Material attribute acquisition method and system based on large language model and BIM

Abstract

The invention discloses a material attribute acquisition method and system based on a large language model and a BIM, and relates to the technical field of digital intelligent buildings. The method comprises the steps that building model information is acquired from a BIM platform, and the building model information comprises component types, component names, component material names, design styles and color preferences; selecting a matched format template according to the building model information, and generating a cue word text containing key information; the cue word text is analyzed and mapped based on a large language model, detailed attribute description about component materials is obtained, and the detailed attribute description comprises RGB color values, transparency, roughness, metallicity and specular reflection intensity; the detailed attribute description of the component material is applied to the model display of the BIM platform, and the corresponding display effect picture is previewed, generated and exported, so that the working efficiency and quality of endowing the building model component with the material are improved, and the flexibility and automation degree of the design process are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of digital intelligent buildings, and in particular to a material attribute acquisition method and system based on a large language model and BIM. Background Art

[0002] Building Information Modeling (BIM) is a method of establishing a building model based on the relevant information data of a construction project. By simulating the real information of a building through digital information, it can realize the information integration and sharing among different participants and the interconnection and interoperability of data in various links, thereby improving the work efficiency of the construction industry. BIM has two characteristics: one is the model and the other is the data. From the model level, BIM is a digital mirror of a physical building, which helps to create a virtual space equivalent to the physical space to achieve a more realistic and immersive virtual reality experience; from the data level, BIM is a basic data asset, the core of realizing the industrialization of construction and smart city business, and will play an increasingly important role with the development of social digitalization. Among them, the assignment of material information needs to consider various factors such as the style of the building, the style of the scene, and the real material of the components.

[0003] The workflows of existing BIM platforms can be roughly divided into three types: one is that in the conventional design process, architectural designers manually select and assign materials, which usually takes a lot of time and effort and is prone to errors, especially in complex building models; the second is to use a preset material library, from which designers can select pre-defined materials and apply them to the building model. However, this method is limited by the material library of specific software and platforms, often cannot meet the needs of specific projects, and lacks flexibility; the third is to achieve the desired effect by manually adjusting material parameters and BIM platform settings. However, this requires designers to have a high level of technical skills, and there are still certain limitations, such as difficulty in achieving the effect of complex materials or the need for continuous debugging and modification. Therefore, the existing workflows generally have problems such as low efficiency, duplication of work, reliance on design experience, lack of quality assurance, and single considerations. Summary of the invention

[0004] The present invention provides a material attribute acquisition method and system based on a large language model and BIM, aiming to improve the work efficiency and quality of assigning materials to building model components and increase the flexibility and automation of the design process.

[0005] In order to achieve the above-mentioned object of the invention, the embodiment of the present invention provides the following technical solutions: In a first aspect, an embodiment of the present invention provides a material attribute acquisition method based on a large language model and BIM, comprising the following steps: S100: Acquire building model information from the BIM platform, where the building model information includes component category, component name, component material name, design style, and color preference; S200: selecting a matching format template according to the building model information, and generating a prompt word text containing key information; S300: parsing and mapping the prompt word text based on the large language model to obtain a detailed attribute description of the component material, wherein the detailed attribute description includes RGB color value, transparency, roughness, metallicity, and specular reflection intensity; S400: Apply the detailed property description of the component material to the model display of the BIM platform, and preview, generate and export the corresponding display effect diagram.

[0006] Furthermore, the step S300 specifically includes: The detailed property description obtained from the large language model is applied to the corresponding components in the building model, and the detailed property description of the component material is updated in real time accordingly.

[0007] Furthermore, after step S400, the method further includes: The prompt word text of the large language model is adjusted or the generated detailed attribute description is modified and applied again to the model display of the BIM platform to preview, generate and export the corresponding display effect diagram in real time.

[0008] Furthermore, before step S100, the method further includes: Classify, encode and integrate the building model information to obtain a structured coding data set; The large language model is trained according to the structured coding data set to establish information mapping between the prompt word text and the detailed attribute description of the component material.

[0009] In a second aspect, an embodiment of the present invention provides a material attribute acquisition system based on a large language model and BIM, comprising: An acquisition module is used to acquire building model information from the BIM platform, wherein the building model information includes component category, component name, component material name, design style and color preference; A generation module, used for selecting a matching format template according to the building model information, and generating a prompt word text containing key information; A parsing and mapping module, used to parse and map the prompt word text based on a large language model to obtain a detailed attribute description of the component material, wherein the detailed attribute description includes RGB color value, transparency, roughness, metallicity, and specular reflection intensity; The display module is used to apply the detailed property description of the component material to the model display of the BIM platform, and preview, generate and export the corresponding display effect diagram.

[0010] Furthermore, the parsing and mapping module is specifically used to apply the detailed attribute description obtained from the large language model to the corresponding components in the building model, and update the detailed attribute description of the component material in real time accordingly.

[0011] Furthermore, the system further comprises: The modification and adjustment module is used to adjust the prompt word text of the large language model or modify the generated detailed attribute description, which is applied again to the model display of the BIM platform to preview, generate and export the corresponding display effect diagram in real time.

[0012] Furthermore, the system further comprises: The coding integration module is used to classify, encode and integrate the building model information to obtain a structured coding data set; The training module is used to train the large language model according to the structured coding data set to establish information mapping between the prompt word text and the detailed attribute description of the component material.

[0013] Compared with the prior art, the present invention has the following beneficial effects: The present invention first obtains building model information from the BIM platform, selects a matching format template, automatically generates prompt word text containing key information, and then parses and maps the prompt word text based on a large language model to obtain a detailed attribute description of the component material, thereby eliminating the need for designers to manually select and adjust, saving a lot of time and reducing the possibility of human error. Finally, the detailed attribute description of the component material is applied to the model display of the BIM platform to preview, generate and export the corresponding display effect diagram.

[0014] In addition, the present invention can also adjust the prompt word text of the large language model or modify the generated detailed attribute description, thereby providing more flexibility, and designers can quickly obtain suitable material attributes according to the needs of specific projects without being restricted by the preset material library. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0016] Figure 1 A flow chart of a method for obtaining material properties based on a large language model and BIM provided in an embodiment of the present invention; Figure 2 A schematic diagram of functional modules of a material attribute acquisition system based on a large language model and BIM provided in an embodiment of the present invention.

[0017] Reference numerals in the figures: Acquisition module 100; Generation module 200; Parsing and mapping module 300; Display module 400. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present invention.

[0019] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance, or implying any such actual relationship or order between these entities or operations.

[0020] like Figure 1 As shown, a material attribute acquisition method based on a large language model and BIM provided by an embodiment of the present invention may include the following steps: Step S100: Acquire building model information from the BIM platform.

[0021] The building model information mainly includes component category, component name, component material name, design style and color preference.

[0022] In one embodiment, before step S100, the method further includes: Classify, encode and integrate the building model information to obtain a structured coding data set; The large language model is trained according to the structured coding data set to establish information mapping between the prompt word text and the detailed attribute description of the component material.

[0023] Specifically, we first use the Autodesk platform application interface Revit API to extract component names and their attributes for automatic encoding, and then perform structural integration based on the encoded information, including the organizational structure of nodes, the extraction of nodes at all levels and their attributes, the unique identification and alignment of nodes, the relationship connection between nodes at all levels and the extraction of relationship attributes, and use entity recognition and relationship extraction algorithms to obtain a structured encoding dataset. Then, the structured encoding dataset is input into the large language model for training, and the information mapping between the prompt word text and the BIM components and their attributes is established, and output in JSON format.

[0024] Relation extraction refers to extracting the relationship between entities from text, such as the category relationship between components, the matching relationship between design style and color preference, etc. Through entity recognition and relationship extraction, the semantically parsed data set can be converted into a structured data set.

[0025] Among them, the Large Language Model (LLM) is a natural language processing model based on artificial intelligence technology. Through deep learning algorithms and training of large-scale text data, it can simulate human language understanding and generation capabilities, and achieve efficient processing and analysis of natural language. Compared with traditional language models, LLM usually adopts more advanced deep learning technologies and algorithms, such as Transformer architecture, attention mechanism, etc. These technologies can help the model better capture the semantic and contextual information in natural language, improve the accuracy and generalization ability of the model; in addition, large language models usually have larger scale and stronger computing power. The large language model is preferably a GPT model.

[0026] Step S200: selecting a matching format template according to the building model information, and generating a prompt word text containing key information.

[0027] In one implementation, after the style words are input, they are combined with the acquired architectural model information to form a unified prompt word text, and subsequent operations are performed based on the big data model.

[0028] Preferably, a SQL parser is used to obtain a detailed property description about the component material from the structured coding data set.

[0029] Step S300: parsing and mapping the prompt word text based on a large language model to obtain a detailed property description of the component material.

[0030] The detailed attribute description includes RGB color value, transparency, roughness, metallicity and specular reflection intensity.

[0031] In one implementation, the step S300 specifically includes: The detailed property description obtained from the large language model is applied to the corresponding components in the building model, and the detailed property description of the component material is updated in real time accordingly.

[0032] Step S400: Apply the detailed property description of the component material to the model display of the BIM platform, and preview, generate and export the corresponding display effect diagram.

[0033] In one embodiment, after step S400, the method further includes: The prompt word text of the large language model is adjusted or the generated detailed attribute description is modified and applied again to the model display of the BIM platform, and the corresponding display renderings are previewed, generated and exported in real time, thereby improving the visual effect and realism of the building model.

[0034] As shown in Table 1, as a specific implementation of an embodiment of the present invention, the BIM platform is preferably an FBIM platform.

[0035] Component name (family, type, category) style Material Name Color Preference RGB returned by the large model Basic Wall - Masonry Brick Antique style - - (120, 72, 64) Curtain Wall Antique style Glass - (112, 147, 143) Floor Slabs - Visual Decorative Decorative Floor Slabs Antique style Aluminum - (193, 154, 107) Curtain wall vertical Antique style Aluminum - (184, 134, 11) Basic wall-column decoration Antique style Wood material - (150, 111, 51) Conventional floor Antique style - - (255, 235, 205) Rectangular beam Antique style Made of steel plate - (136, 69, 53) Shutters Antique style Solid Wood - (139, 69, 19) sofa Antique style - - (139, 0, 0) Sofa legs Antique style - - (184, 134, 11) Air duct Antique style - Light (205, 175, 149) Outdoor unit Antique style - - (105, 105, 105) Weather Monitor Antique style - - (184, 134, 11) Basic wall-shear wall Antique style Concrete - (210, 180, 140) Double leaf door, single leaf door Antique style Woody - (101, 67, 33) Green plants Antique style - - (60, 179, 113) soil Antique style - - (101, 67, 33) flowerpot Antique style - antiquity (102, 93, 30) Specifically, the material attribute acquisition method may include the following steps: Step 1: Extract the category and name of each component, material name and other information from multiple BIM models stored in the FBIM platform, and obtain the design style and color preference; Step 2: Call the large language model to generate corresponding color attributes, where the color attributes are RGB color values; Step 3: Use the API of the FBIM platform to apply the generated color attributes to the corresponding components, view the effect preview in FBIM, make detailed adjustments as needed, and generate the final display effect diagram.

[0036] It should be noted that the BIM platform provided in this embodiment may also be a BIM lightweight graphic platform such as the BIMFACE platform and the EveryBIM platform.

[0037] This embodiment first obtains building model information from the BIM platform, selects a matching format template, automatically generates prompt word text containing key information, and then parses and maps the prompt word text based on the large language model to obtain detailed attribute descriptions of component materials, thereby eliminating the need for designers to manually select and adjust, saving a lot of time and reducing the possibility of human errors. Finally, the detailed attribute description of the component material is applied to the model display of the BIM platform, and the corresponding display effect diagram is previewed, generated, and exported.

[0038] like Figure 2 As shown, a material attribute acquisition system based on a large language model and BIM is provided in an embodiment of the present invention, which is used to implement the material attribute acquisition method based on a large language model and BIM. The material attribute acquisition system may include an acquisition module 100, a generation module 200, a parsing and mapping module 300, and a display module 400.

[0039] In this embodiment, the acquisition module 100 is used to acquire building model information from the BIM platform, wherein the building model information includes component category, component name, component material name, design style and color preference.

[0040] The generating module 200 is used to select a matching format template according to the building model information and generate a prompt word text containing key information.

[0041] The parsing and mapping module 300 is used to parse and map the prompt word text based on the large language model to obtain a detailed attribute description of the component material, wherein the detailed attribute description includes RGB color value, transparency, roughness, metallicity, and specular reflection intensity.

[0042] The display module 400 is used to apply the detailed property description of the component material to the model display of the BIM platform, and preview, generate and export the corresponding display effect diagram.

[0043] In one implementation, the parsing and mapping module 300 is specifically configured to apply the detailed property description obtained from the large language model to the corresponding components in the building model, and correspondingly update the detailed property description of the component material in real time.

[0044] Furthermore, the system also includes a modification and adjustment module, which is used to adjust the prompt word text of the large language model or modify the generated detailed attribute description, and apply it again to the model display of the BIM platform to preview, generate and export the corresponding display effect diagram in real time.

[0045] Furthermore, the system also includes a coding integration module and a training module. The coding integration module is used to classify, encode and integrate the building model information to obtain a structured coding data set.

[0046] The training module is used to train the large language model according to the structured coding data set to establish information mapping between the prompt word text and the detailed attribute description of the component material.

[0047] In summary, the embodiment of the present invention first obtains the building model information from the BIM platform, selects a matching format template, automatically generates a prompt word text containing key information, and then parses and maps the prompt word text based on the large language model to obtain a detailed attribute description of the component material, thereby eliminating the need for designers to manually select and adjust, saving a lot of time and reducing the possibility of human error. Finally, the detailed attribute description of the component material is applied to the model display of the BIM platform, and the corresponding display effect diagram is previewed, generated, and exported.

[0048] In addition, the embodiment of the present invention can also adjust the prompt word text of the large language model or modify the generated detailed attribute description, thereby providing more flexibility. Designers can quickly obtain suitable material attributes according to the needs of specific projects without being restricted by the preset material library.

[0049] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A material attribute acquisition method based on a large language model and BIM, characterized in that: The following steps are involved: S100: Acquire building model information from the BIM platform, where the building model information includes component category, component name, component material name, design style, and color preference; S200: selecting a matching format template according to the building model information, and generating a prompt word text containing key information; S300: parsing and mapping the prompt word text based on the large language model to obtain a detailed attribute description of the component material, wherein the detailed attribute description includes RGB color value, transparency, roughness, metallicity, and specular reflection intensity; S400: Apply the detailed property description of the component material to the model display of the BIM platform, and preview, generate and export the corresponding display effect diagram.

2. The material attribute acquisition method based on a large language model and BIM according to claim 1, characterized in that: The step S300 specifically includes: The detailed property description obtained from the large language model is applied to the corresponding components in the building model, and the detailed property description of the component material is updated in real time accordingly.

3. The material attribute acquisition method based on a large language model and BIM according to claim 1, characterized in that: After step S400, the method further includes: The prompt word text of the large language model is adjusted or the generated detailed attribute description is modified and applied again to the model display of the BIM platform to preview, generate and export the corresponding display effect diagram in real time.

4. The material attribute acquisition method based on a large language model and BIM according to claim 1, characterized in that: Before step S100, the method further includes: Classify, encode and integrate the building model information to obtain a structured coding data set; The large language model is trained according to the structured coding data set to establish information mapping between the prompt word text and the detailed attribute description of the component material.

5. A material attribute acquisition system based on a large language model and BIM, characterized in that: The system comprises: An acquisition module is used to acquire building model information from the BIM platform, wherein the building model information includes component category, component name, component material name, design style and color preference; A generation module, used for selecting a matching format template according to the building model information, and generating a prompt word text containing key information; A parsing and mapping module, used to parse and map the prompt word text based on a large language model to obtain a detailed attribute description of the component material, wherein the detailed attribute description includes RGB color value, transparency, roughness, metallicity, and specular reflection intensity; The display module is used to apply the detailed property description of the component material to the model display of the BIM platform, and preview, generate and export the corresponding display effect diagram.

6. The material attribute acquisition system based on a large language model and BIM according to claim 5 is characterized in that: The parsing and mapping module is specifically used to apply the detailed attribute description obtained from the large language model to the corresponding components in the building model, and to update the detailed attribute description of the component material in real time accordingly.

7. The material attribute acquisition system based on a large language model and BIM according to claim 5, characterized in that: The system further comprises: The modification and adjustment module is used to adjust the prompt word text of the large language model or modify the generated detailed attribute description, which is applied again to the model display of the BIM platform to preview, generate and export the corresponding display effect diagram in real time.

8. The material attribute acquisition system based on a large language model and BIM according to claim 5, characterized in that: The system further comprises: The coding integration module is used to classify, encode and integrate the building model information to obtain a structured coding data set; The training module is used to train the large language model according to the structured coding data set to establish information mapping between the prompt word text and the detailed attribute description of the component material.

Images