BIM-based three-dimensional model rapid generation method, device, system and medium

Abstract

The application discloses a BIM-based three-dimensional model rapid generation method, device, system and medium, the method extracts language features from a user requirement document, and combines project basic information to generate project building information based on a programming language, which is used for generating two-dimensional drawings, greatly reducing the design time of the project, and setting a professional area as a non-editable constraint area in the BIM system, so that overdesign of the professional area by the designer under the premise of insufficient communication is avoided, thereby reducing the workload of later modification, and balancing the design progress and the owner demand.

Description

Technical Field

[0001] The present invention relates to the technical field of building models, and in particular to a method, device, system and medium for quickly generating a three-dimensional model based on BIM. Background Art

[0002] BIM (Building Information Modeling) is an indispensable information technology tool in engineering design, construction, and management. By centrally integrating digital and information-based models of building projects, it enables information sharing and efficient communication throughout the project lifecycle. BIM technology enables engineering and construction teams to more accurately understand and respond to various construction information, providing a solid foundation for collaborative work.

[0003] Although BIM is an efficient information model, its application involves a lot of repetitive work. For example, in order to obtain a preliminary design plan, the owner will require the design unit to introduce a BIM management system in the early stages of the project and use it as the basis for subsequent modifications. However, it is difficult for project designers to consider the impact of the installation locations of equipment, pipelines, etc. on operations from the perspective of the owner, and the owner often only has a basis and direction for modification after seeing the specific design plan. As a result, the BIM model needs to be revised several times in the future to meet the requirements of the owner. Moreover, the more detailed the BIM model is in the early stages, the more difficult it will be to modify it in the later stages, which greatly increases the workload of the project designer in the early stages and slows down the progress of the project. Summary of the Invention

[0004] In response to the above problems, the present invention proposes a method, device, system and medium for rapid generation of three-dimensional models based on BIM, which mainly solves the problems of the background technology.

[0005] To solve the above technical problems, the first aspect of the present invention proposes a method for quickly generating a three-dimensional model based on BIM, comprising the following steps:

[0006] Obtaining a user requirement document, and extracting language features from the user requirement document;

[0007] The language features are combined with basic project information to generate project building information based on the programming language;

[0008] Calling the drawing API interface of the two-dimensional drawing software to generate a two-dimensional drawing according to the project building information;

[0009] The two-dimensional drawings are input into the BIM system, the two-dimensional elements are associated and matched with the BIM family library through the spatial topological relationship mapping algorithm, and the professional areas in the basic information of the project are set as non-editable constraint areas, and a three-dimensional model that complies with the IFC standard is automatically generated.

[0010] In some implementations, the language feature extraction process includes:

[0011] Preprocessing the user demand document, removing special symbols in the document and performing word segmentation processing to obtain a preprocessed text;

[0012] Construct an entity vocabulary in the architectural field. This entity vocabulary adopts a dynamic update mechanism, uses a BiLSTM-CRF model to incrementally learn unregistered words, and uses context to perform entity disambiguation. The entity vocabulary includes at least space types, materials, and regulatory requirements.

[0013] The entity vocabulary is integrated into the pre-trained language model feature space, domain adaptation is achieved through transfer learning, entity recognition is performed on the pre-processed text, and the language features are obtained.

[0014] In some embodiments, the basic project information further includes a pre-processing step, the process including:

[0015] Structuring the basic project information to generate site feature analysis data and building space feature analysis data based on programming language;

[0016] Performing geometric parameter conversion on the irregular terrain in the site feature analysis data to generate polygonal data;

[0017] The site feature analysis data, the building space feature analysis data and the polygon data are fused to obtain the processed basic information of the project.

[0018] In some implementations, the process of generating the project building information includes:

[0019] Dependency syntax analysis is used to extract the spatial relationship and attribution parameters of each feature in the language features, and each feature is combined into multiple feature groups according to the spatial relationship and the attribution parameters. Data related to the feature groups is searched from the basic project information. A spatial relationship graph inference engine is used to convert the feature groups into direction and size parameters based on a programming language, and the parameters are used to modify the corresponding data in the basic project information to form the project building information.

[0020] In some embodiments, the process of generating the two-dimensional drawing includes:

[0021] The drawing API interface of the two-dimensional drawing software is called to extract the site information, site available area information, building space, and facility location and equipment location preference information contained in the project building information within the two-dimensional drawing software, and a mixed integer programming algorithm is used to optimize the spatial layout. Hard distance constraints are set to limit each facility to be completely located within the site, each equipment to be completely located within the corresponding building space, and facilities and equipment to not be allowed to overlap. Finally, the corresponding two-dimensional drawing is output.

[0022] In some embodiments, the process of generating the three-dimensional model includes:

[0023] Detecting the professional areas included in the basic information of the project, implementing digital signature verification on the professional areas using a role-based access control mechanism, and generating the three-dimensional model;

[0024] Locking the editable elements of the professional area and setting a first mark to highlight the position of the professional area;

[0025] A second sign is provided to highlight pipelines that cannot be bypassed from the professional area.

[0026] In some embodiments, an error correction mechanism is also set up, one or more open spaces are preset, and a constraint propagation network is used to detect whether there is a spatial topological conflict in the basic information of the project. When the conflict degree is greater than a preset threshold, the error correction mechanism is triggered, and the functional areas with demand conflicts in the three-dimensional model are separated by the open spaces.

[0027] A second aspect of the present invention provides a BIM-based 3D model rapid generation system, comprising:

[0028] A natural language analysis module, used to obtain user requirement documents and extract language features from the user requirement documents;

[0029] A project construction information generation module, configured to generate project construction information based on a programming language by combining the language features with basic project information;

[0030] A two-dimensional drawing generation module is used to call the drawing API interface of the two-dimensional drawing software to generate a two-dimensional drawing according to the project building information;

[0031] The three-dimensional model generation module is used to input the two-dimensional drawings into the BIM system, associate and match the two-dimensional elements with the BIM family library through a spatial topological relationship mapping algorithm, and set the professional area in the basic information of the project as an uneditable constraint area to automatically generate a three-dimensional model that complies with the IFC standard.

[0032] The third aspect of the present invention provides a device for quickly generating a three-dimensional model based on BIM, the device comprising a memory, a processor and a communication module, wherein:

[0033] The memory is used to store executable program code;

[0034] The processor is coupled to the memory;

[0035] The processor calls the executable program code stored in the memory to execute the above-mentioned BIM-based three-dimensional model rapid generation method.

[0036] A fourth aspect of the present invention provides a computer-storable medium, which stores computer instructions. When the computer instructions are called, they are used to execute the above-mentioned BIM-based three-dimensional model rapid generation method.

[0037] The beneficial effects of the present invention are: by extracting language features from user requirement documents and combining them with basic project information to generate project building information based on a programming language for generating two-dimensional drawings, the design time of the project is greatly reduced. In addition, setting professional areas as non-editable constraint areas in the BIM system can avoid designers from over-designing professional areas without sufficient communication, thereby reducing the workload of later modifications and balancing the design progress and owner needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 A schematic diagram of a process for rapidly generating a three-dimensional model based on BIM is disclosed in the first embodiment of the present invention;

[0039] Figure 2 A schematic diagram of the structure of a BIM-based 3D model rapid generation system is disclosed as a second embodiment of the present invention;

[0040] Figure 3 This is a structural schematic diagram of a BIM-based three-dimensional model rapid generation device according to the third embodiment of the present invention. DETAILED DESCRIPTION

[0041] To make the objectives, technical solutions, and advantages of the present invention more clear and distinct, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the present invention and are not intended to limit the present invention. It should also be noted that, for ease of description, the accompanying drawings only illustrate portions relevant to the present invention, not all of the present invention.

[0042] Example 1

[0043] This embodiment proposes a method for rapid generation of three-dimensional models based on BIM, which is applicable to the field of civil construction and does not include linear projects such as bridges and tunnels. It aims to extract language features from user requirement documents and generate project building information based on programming language in combination with basic project information for generating two-dimensional drawings, thereby greatly reducing the design time of the project. In addition, setting professional areas as non-editable constraint areas in the BIM system can prevent designers from over-designing professional areas without sufficient communication, thereby reducing the workload of later modifications and balancing the design progress and owner needs.

[0044] This method mainly includes the following steps:

[0045] S1, obtain user requirement documents and extract language features from the user requirement documents.

[0046] In S1, the user requirement document can be a project book provided by the owner or minutes of a collective meeting. To improve recognition efficiency, the keyword extraction part needs to be processed using a natural language analysis model. For example, the user's requirement description may mention "open layout," "natural lighting," "conference room," "storage space," etc. These keywords need to be identified using a natural language library such as NLTK, spaCy, or more advanced models such as BERT. In one example, the language feature extraction process includes:

[0047] S101, pre-processing the user demand document, removing special symbols in the document and performing word segmentation processing to obtain a pre-processed text.

[0048] S102: Build an entity vocabulary for the architectural domain. This vocabulary uses a dynamic update mechanism, incrementally learning unregistered terms using a BiLSTM-CRF model and disambiguating entities using context. The entity vocabulary includes at least space type, material, and regulatory requirements. Dynamically updating the architectural entity vocabulary using the BiLSTM-CRF model improves the recognition accuracy of unregistered terms. Optionally, set a trigger mechanism for incremental learning: when the entity recognition confidence level falls below 80%, manual review is automatically triggered. Approved data is added to the training set for incremental model updates.

[0049] In step S103, the entity vocabulary is integrated into the feature space of the pre-trained language model. Domain adaptation is achieved through transfer learning. Entity recognition is performed on the pre-processed text to identify nouns or domain-specific terms, thereby obtaining language features. In this step, transfer learning technology adapts the general NLP model to the architectural domain, reducing entity disambiguation errors. The pre-trained language model can be based on the BERT-Base model, freezing the parameters of the first six layers of the network and fine-tuning the top layer using architectural domain data.

[0050] S2, language features are combined with basic project information to generate project building information based on programming language.

[0051] The project basic information also includes preprocessing steps, which include:

[0052] S201: Structuring the basic project information to generate site feature analysis data and building space feature analysis data based on a programming language. The basic project information includes information such as the plot size, shape, and permitted design location of the building. The site feature analysis data and building space feature analysis data are generally presented in text or tables and need to be structured and parsed into a format that can be understood by CAD software, such as coordinates, lengths, angles, etc. Using regular expressions or specific parsing libraries to process these data can reduce the time spent on format conversion and the workload of manual annotation.

[0053] S202 performs geometric parameter conversion on the irregular terrain in the site feature analysis data to generate polygon data. The purpose is to convert a string of coordinate points described by text into a list of coordinate points. The polygon terrain conversion supports direct import into CAD software to improve the efficiency of terrain modeling.

[0054] S203, fusing the site feature analysis data, the building space feature analysis data and the polygon data to obtain processed basic project information. Optionally, the basic project information is constructed using a Python program.

[0055] Specifically, the process of generating project building information includes:

[0056] S204 uses dependency parsing to extract the spatial relationships and attribution parameters of each feature in the language features. The features are then grouped into multiple feature groups based on their spatial relationships and attribution parameters. Data related to the feature groups is searched for from the project's basic information. A spatial relationship graph inference engine is used to convert the feature groups into programming language-based direction and size parameters. These are then used to modify the corresponding data in the project's basic information to form the project's building information. Specifically, the layout requirements mentioned in the user requirements document included in S1 may include functional zoning, flow lines, orientation, spacing, etc. It is necessary to first analyze how to extract these layout-related keywords from the text. For example, NLP can be used to identify specific entities, such as "active and static zoning" and "north-south permeability," as well as the functional areas corresponding to these layouts. For example, if a paragraph states, "The outpatient department and CT laboratory are at least 50 meters apart, and the outpatient department windows face south," spatial relationship graph inference can be used to group "outpatient department," "CT laboratory," and "50 meters" into a feature group and converted into a Python program for precise parameter mapping. The Python program converted from the above feature group is input into the established project building information for search, locating the "outpatient room" and "CT laboratory" positions, and replacing the original values ​​with "50 meters" as the modified value, automatically modifying the distance between the two functional areas. The direction parameter "south" of the "outpatient room window" is expressed in Euler angles under the world coordinate system, and the size parameters comply with the building module coordination standard, forming the final project building information for subsequent 2D drawing generation.

[0057] S3, calls the drawing API interface of the two-dimensional drawing software to generate a two-dimensional drawing based on the project building information.

[0058] The process of generating 2D drawings includes:

[0059] By invoking the drawing API of the 2D drawing software, the project building information, including site information, site availability, building space, and facility and equipment location preferences, is extracted. A mixed integer programming algorithm is then used to optimize the spatial layout, imposing hard distance constraints such as requiring each facility to be completely within the site, each piece of equipment to be completely within its corresponding building space, and preventing overlap between facilities and equipment. The resulting 2D drawings are then output, reducing the time required to generate them. Alternatively, the pyautocad library (AutoCAD) or the FreeCAD API is recommended. Monte Carlo simulations are performed on the equipment location preferences to generate layout plans that meet a ±5% position tolerance, minimizing rework.

[0060] As another important aspect of the present invention, the method for calculating the device location preference information mainly includes:

[0061] S301, calculate the location preference probability density of the device:

[0062] (1);

[0063] Where K is the total number of building usable areas, d k is the actual distance from device i to the available area of ​​the kth building, μ k is the ideal distance from device i to the kth building’s available area, σ k To allow deviation.

[0064] S302, setting association preferences between devices:

[0065] (2);

[0066] Among them, p i is the center coordinate of device i, p j is the center coordinate of device j, L max is the maximum allowable spacing, L min The minimum safety distance.

[0067] S303, Latin hypercube sampling generates a candidate set of device locations:

[0068] (3);

[0069] Among them, x i （k） and y i （k） are the candidate horizontal and vertical coordinates of the k-th building available area, respectively, x min and y min is the minimum horizontal coordinate and minimum vertical coordinate of the k-th building usable area, s k is the offset of the k-th building’s available area, Δx and Δy are the length and width of the k-th building’s available area, respectively, and rand∼U(0, 1) represents a uniformly distributed random number.

[0070] S304, perform feasibility check on each candidate sample:

[0071] (4);

[0072] Where C = {c1, c2, ...., cm} represents a set of hard location constraints (such as prohibiting overlap and setting a minimum spacing of 0.5 m), and P = {p1, p2, ...., pn} represents a set of preference weights (such as setting the probability weight of a CT device near a power source to 0.7).

[0073] S305, setting a comprehensive scoring function for device layout;

[0074] (5);

[0075] Among them, 0.6 is the preset weight coefficient of preference, 0.4 is the preset weight coefficient of cost, α+β=1, f pref （k） (i) is the location preference probability density of device i, f cost （k） (i) is the placement cost of device i, and Norm(⋅) represents the normalization function.

[0076] After the above calculations, an equipment layout plan with extremely high preference satisfaction can be obtained, thereby reducing the number of rework times.

[0077] S4, input the two-dimensional drawings into the BIM system, associate and match the two-dimensional elements with the BIM family library through the spatial topology relationship mapping algorithm, and set the professional area in the basic information of the project as an uneditable constraint area to automatically generate a three-dimensional model that complies with the IFC standard.

[0078] The process of generating a 3D model includes:

[0079] S401, using a role-based access control mechanism to verify the digital signature of the professional areas included in the basic information of the inspection project, and generating the three-dimensional model. The professional areas described in this embodiment refer to the areas where the main professional business of the owner unit is located. Taking a medical building as an example, its CT examination room, biochemical laboratory, biochemical waste treatment room and other functional areas belong to the professional areas. Under the premise of compliance, the owner unit generally has some specific requirements, which are recorded in the above-mentioned user requirements document.

[0080] S402: Lock the editable elements in the professional area and set a first mark to highlight the position of the professional area.

[0081] S403: Setting a second mark to highlight pipelines that cannot be bypassed from the professional area.

[0082] Among them, the above-mentioned first identification / second identification can optionally add a ConstraintZone mark in the IFC property set, and render the red / yellow translucent warning area through the BIM visualization engine, such as using the BlenderBIM plug-in based on the OpenBIM workflow to implement custom property rendering.

[0083] S404 also includes setting up a correction mechanism by pre-setting one or more open spaces. These open spaces serve as buffer zones, with an area no less than 15% of the total area of ​​the conflicting areas and meeting bidirectional evacuation width requirements. A constraint propagation network (e.g., abstracting building elements into nodes and using MiniZinc to build a CSP model with distance and orientation constraints) is used to detect whether there are spatial topological conflicts within the project's basic information. This is intended to reduce the need for manual review. When the conflict level exceeds a preset threshold, a correction mechanism is triggered (based on a confusion matrix analysis of 100 test cases, the threshold value at the maximum F1-score is selected as the preset threshold). Functional areas within the 3D model with conflicting requirements are separated by open spaces. In one example, a soundproof corridor can be used as the open space to achieve active and quiet zoning. This means that a soundproof corridor is set between areas requiring quiet (such as outpatient clinics) and areas likely to generate noise (such as the check-in counter) to reduce noise transmission and thereby improve living or working comfort. Alternatively, buffer courtyards are recommended to separate different functional areas. By designing one or more courtyards or open spaces, different functional areas within a building (such as the outpatient department and the inpatient department) can be effectively separated. The courtyards also provide natural light and ventilation, increasing the sense of openness and comfort of the space. The above zoning methods aim to resolve conflicts or irrationalities in the layout through physical means (soundproof corridors and buffer courtyards), thereby optimizing space efficiency and work environment quality, and reducing the number of design revisions.

[0084] Example 2

[0085] This embodiment proposes a BIM-based 3D model rapid generation system. Figure 2 Shown, including:

[0086] The natural language analysis module 201 is used to obtain user requirement documents and extract language features from the user requirement documents;

[0087] The project building information generating module 202 is used to generate the project building information based on the programming language by combining the language features with the basic project information;

[0088] The two-dimensional drawing generation module 203 is used to call the drawing API interface of the two-dimensional drawing software to generate a two-dimensional drawing according to the project building information;

[0089] The three-dimensional model generation module 204 is used to input the two-dimensional drawings into the BIM system, associate and match the two-dimensional elements with the BIM family library through the spatial topological relationship mapping algorithm, and set the professional area in the basic information of the project as an uneditable constraint area to automatically generate a three-dimensional model that complies with the IFC standard.

[0090] The functions implemented by the above modules can be referred to S1-S4 described in Example 1, and will not be repeated here.

[0091] Example 3

[0092] See also Figure 3 , Figure 3 This is a schematic diagram of the structure of another BIM-based three-dimensional model rapid generation device disclosed in an embodiment of the present invention. Figure 3 As shown, the control platform may include:

[0093] A memory 301 storing executable program code;

[0094] a processor 302 coupled to the memory 301;

[0095] The processor 302 calls the executable program code stored in the memory 301 to execute the steps of the BIM-based three-dimensional model rapid generation method described in the first embodiment.

[0096] Example 4

[0097] An embodiment of the present invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to enable a computer to execute the steps of the BIM-based three-dimensional model rapid generation method described in Example 1.

[0098] The device embodiments described above are merely illustrative, wherein the modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules, i.e., they may be located in one place or distributed across multiple network modules. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Those skilled in the art can understand and implement the present invention without inventive effort.

[0099] Through the detailed description of the above embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, or of course, by means of hardware. Based on this understanding, the above technical solution, in essence, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM), or other optical disk storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0100] Finally, it should be noted that the method, device, system and medium for rapid generation of three-dimensional models based on BIM disclosed in the embodiments of the present invention are only preferred embodiments of the present invention, and are only used to illustrate the technical solutions of the present invention, rather than to limit them. Although the present invention has been described in detail with reference to the aforementioned embodiments, it should be understood by those skilled in the art that the technical solutions described in the aforementioned embodiments can still be modified, or some of the technical features therein can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments.

Claims

1. A method for rapid generation of three-dimensional models based on BIM, characterized in that: The following steps are involved: Obtaining a user requirement document, and extracting language features from the user requirement document; The language feature extraction process includes: preprocessing the user demand document, removing special symbols in the document, and performing word segmentation to obtain a preprocessed text; constructing an entity vocabulary in the architectural field, wherein the entity vocabulary adopts a dynamic update mechanism, incrementally learns unregistered words through a BiLSTM-CRF model, and uses context to perform entity disambiguation, and the entity vocabulary includes at least space type, material, and specification requirements; integrating the entity vocabulary into the pretrained language model feature space, achieving domain adaptation through transfer learning, and performing entity recognition on the preprocessed text to obtain the language features; The language features are combined with basic project information to generate project building information based on the programming language; Calling the drawing API interface of the two-dimensional drawing software to generate a two-dimensional drawing according to the project building information; The two-dimensional drawings are input into the BIM system, the two-dimensional elements are associated and matched with the BIM family library through the spatial topological relationship mapping algorithm, and the professional areas in the basic information of the project are set as non-editable constraint areas, and a three-dimensional model that complies with the IFC standard is automatically generated.

2. The method for rapid generation of a three-dimensional model based on BIM according to claim 1, wherein: The basic project information also includes pre-processing steps, which include: Structuring the basic project information to generate site feature analysis data and building space feature analysis data based on programming language; Performing geometric parameter conversion on the irregular terrain in the site feature analysis data to generate polygonal data; The site feature analysis data, the building space feature analysis data and the polygon data are fused to obtain the processed basic information of the project.

3. The method for rapid generation of a three-dimensional model based on BIM according to claim 1, wherein: The process of generating the project building information includes: Dependency syntax analysis is used to extract the spatial relationship and attribution parameters of each feature in the language features, and each feature is combined into multiple feature groups according to the spatial relationship and the attribution parameters. Data related to the feature groups is searched from the basic project information. A spatial relationship graph inference engine is used to convert the feature groups into direction and size parameters based on a programming language, and the parameters are used to modify the corresponding data in the basic project information to form the project building information.

4. The method for rapid generation of a three-dimensional model based on BIM according to claim 1, wherein: The process of generating the two-dimensional drawing includes: The drawing API interface of the two-dimensional drawing software is called to extract the site information, site available area information, building space, and facility location and equipment location preference information contained in the project building information within the two-dimensional drawing software, and a mixed integer programming algorithm is used to optimize the spatial layout. Hard distance constraints are set to limit each facility to be completely located within the site, each equipment to be completely located within the corresponding building space, and facilities and equipment to not be allowed to overlap. Finally, the corresponding two-dimensional drawing is output.

5. The method for rapid generation of a three-dimensional model based on BIM according to claim 1, wherein: The generation process of the three-dimensional model includes: Detecting the professional areas included in the basic information of the project, implementing digital signature verification on the professional areas using a role-based access control mechanism, and generating the three-dimensional model; Locking the editable elements of the professional area and setting a first mark to highlight the position of the professional area; A second sign is provided to highlight pipelines that cannot be bypassed from the professional area.

6. The method for rapid generation of a three-dimensional model based on BIM according to claim 5, characterized in that: It also includes setting up an error correction mechanism, presetting one or more open spaces, and detecting whether there is a spatial topology conflict in the basic information of the project through a constraint propagation network. When the conflict degree is greater than a preset threshold, the error correction mechanism is triggered, and the functional areas with demand conflicts in the three-dimensional model are separated by the open space.

7. A BIM-based 3D model rapid generation system, characterized by: include: A natural language analysis module, used to obtain user requirement documents and extract language features from the user requirement documents; The language feature extraction process includes: preprocessing the user demand document, removing special symbols in the document, and performing word segmentation to obtain a preprocessed text; constructing an entity vocabulary in the architectural field, wherein the entity vocabulary adopts a dynamic update mechanism, incrementally learns unregistered words through a BiLSTM-CRF model, and uses context to perform entity disambiguation, and the entity vocabulary includes at least space type, material, and specification requirements; integrating the entity vocabulary into the pretrained language model feature space, achieving domain adaptation through transfer learning, and performing entity recognition on the preprocessed text to obtain the language features; A project construction information generation module, configured to generate project construction information based on a programming language by combining the language features with basic project information; A two-dimensional drawing generation module is used to call the drawing API interface of the two-dimensional drawing software to generate a two-dimensional drawing according to the project building information; The three-dimensional model generation module is used to input the two-dimensional drawings into the BIM system, associate and match the two-dimensional elements with the BIM family library through a spatial topological relationship mapping algorithm, and set the professional area in the basic information of the project as an uneditable constraint area to automatically generate a three-dimensional model that complies with the IFC standard.

8. A BIM-based 3D model rapid generation device, characterized in that: The device includes a memory, a processor and a communication module, wherein: The memory is used to store executable program code; The processor is coupled to the memory; The processor calls the executable program code stored in the memory to execute the BIM-based three-dimensional model rapid generation method according to any one of claims 1 to 6.

9. A computer storable medium, characterized in that The computer storable medium stores computer instructions, and when the computer instructions are called, they are used to execute the BIM-based three-dimensional model rapid generation method according to any one of claims 1 to 6.

Images