to-one matching on the ground point cloud, the space point cloud and the building point cloud, and a weighted average method is used to fuse the matched multi-source point cloud data into a point cloud data set;

[0023] The point cloud data set is integrated into the BIM platform to construct a three-dimensional space model corresponding to the building to be constructed.

[0024] The present invention first performs correction and denoising processing on the collected multi-source spatial data to enhance the accuracy of the data, and then extracts point cloud data from the data to integrate data from different data sources into unified data for modeling based on the point cloud data, thereby ensuring the accuracy of the three-dimensional spatial model.

[0025] As a preferred example, the building planning data is parsed by the pre-trained layout constraint model to obtain the identification nodes corresponding to the rooms in the building to be constructed and the constraints between the identification nodes, including:

[0026] Acquire drawing data corresponding to the room layout requirements from the building planning data, identify and parse the drawing data, and acquire each marked point in the drawing data and the connection relationship and position relationship between the marked points; wherein the marked points are used to identify different rooms and different structures set in each room;

[0027] According to the connection relationship and the position relationship, the adjacent marked points are connected into lines to form a tree structure, thereby constructing a room layout constraint graph corresponding to the room layout requirement;

[0028] By parsing the room layout constraint graph, the identification nodes corresponding to each room in the building to be constructed and the constraints between the identification nodes are obtained; wherein the constraints include adjacent constraints between the room identification nodes and newly added room constraints; wherein the newly added room constraints include newly added room position constraints, newly added room shape constraints and newly added room size constraints.

[0029] The present invention first obtains drawing data corresponding to the layout requirements from the building planning data, and then obtains the room layout constraints corresponding to the building to be modeled by parsing the drawing data. When the room layout is executed according to the room layout constraints, it is ensured that the room layout meets the design requirements in the building planning, and the amount of data to be processed is reduced, thereby improving the efficiency and accuracy of modeling.

[0030] As a preferred example, the obtaining of several room layouts corresponding to the building to be constructed according to the identification node and the constraint factor includes:

[0031] Obtaining newly added room constraints corresponding to the room layout requirements;

[0032] According to the newly added room constraint, the identification node of the room is adjusted to obtain an adjusted room identification node;

[0033] Several room layouts are generated according to the adjusted room identification nodes and the corresponding constraints.

[0034] The present invention utilizes the newly added room constraints corresponding to the room layout requirements to iteratively generate rooms for the three-dimensional space model, thereby realizing automatic layout of the room and improving the efficiency of the room layout while meeting actual design requirements.

[0035] As a preferred example, extracting architectural style information corresponding to the building to be constructed from the architectural planning data, so as to determine a target room layout corresponding to the building to be constructed from a plurality of room layouts according to the architectural style information, includes:

[0036] Acquire design drawings corresponding to the style design requirements from the architectural planning data, and identify and analyze the design drawings to obtain design style information corresponding to the design drawings;

[0037] Establishing corresponding search tags for the design drawings according to the design style information, so as to perform search and analysis from a database according to the search tags to obtain the architectural style information corresponding to the building to be constructed and the style room layout corresponding to the architectural style information;

[0038] A second similarity between each of the room layouts and the style room layout is calculated to determine the target room layout from a plurality of the room layouts according to the second similarity.

[0039] After determining several layout requirements based on the room layout requirements, the present invention determines a target room layout that meets the style requirements from the several room layouts based on the requirements for the construction style in the architectural planning data, thereby improving the accuracy of the room layout. A database is used to save the style room layout corresponding to each architectural style, so as to improve the efficiency of selecting the target room layout by searching the database.

[0040] As a preferred example, generating a building facade image corresponding to the building to be constructed according to the target room layout includes:

[0041] Generating a room layout diagram corresponding to the building to be constructed according to the target room layout and the three-dimensional space model;

[0042] Inputting the room layout diagram into a pre-trained facade drawing model so that the facade drawing model extracts facade attributes of the room layout diagram, and then generates a first building facade image corresponding to the building to be constructed according to the facade attributes;

[0043] A corresponding second building facade image is obtained from the database according to the building style information, and the first building facade image is corrected according to the second building facade image to obtain a building facade image corresponding to the building to be constructed.

[0044] After the room layout is performed, the present invention needs to add building facades according to the room parts to improve the details of the modeling. In the process of improving the details, the corresponding building facade image is generated using the room layout diagram and the pre-trained facade drawing model to improve the details of the model. Furthermore, the second building facade image corresponding to the architectural style information is used to correct the first building facade image output by the facade drawing model, and information such as facade details can be added to improve the accuracy of the building facade image.

[0045] As a preferred example, the parsing of the three-dimensional layout model and obtaining a plurality of components corresponding to the building facade image and component information of each of the components from the building planning data include:

[0046] Analyzing the three-dimensional layout model to obtain spatial information and attribute information corresponding to each building facade in the three-dimensional layout model;

[0047] According to the spatial information and the attribute information, a plurality of components corresponding to each building facade and component information of each of the components are matched from the building planning data based on a pre-trained natural language processing model.

[0048] The present invention analyzes the three-dimensional layout model and the building planning data, and then matches the components with the building facades according to the spatial information and attribute information corresponding to each building facade and the component information corresponding to each component, so as to facilitate adding the components to the corresponding building facades, thereby improving the accuracy of modeling by improving the details of the model.

[0049] As a preferred example, the extracting of a plurality of semantic regions corresponding to each of the building facade images, so as to add the components to the three-dimensional layout model according to the semantic regions and the component information, includes:

[0050] Performing semantic segmentation on each of the building facades according to a preset SAM image segmentation network, thereby obtaining a plurality of semantic regions corresponding to the building facade;

[0051] Converting each of the semantic regions into a text description, and calculating the similarity between the text description and the component information, so as to match the semantic regions with the components one by one according to the similarity;

[0052] The component is added to the three-dimensional layout model according to the matching result to obtain a target three-dimensional model corresponding to the building to be constructed.

[0053] The present invention performs semantic segmentation on the building facade through the SAM image segmentation network, matches the corresponding added area for each component according to the result of the semantic segmentation, ensures the accuracy of the component configuration, and further improves the accuracy of the target three-dimensional model.

[0054] On the other hand, the present invention also discloses a modeling and construction system based on a BIM building model, the system comprising a space modeling module, a layout constraint module, a room layout module, a style extraction module, a function modeling module, a component analysis module and a component modeling module;

[0055] The spatial modeling module is used to obtain the building space information corresponding to the building to be constructed from multiple data sources respectively, and perform decision-level fusion on the building space information through a preset data fusion algorithm to construct a three-dimensional spatial model corresponding to the building to be constructed;

[0056] The layout constraint module is used to obtain the building planning data corresponding to the building to be constructed, and to parse the building planning data through the pre-trained layout constraint model to obtain the identification nodes corresponding to each room in the building to be constructed and the constraints between the identification nodes;

[0057] The room layout module is used to obtain several room layouts corresponding to the building to be constructed according to the identification node and the constraint factor;

[0058] The style extraction module is used to extract the architectural style information corresponding to the building to be constructed from the architectural planning data, so as to determine the target room layout corresponding to the building to be constructed from the several room layouts according to the architectural style information;

[0059] The functional modeling module is used to generate a building facade image corresponding to the building to be constructed according to the target room layout, and add the building facade image to the three-dimensional space model to construct a three-dimensional layout model corresponding to the building to be constructed;

[0060] The component parsing module is used to parse the three-dimensional layout model and obtain a plurality of components corresponding to the building facade image and component information of each of the components from the building planning data;

[0061] The component modeling module is used to extract several semantic areas corresponding to the building facade image, so as to add the components to the three-dimensional layout model according to the semantic areas and the component information, and obtain a target three-dimensional model corresponding to the building to be constructed.

[0062] The present invention discloses a modeling and construction system based on a BIM building model. First, basic building space information corresponding to a building to be constructed is obtained from multiple data sources to solve the technical problem of single data collection in the prior art. Further, the obtained multi-source spatial data is fused at the decision level to form a unified and comprehensive building space data for modeling, thereby improving the accuracy of the three-dimensional spatial model after modeling. Secondly, the layout constraints between the rooms corresponding to the building to be constructed are obtained from the architectural planning data corresponding to the building to be constructed, so as to layout the rooms in the three-dimensional spatial model, so that the three-dimensional layout model finally established fits the actual functional requirements of the user and improves the practicality of modeling. Among them, when laying out each of the rooms, the corresponding architectural style information is extracted from the architectural planning data to determine the final room layout and ensure the accuracy of the room layout. Finally, on the three-dimensional layout model combining the room layout and the spatial data, the corresponding components are matched for each building facade in the three-dimensional layout model from the architectural planning data, and then the components are added to the model to obtain an accurate target three-dimensional model. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 : A flow chart of a modeling and construction method based on a BIM building model provided by an embodiment of the present invention;

[0064] Figure 2 : A structural schematic diagram of a modeling and construction system based on a BIM building model provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0065] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0066] Reference Figure 1 , is a schematic diagram of a specific implementation process of a modeling and construction method based on a BIM building model provided by an embodiment of the present invention. Specifically, the construction method mainly includes steps 101 to 106, and the steps are mainly:

[0067] Step 101: Acquire architectural space information corresponding to the building to be constructed from multiple data sources respectively, and perform decision-level fusion on the architectural space information through a preset data fusion algorithm to construct a three-dimensional space model corresponding to the building to be constructed.

[0068] In this embodiment, the steps mainly include: traversing the planned walking path through a preset mobile measurement system, mapping the ground area of ​​the building to be constructed, and obtaining mobile measurement system data with position marks; wherein the mobile measurement system data includes ground geometry data and ground texture data; scanning the internal space structure and space details of the building to be constructed through a preset three-dimensional laser, and obtaining three-dimensional laser data with position marks; wherein the three-dimensional laser data includes internal space structure data and internal space detail data; obtaining a satellite remote sensing image of the building to be constructed, and performing image processing on the satellite remote sensing image to obtain remote sensing image data of the building to be constructed; wherein the remote sensing image data includes appearance data, contour data and surrounding environment information.

[0069] Furthermore, the mobile measurement system data, the three-dimensional laser data and the remote sensing image data are corrected and denoised to obtain first mobile measurement system data, first three-dimensional laser data and first remote sensing image data; the ground point cloud corresponding to the first mobile measurement system data, the space point cloud corresponding to the first three-dimensional laser data and the building point cloud corresponding to the first remote sensing image data are extracted; the ground point cloud, the space point cloud and the building point cloud are matched one by one using an iterative closest point algorithm, and the matched multi-source point cloud data are fused into a point cloud data set using a weighted average method; the point cloud data set is integrated into a BIM platform to construct a three-dimensional space model corresponding to the building to be constructed.

[0070] In some implementations of this embodiment, satellite remote sensing, ground mobile measurement systems, and indoor 3D laser scanners can be used to perform multi-source collection of basic data of buildings. Specifically, satellite remote sensing can be used to obtain the overall appearance, outline, and surrounding environment information of the building, the ground mobile measurement system can be used to obtain the detailed geometric shape and texture of the near-ground part of the building, and indoor 3D laser scanning can be used to obtain the internal spatial structure and details of the building.

[0071] After acquiring different spatial data of the building through the multi-source acquisition device, in order to enhance the accuracy of the data and delete abnormal or useless data in the data, the image data acquired by satellite remote sensing can be subjected to radiation correction, geometric correction and image enhancement processing, the data acquired by the ground mobile measurement system can be subjected to trajectory correction, point cloud filtering and noise removal, and the data acquired by indoor three-dimensional laser scanning can be aligned, spliced ​​and denoised, so as to improve the accuracy of the data.

[0072] After data enhancement is performed on the multi-source data, feature points of different data sources may be matched using a feature matching algorithm to fuse the multi-source data into a unified data set, so as to perform building modeling based on the unified data set.

[0073] Specifically, the weighted average method can be used to fuse the data. The calculation expression of the weighted average method is: ;

[0074] Wherein, F is the fused data value, Represents the weight corresponding to the nth data source. represents the data of the nth data source, and .

[0075] Furthermore, after obtaining the fused unified data, the unified spatial data is integrated into the BIM platform, and the BIM platform constructs a three-dimensional spatial model corresponding to the building to be constructed according to the location identifier contained in the data; the three-dimensional spatial model includes the coordinates of each location point.

[0076] In this embodiment, this step uses a mobile measurement system to obtain the ground data of the building, uses a three-dimensional laser to obtain the spatial structure data and spatial detail data inside the building, and uses satellite remote sensing to obtain the external overall data of the building, so as to improve the accuracy of the three-dimensional spatial modeling corresponding to the building to be constructed by comprehensively collecting the basic spatial data of the building. Then, the collected multi-source spatial data is first corrected and denoised to enhance the accuracy of the data, and then point cloud data in the data is extracted to integrate data from different data sources into unified data for modeling based on the point cloud data, thereby ensuring the accuracy of the three-dimensional spatial model.

[0077] Step 102: Acquire architectural planning data corresponding to the building to be constructed, and parse the architectural planning data through a pre-trained layout constraint model to obtain identification nodes corresponding to each room in the building to be constructed and constraints between the identification nodes.

[0078] In this embodiment, this step mainly includes: obtaining drawing data corresponding to the room layout requirements from the building planning data, identifying and parsing the drawing data, and obtaining each marked point in the drawing data and the connection relationship and position relationship between the marked points; wherein the marked points are used to identify different rooms and different structures set up in each room; according to the connection relationship and the position relationship, the adjacent marked points are connected into lines to form a tree structure, and then a room layout constraint map corresponding to the room layout requirements is constructed; by parsing the room layout constraint map, the identification nodes corresponding to each room in the building to be constructed and the constraints between the identification nodes are obtained; wherein the constraints include adjacent constraints between the room identification nodes and newly added room constraints; wherein the newly added room constraints include newly added room position constraints, newly added room shape constraints and newly added room size constraints.

[0079] In some implementations of this embodiment, by obtaining an annotated drawing corresponding to the room layout requirements, the annotated drawing is identified and parsed based on the first network channel of the trained layout constraint model to obtain each annotated point carried by the annotated drawing, as well as the connection relationship and position relationship between each annotated point. Then, based on each annotated point, the connection relationship and position relationship between each annotated point, a room layout constraint graph corresponding to the design requirements is generated. The first network channel of the layout constraint model can be trained based on the pre-annotated graph topology structure.

[0080] Among them, the annotated drawings corresponding to the room layout requirements can be obtained by the staff in advance by simple annotation according to the design requirements, wherein the annotated drawings carry multiple annotation points, with the connection relationship and position relationship between the annotation points. Specifically, the connection relationship between the annotation points is used to indicate which specific annotation points have a connection relationship, and the position relationship between the annotation points specifically includes the specific position of the annotation points on the annotated drawings, and the specific position can be the left side, right side, top and bottom of the drawing, as well as the relative position between different annotation points, such as how many other annotation points exist near a annotation point, and in which direction the other annotation points are located of the annotation point.

[0081] Furthermore, by inputting the acquired room layout constraint graph into the layout generator of the layout constraint model, room identification nodes corresponding to the room layout constraint graph and adjacent constraints between the room identification nodes can be generated.

[0082] The layout generator may be trained using a convolutional message transfer neural network (Conv-MPN), wherein a node represents a room as a feature quantity of the design space, and the features of the design space are updated through convolution to form a layout and adjacent constraints for verifying whether the rooms are in an adjacent relationship.

[0083] In this embodiment, this step first obtains drawing data corresponding to the layout requirements from the building planning data, and then obtains the room layout constraints corresponding to the building to be modeled by parsing the drawing data. When the room layout is executed according to the room layout constraints, it ensures that the room layout meets the design requirements in the building planning, reduces the amount of data to be processed, and improves the efficiency and accuracy of modeling.

[0084] Step 103: According to the identification node and the constraint factor, several room layouts corresponding to the building to be constructed are obtained.

[0085] In this embodiment, the steps mainly include: obtaining newly added room constraints corresponding to the room layout requirements; adjusting the identification nodes of the room according to the newly added room constraints to obtain adjusted room identification nodes; and generating several room layouts according to the adjusted room identification nodes and the corresponding constraints.

[0086] Specifically, in some implementations of this embodiment, based on a trained adversarial generative network, several corresponding room layouts may be generated according to room identification nodes, adjacent constraints between room identification nodes, and newly added user constraints.

[0087] In this embodiment, this step utilizes the newly added room constraints corresponding to the room layout requirements to iteratively generate rooms for the three-dimensional space model, thereby realizing automatic room layout while meeting actual design requirements and improving the efficiency of room layout.

[0088] Step 104: extracting architectural style information corresponding to the building to be constructed from the architectural planning data, so as to determine a target room layout corresponding to the building to be constructed from the plurality of room layouts according to the architectural style information.

[0089] In this embodiment, this step mainly includes: obtaining design drawings corresponding to the style design requirements from the architectural planning data, and identifying and analyzing the design drawings to obtain design style information corresponding to the design drawings; establishing corresponding retrieval tags for the design drawings according to the design style information, and performing retrieval and analysis from the database according to the retrieval tags to obtain the architectural style information corresponding to the building to be constructed and the style room layout corresponding to the architectural style information; calculating the second similarity between each of the room layouts and the style room layout, and determining the target room layout from a number of the room layouts according to the second similarity.

[0090] In some implementations of this embodiment, a natural language processing model is introduced to extract architectural style information representing the architectural style from the architectural planning data, wherein the architectural style information includes structural design features, structural design characteristics, building types, building unit types, architectural styles, and segmented prefabrication information of building modules in the design drawings.

[0091] According to the architectural style information, a corresponding search tag is established for the design drawings in the architectural planning data, and then a query is performed in a pre-built database according to the search tag to obtain architectural style information matching the search tag. The database includes a style tag, a style room layout, and a corresponding building facade image corresponding to each type of architectural style information.

[0092] Specifically, the field similarity between the search tag and the style tag is calculated to obtain the target building style information corresponding to the search tag according to the field similarity, and then obtain the style room layout corresponding to the target building style information. Further, the similarity between the style room layout and each of the room layouts is calculated, and the room layout with the largest similarity value is selected as the target room layout corresponding to the building to be constructed.

[0093] In this embodiment, after determining several layout requirements based on the room layout requirements, this step determines a target room layout that meets the style requirements from the several room layouts based on the requirements for the construction style in the architectural planning data, thereby improving the accuracy of the room layout. In particular, a database is used to save the style room layout corresponding to each architectural style, so as to improve the efficiency of selecting the target room layout by searching the database.

[0094] Step 105: Generate a building facade image corresponding to the building to be constructed according to the target room layout, and add the building facade image to the three-dimensional space model to construct a three-dimensional layout model corresponding to the building to be constructed.

[0095] In this embodiment, the step mainly includes: generating a room layout diagram corresponding to the building to be constructed according to the target room layout and the three-dimensional space model; inputting the room layout diagram into a pre-trained facade drawing model so that the facade drawing model extracts the facade attributes of the room layout diagram, and then generates a first building facade image corresponding to the building to be constructed according to the facade attributes; obtaining a corresponding second building facade image from the database according to the architectural style information, and correcting the first building facade image according to the second building facade image to obtain the building facade image corresponding to the building to be constructed.

[0096] In some implementations of this embodiment, according to the target room layout, the area corresponding to each room in the three-dimensional space model is displayed differently to establish a room layout diagram corresponding to the building to be constructed.

[0097] Furthermore, the room layout diagram is input into a pre-trained facade drawing model, so that the facade drawing model extracts facade attributes corresponding to the room layout diagram, and then generates a corresponding first building facade image according to the facade attributes. Specifically, the facade drawing model may adopt Stable-Diffusion.

[0098] Optionally, after generating the first building facade image, in order to increase the details of the facade image, a corresponding second building facade image can be obtained from the database according to the architectural style information, and the first building facade image can be corrected according to the second building facade image to obtain a building facade image corresponding to the building to be constructed.

[0099] Specifically, for each pixel in the second building facade image, a horizontal distance and a vertical distance between the pixel and a reference point in the second building facade image may be calculated; a horizontal scale parameter and a vertical scale parameter may be determined according to the horizontal distance and the vertical distance between the pixel and the reference point in the original image, and an approximate nonlinear regression calculation may be performed on the horizontal scale parameter, the vertical scale parameter and the coordinates of the pixel to correct the first building facade image.

[0100] In this embodiment, after the room layout is performed, the building facade needs to be added according to the room part to improve the modeling details. In the process of improving the details, the corresponding building facade image is generated using the room layout diagram and the pre-trained facade drawing model to improve the details of the model. Furthermore, the first building facade image output by the facade drawing model is corrected using the second building facade image corresponding to the architectural style information, and information such as facade details can be added to improve the accuracy of the building facade image.

[0101] Step 106: parsing the three-dimensional layout model, and acquiring a plurality of components corresponding to the building facade image and component information of each of the components from the building planning data.

[0102] In this embodiment, this step mainly includes: parsing the three-dimensional layout model to obtain spatial information and attribute information corresponding to each building facade in the three-dimensional layout model; matching a number of components corresponding to each building facade and component information of each of the components from the building planning data based on the pre-trained natural language processing model according to the spatial information and the attribute information.

[0103] In some implementations of this embodiment, the three-dimensional layout model is parsed to obtain floor information in the three-dimensional layout model, and the floor information includes the location information and functional attribute information corresponding to each building facade in each floor. A natural language processing model is introduced to obtain component information corresponding to each component in the building to be constructed from the building planning data; wherein the component information includes component function information and component space information.

[0104] A first feature corresponding to each building facade is extracted according to the position information and the functional attribute information; a second feature corresponding to each component is extracted according to the component function information and the component space information; a similarity between the first feature and the second feature is calculated, and the components are matched one by one with the building facade according to the similarity.

[0105] In this embodiment, this step parses the three-dimensional layout model and the building planning data, and then matches the components with the building facades according to the spatial information and attribute information corresponding to each building facade and the component information corresponding to each component, so as to facilitate adding the components to the corresponding building facades, thereby improving the details of the model and thus improving the accuracy of modeling.

[0106] Step 107: extracting a plurality of semantic regions corresponding to the building facade image, adding the components to the three-dimensional layout model according to the semantic regions and the component information, and obtaining a target three-dimensional model corresponding to the building to be constructed.

[0107] In this embodiment, this step mainly includes: performing semantic segmentation on each of the building facades according to a preset SAM image segmentation network, so as to obtain a plurality of semantic regions corresponding to the building facades; converting each of the semantic regions into a text description, and calculating the similarity between the text description and the component information, so as to match the semantic regions with the components one by one according to the similarity; adding the components to the three-dimensional layout model according to the matching results, so as to obtain a target three-dimensional model corresponding to the building to be constructed.

[0108] In some implementations of this embodiment, the building facade is semantically segmented by a SAM image segmentation network, and semantic regions including building components such as windows, balconies, air conditioner units, sun visors, doors, and walls that constitute the building facade are located. Each semantic type has a unique label for distinguishing other semantics within the facade. Then, the type corresponding to each semantic region is converted into a text description to calculate the similarity between the text description and the attribute information in each component, select the component with the highest similarity value, and add the component to the three-dimensional layout model to form the final target three-dimensional model.

[0109] In this embodiment, this step performs semantic segmentation on the building facade through the SAM image segmentation network, matches the corresponding added area for each component according to the result of the semantic segmentation, ensures the accuracy of the component configuration, and thus improves the accuracy of the target three-dimensional model.

[0110] Reference Figure 2 This embodiment also discloses a modeling and construction system based on the BIM building model, the system comprising a space modeling module 201, a layout constraint module 202, a room layout module 203, a style extraction module 204, a function modeling module 205, a component analysis module 206 and a component modeling module 207.

[0111] The space modeling module 201 is used to obtain the building space information corresponding to the building to be constructed from multiple data sources respectively, and perform decision-level fusion on the building space information through a preset data fusion algorithm to construct a three-dimensional space model corresponding to the building to be constructed.

[0112] The layout constraint module 202 is used to obtain the architectural planning data corresponding to the building to be constructed, so as to parse the architectural planning data through a pre-trained layout constraint model to obtain the identification nodes corresponding to each room in the building to be constructed and the constraints between the identification nodes.

[0113] The room layout module 203 is used to obtain several room layouts corresponding to the building to be constructed according to the identification nodes and the constraint factors.

[0114] The style extraction module 204 is used to extract the architectural style information corresponding to the building to be constructed from the architectural planning data, so as to determine the target room layout corresponding to the building to be constructed from the several room layouts according to the architectural style information.

[0115] The functional modeling module 205 is used to generate a building facade image corresponding to the building to be constructed according to the target room layout, and add the building facade image to the three-dimensional space model to construct a three-dimensional layout model corresponding to the building to be constructed.

[0116] The component parsing module 206 is used to parse the three-dimensional layout model and obtain a number of components corresponding to the building facade image and component information of each of the components from the building planning data.

[0117] The component modeling module 207 is used to extract a plurality of semantic regions corresponding to the building facade image, and to add the components to the three-dimensional layout model according to the semantic regions and the component information, so as to obtain a target three-dimensional model corresponding to the building to be constructed.

[0118] In certain implementations of this embodiment, an electronic device and a computer-readable storage medium are also disclosed, wherein the electronic device includes a memory and at least one processor: the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory so that the device performs a modeling and construction method based on a BIM building model; the computer-readable storage medium stores a computer program, and when the program is executed by the processor, the modeling and construction method based on a BIM building model is implemented.

[0119] In certain implementations of this embodiment, a modeling construction method and system based on a BIM building model is provided. First, basic building space information corresponding to the building to be constructed is obtained from multiple data sources to solve the technical problem of single data collection in the prior art. Further, the obtained multi-source spatial data is fused at the decision level to form a unified and comprehensive building space data for modeling, thereby improving the accuracy of the three-dimensional spatial model after modeling. Secondly, the layout constraints between the rooms corresponding to the building to be constructed are obtained from the architectural planning data corresponding to the building to be constructed, so as to layout each room in the three-dimensional spatial model, so that the three-dimensional layout model finally established fits the actual functional requirements of the user and improves the practicality of modeling. Among them, when laying out each of the rooms, the corresponding architectural style information is extracted from the architectural planning data to determine the final room layout and ensure the accuracy of the room layout. Finally, on the three-dimensional layout model combining the room layout and the spatial data, the corresponding components are matched for each building facade in the three-dimensional layout model from the architectural planning data, and then the components are added to the model to obtain an accurate target three-dimensional model.

[0120] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A modeling and construction method based on BIM building model, characterized in that: include: Acquire architectural space information corresponding to the building to be constructed from multiple data sources respectively, and perform decision-level fusion on the architectural space information through a preset data fusion algorithm to construct a three-dimensional space model corresponding to the building to be constructed; Acquire architectural planning data corresponding to the building to be constructed, and parse the architectural planning data through a pre-trained layout constraint model to obtain identification nodes corresponding to each room in the building to be constructed and constraints between the identification nodes; According to the identified nodes and the constraint factors, obtaining several room layouts corresponding to the building to be constructed; Extracting architectural style information corresponding to the building to be constructed from the architectural planning data, so as to determine a target room layout corresponding to the building to be constructed from the plurality of room layouts according to the architectural style information; Generating a building facade image corresponding to the building to be constructed according to the target room layout, and adding the building facade image to the three-dimensional space model to construct a three-dimensional layout model corresponding to the building to be constructed; Parsing the three-dimensional layout model, and acquiring a plurality of components corresponding to the building facade image and component information of each of the components from the building planning data; Extracting a plurality of semantic regions corresponding to the building facade image, adding the components to the three-dimensional layout model according to the semantic regions and the component information, and obtaining a target three-dimensional model corresponding to the building to be constructed; The method of parsing the building planning data through the pre-trained layout constraint model to obtain the identification nodes corresponding to each room in the building to be constructed and the constraints between the identification nodes includes: obtaining drawing data corresponding to the room layout requirements from the building planning data to identify and parse the drawing data to obtain each annotation point in the drawing data and the connection relationship and position relationship between the annotation points; wherein the annotation points are used to identify different rooms and different structures set in each room; according to the connection relationship and the position relationship, the adjacent annotation points are connected into lines to form a tree structure, and then a room layout constraint map corresponding to the room layout requirements is constructed; by parsing the room layout constraint map, the identification nodes corresponding to each room in the building to be constructed and the constraints between the identification nodes are obtained; wherein the constraints include adjacent constraints between room identification nodes and newly added room constraints; wherein the newly added room constraints include newly added room position constraints, newly added room shape constraints and newly added room size constraints; Among them, the extracting of architectural style information corresponding to the building to be constructed from the architectural planning data to determine a target room layout corresponding to the building to be constructed from several room layouts according to the architectural style information includes: obtaining design drawings corresponding to style design requirements from the architectural planning data, and identifying and analyzing the design drawings to obtain design style information corresponding to the design drawings; establishing corresponding retrieval tags for the design drawings according to the design style information, and performing retrieval and analysis from a database according to the retrieval tags to obtain the architectural style information corresponding to the building to be constructed and the style room layout corresponding to the architectural style information; calculating a second similarity between each room layout and the style room layout, and determining the target room layout from several room layouts according to the second similarity.

2. A modeling and construction method based on a BIM building model according to claim 1, characterized in that: The step of respectively obtaining the building space information corresponding to the building to be constructed from the multiple data sources includes: The ground area of ​​the building to be constructed is surveyed and mapped by traversing the planned walking path through a preset mobile measurement system, and mobile measurement system data with position marks is obtained; wherein the mobile measurement system data includes ground geometry data and ground texture data; Scanning the internal space structure and space details of the building to be constructed by a preset three-dimensional laser to obtain three-dimensional laser data with position marks; wherein the three-dimensional laser data includes internal space structure data and internal space detail data; Obtain a satellite remote sensing image of the building to be constructed, and perform image processing on the satellite remote sensing image to obtain remote sensing image data of the building to be constructed; wherein the remote sensing image data includes appearance data, contour data and surrounding environment information.

3. A modeling and construction method based on a BIM building model according to claim 2, characterized in that: The decision-level fusion of the building space information by a preset data fusion algorithm to construct a three-dimensional space model corresponding to the building to be constructed includes: Performing correction and denoising processing on the mobile measurement system data, the three-dimensional laser data and the remote sensing image data to obtain first mobile measurement system data, first three-dimensional laser data and first remote sensing image data; Extracting a ground point cloud corresponding to the first mobile measurement system data, a space point cloud corresponding to the first three-dimensional laser data, and a building point cloud corresponding to the first remote sensing image data; An iterative closest point algorithm is used to perform one-to-one matching on the ground point cloud, the space point cloud and the building point cloud, and a weighted average method is used to fuse the matched multi-source point cloud data into a point cloud data set; The point cloud data set is integrated into the BIM platform to construct a three-dimensional space model corresponding to the building to be constructed.

4. The modeling and construction method based on the BIM building model according to claim 1, characterized in that: The step of obtaining, according to the identification node and the constraint factor, several room layouts corresponding to the building to be constructed includes: Obtaining newly added room constraints corresponding to the room layout requirements; According to the newly added room constraint, the identification node of the room is adjusted to obtain an adjusted room identification node; Several room layouts are generated according to the adjusted room identification nodes and the corresponding constraints.

5. The modeling and construction method based on the BIM building model according to claim 1, characterized in that: The step of generating a building facade image corresponding to the building to be constructed according to the target room layout includes: Generating a room layout diagram corresponding to the building to be constructed according to the target room layout and the three-dimensional space model; Inputting the room layout diagram into a pre-trained facade drawing model so that the facade drawing model extracts facade attributes of the room layout diagram, and then generates a first building facade image corresponding to the building to be constructed according to the facade attributes; A corresponding second building facade image is obtained from the database according to the building style information, and the first building facade image is corrected according to the second building facade image to obtain a building facade image corresponding to the building to be constructed.

6. A modeling and construction method based on a BIM building model according to claim 1, characterized in that: The step of analyzing the three-dimensional layout model and obtaining a plurality of components corresponding to the building facade image and component information of each of the components from the building planning data includes: Analyzing the three-dimensional layout model to obtain spatial information and attribute information corresponding to each building facade in the three-dimensional layout model; According to the spatial information and the attribute information, a plurality of components corresponding to each building facade and component information of each of the components are matched from the building planning data based on a pre-trained natural language processing model.

7. The modeling and construction method based on the BIM building model according to claim 1, characterized in that: The extracting a plurality of semantic regions corresponding to each of the building facade images, and adding the components to the three-dimensional layout model according to the semantic regions and the component information, comprises: Performing semantic segmentation on each of the building facades according to a preset SAM image segmentation network, thereby obtaining a plurality of semantic regions corresponding to the building facade; Converting each of the semantic regions into a text description, and calculating the similarity between the text description and the component information, so as to match the semantic regions with the components one by one according to the similarity; The component is added to the three-dimensional layout model according to the matching result to obtain a target three-dimensional model corresponding to the building to be constructed.

8. A modeling and construction system based on a BIM building model, characterized in that: The system includes a space modeling module, a layout constraint module, a room layout module, a style extraction module, a function modeling module, a component parsing module and a component modeling module; The spatial modeling module is used to obtain the building space information corresponding to the building to be constructed from multiple data sources respectively, and perform decision-level fusion on the building space information through a preset data fusion algorithm to construct a three-dimensional spatial model corresponding to the building to be constructed; The layout constraint module is used to obtain the building planning data corresponding to the building to be constructed, and to parse the building planning data through the pre-trained layout constraint model to obtain the identification nodes corresponding to each room in the building to be constructed and the constraints between the identification nodes; The room layout module is used to obtain several room layouts corresponding to the building to be constructed according to the identification node and the constraint factor; The style extraction module is used to extract the architectural style information corresponding to the building to be constructed from the architectural planning data, so as to determine the target room layout corresponding to the building to be constructed from the several room layouts according to the architectural style information; The functional modeling module is used to generate a building facade image corresponding to the building to be constructed according to the target room layout, and add the building facade image to the three-dimensional space model to construct a three-dimensional layout model corresponding to the building to be constructed; The component parsing module is used to parse the three-dimensional layout model and obtain a plurality of components corresponding to the building facade image and component information of each of the components from the building planning data; The component modeling module is used to extract a plurality of semantic areas corresponding to the building facade image, so as to add the components to the three-dimensional layout model according to the semantic areas and the component information, and obtain a target three-dimensional model corresponding to the building to be constructed; The method of parsing the building planning data through the pre-trained layout constraint model to obtain the identification nodes corresponding to each room in the building to be constructed and the constraints between the identification nodes includes: obtaining drawing data corresponding to the room layout requirements from the building planning data to identify and parse the drawing data to obtain each annotation point in the drawing data and the connection relationship and position relationship between the annotation points; wherein the annotation points are used to identify different rooms and different structures set in each room; according to the connection relationship and the position relationship, the adjacent annotation points are connected into lines to form a tree structure, and then a room layout constraint map corresponding to the room layout requirements is constructed; by parsing the room layout constraint map, the identification nodes corresponding to each room in the building to be constructed and the constraints between the identification nodes are obtained; wherein the constraints include adjacent constraints between room identification nodes and newly added room constraints; wherein the newly added room constraints include newly added room position constraints, newly added room shape constraints and newly added room size constraints; Among them, the extracting of architectural style information corresponding to the building to be constructed from the architectural planning data to determine a target room layout corresponding to the building to be constructed from several room layouts according to the architectural style information includes: obtaining design drawings corresponding to style design requirements from the architectural planning data, and identifying and analyzing the design drawings to obtain design style information corresponding to the design drawings; establishing corresponding retrieval tags for the design drawings according to the design style information, and performing retrieval and analysis from a database according to the retrieval tags to obtain the architectural style information corresponding to the building to be constructed and the style room layout corresponding to the architectural style information; calculating a second similarity between each room layout and the style room layout, and determining the target room layout from several room layouts according to the second similarity.

Images