Construction atlas optimal scheme evaluation method and system based on large language model

Abstract

The invention discloses a construction atlas optimal scheme evaluation method and system based on a large language model. The method comprises the steps of construction atlas detection blocking, block atlas analysis, block atlas knowledge graph construction, block atlas process practice advantages and disadvantages analysis, block atlas process practice difference analysis and block atlas optimal construction scheme evaluation. According to the method, intelligent analysis of the optimal scheme of the construction atlas is carried out by using the large language model, field construction is guided, the evaluation efficiency and accuracy of determining the optimal scheme are improved, and the problem of low evaluation accuracy caused by omission of the construction atlas in manual evaluation is avoided.

Description

Technical Field

[0001] The present invention relates to the field of building construction drawing analysis, and in particular to a construction drawing optimal solution evaluation method and system based on a large language model. Background Art

[0002] In recent years, large language model technology has been widely used in finance, the Internet, and biomedicine, and has been applied and promoted in text data retrieval and knowledge question and answer in construction projects. However, at present, whether it is a general large language model or other professional large language models in the construction industry, the understanding of construction drawings is still relatively superficial, and it is difficult to read complex construction drawings and give professional answers. Since construction drawings are an important basis for guiding construction, understanding construction drawings is extremely important for on-site engineering and technical personnel. At this stage, the understanding of construction drawings still depends on the experience of engineering and technical personnel. When the atlas data is complex and there are many atlas versions, it is difficult for engineering and technical personnel to compare one by one in complex atlas data and many versions of atlases, so as to guide construction on site. Summary of the invention

[0003] The purpose of the present invention is to provide a method and system for evaluating the optimal solution of construction drawings based on a large language model, so as to solve the problem that it is difficult for engineering and technical personnel to comprehensively check the structure and process practices of each construction node from construction drawings with complex construction drawings, numerous atlas versions and large amount of atlas data to determine the optimal solution to guide on-site construction.

[0004] In order to solve the above technical problems, the present invention provides a construction drawing set optimal solution evaluation method based on a large language model, comprising:

[0005] Construction drawing detection and segmentation: The large language model is used to detect each unstructured construction drawing using a multi-scale page detection algorithm, extract the process structure construction schematic diagram, process annotation content, and the drawing name and drawing number to form a segmented drawing, and label the segmented drawing.

[0006] Block atlas analysis: Use text recognition technology to perform text recognition on the process annotation content and atlas name of each block atlas, and use a large language model to understand the process structure construction schematic diagram of each block atlas to extract key text information;

[0007] Constructing a knowledge graph of the block atlas: Reconstruct the text recognition content and key text information of each block atlas in the form of a structure tree to establish a knowledge graph of the block atlas content;

[0008] Analyze the advantages and disadvantages of the process practices in the block atlas: Use the large language model to compare and analyze the process annotation content in each knowledge graph to analyze the advantages and disadvantages of various process practices;

[0009] Analyze the differences in process practices of block atlases: Use the large language model to compare the process structure construction schematics of the same construction process in each knowledge graph in each block atlas, find out the differences between the two, and analyze the impact of the differences on the process practices;

[0010] Evaluate the optimal construction plan for a block atlas: The large language model generates the optimal implementation plan for the same block atlas based on the two dimensions of the advantages and disadvantages of the process practices and the impact of differences in process practices.

[0011] Furthermore, the large language model-based construction drawing optimal solution evaluation method provided by the present invention performs text recognition through optical character recognition technology.

[0012] Furthermore, the present invention provides a method for evaluating the optimal solution of construction drawings based on a large language model, and the block atlas knowledge graph constructed is the global content of the construction drawings, partial construction drawings of more than one construction node, and / or the related content of the same process practices in different versions of construction drawings.

[0013] Furthermore, the present invention provides a method for evaluating the optimal solution of construction drawings based on a large language model. The large language model understands the meaning of the construction drawings by using the block atlas knowledge graph to determine the name, atlas number, and annotation content of the construction drawings, thereby giving a graphic and text-based answer and explanation.

[0014] In order to solve the above technical problems, the present invention also provides a construction drawing set optimal solution evaluation system based on a large language model, comprising:

[0015] The construction atlas parsing unit includes a construction atlas detection block sub-unit, a block atlas parsing sub-unit, and a block atlas knowledge graph building sub-unit, wherein:

[0016] The construction atlas detection block subunit is used to detect each unstructured construction atlas using a large language model and a multi-scale page detection algorithm, extract the process structure construction schematic diagram, process annotation content, atlas name, atlas number and other contents to form a block atlas, and assign labels to the block atlas;

[0017] The block atlas parsing subunit is used to perform text recognition on the process annotation content and atlas name of each block atlas through text recognition technology, and to understand the process structure construction schematic diagram of each block atlas through a large language model to extract key text information;

[0018] Construct a block atlas knowledge graph subunit, which is used to reconstruct the text recognition content and key text information of each block atlas in the form of a structure tree to establish a knowledge graph of the block atlas content;

[0019] The large language model processing unit of the construction drawing set includes a sub-unit for analyzing the advantages and disadvantages of the process practices of the block drawing set, a sub-unit for analyzing the differences in the process practices of the block drawing set, and a sub-unit for evaluating the optimal construction plan of the block drawing set, among which:

[0020] Analyze the advantages and disadvantages of process practices in the block atlas sub-unit, which is used to compare and analyze the process annotation content in each knowledge graph through a large language model to analyze the advantages and disadvantages of various process practices;

[0021] The sub-unit for analyzing differences in process practices in block atlases is used to compare the process structure construction schematic diagrams of the same construction process in each knowledge graph through a large language model, find out the differences between the two, and analyze the impact of the differences on the process practices;

[0022] The sub-unit for evaluating the optimal construction plan of the block atlas is used for the large language model to generate the optimal implementation plan for the same block atlas based on the two dimensions of the advantages and disadvantages of the process practices and the impact of differences in process practices.

[0023] Compared with the prior art, the present invention has the following beneficial effects:

[0024] The method and system for evaluating the optimal solution of construction drawings based on a large language model provided by the present invention utilize construction drawing detection, segmentation, analysis, and knowledge graph construction to convert the structured characteristics of the construction drawings into unstructured text content to form a language and format that can be understood by the large language model. The large language model is used to analyze the advantages and disadvantages of process practices and the differences in process practices of the text content of the unstructured drawings to understand and evaluate the optimal construction solution of the construction drawings, thereby realizing intelligent analysis of the optimal solution of the construction drawings for guiding on-site construction, improving the evaluation efficiency and accuracy of determining the optimal solution, and avoiding the problem of low evaluation accuracy caused by omissions of construction drawings in manual evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a flowchart of the optimal solution evaluation method for construction drawings based on the large language model;

[0026] Figure 2 It is a structural composition relationship diagram of the construction drawing set optimal solution evaluation system based on the large language model;

[0027] As shown in the figure:

[0028] 1. Construction drawing parsing unit, 11. Construction drawing detection block sub-unit, 12. Block atlas parsing sub-unit, 13. Block atlas knowledge graph construction sub-unit; 2. Construction drawing large language model processing unit, 21. Block atlas process pros and cons analysis sub-unit, 22. Block atlas process difference analysis sub-unit, 23. Block atlas optimal construction plan evaluation sub-unit. DETAILED DESCRIPTION

[0029] The present invention is described in detail below in conjunction with the accompanying drawings: The advantages and features of the present invention will become more apparent from the following description. It should be noted that the accompanying drawings are in very simplified form and in non-precise proportions, and are only used to conveniently and clearly assist in explaining the purpose of the embodiments of the present invention.

[0030] Please refer to Figure 1 , an embodiment of the present invention provides a method for evaluating the optimal solution of a construction drawing set based on a large language model, which may include:

[0031] Step S101, construction atlas detection and segmentation: the unstructured construction atlases are detected by using a large language model and a multi-scale page detection algorithm, and the process structure construction schematic diagram, process annotation content, atlas name, atlas number and other contents are extracted to form segmented atlases, and labels are assigned to the segmented atlases.

[0032] Step S102, block atlas analysis: text recognition is performed on the process annotation content and atlas name of each block atlas through text recognition technology, and the process structure construction schematic diagram of each block atlas is understood by the large language model to extract key text information. Specifically, the block atlas can be understood by the graphic understanding model in the large language model. In order to perform text recognition, text recognition can be performed by optical character recognition (OCR) technology.

[0033] Step S103, constructing a knowledge graph of the block atlas: reconstructing the text recognition content and key text information of each block atlas in the form of a structure tree to establish a knowledge graph of the block atlas content. The constructed block atlas knowledge graph may be the global content of the construction atlas, partial construction atlases of more than one construction node, and / or the associated content of the same process in different versions of construction atlases.

[0034] Step S104, analyzing the advantages and disadvantages of the process practices in the block atlas: using a large language model to compare and analyze the process annotation content in each knowledge graph, and analyzing the advantages and disadvantages of each type of process practice.

[0035] Step S105, analyzing the differences in process practices in the block atlas: using a large language model to compare the process structure construction schematics of the same construction process in each knowledge graph in each block atlas, find out the differences between the two and analyze the impact of the differences on the process practices.

[0036] In order to improve the understanding of the construction drawings, in step S104 and step S105, the large language model uses the block atlas knowledge graph to understand the meaning of the construction drawings by the name, atlas number, and annotation content of the atlas, and then gives a picture-text answer and explanation. The answer and explanation include the advantages and disadvantages of the process, the differences in process and their impact.

[0037] Step S106, evaluate the optimal construction plan for the block atlas: the large language model generates the optimal implementation plan for the same block atlas based on the advantages and disadvantages of the process and the impact of the differences in the process. After analyzing the advantages and disadvantages of the process, the advantages are selected and the disadvantages are avoided; after analyzing the differences and impacts of the process, the favorable effects are selected and the unfavorable effects are avoided.

[0038] Please refer to Figure 2 The embodiment of the present invention further provides a construction drawing set optimal solution evaluation system based on a large language model, which adopts the method of the above embodiment, including a construction drawing set parsing unit 1 and a construction drawing set large language model processing unit 2, wherein:

[0039] The construction atlas parsing unit 1 includes a construction atlas detection block subunit 11, a block atlas parsing subunit 12, and a block atlas knowledge graph building subunit 13, wherein:

[0040] The construction atlas detection block subunit 11 is used to detect each unstructured construction atlas using a multi-scale page detection algorithm through a large language model, extract the process structure construction schematic diagram, process annotation content, atlas name, atlas number and other contents to form a block atlas, and assign labels to the block atlas.

[0041] The block atlas parsing subunit 12 is used to perform text recognition on the process annotation content and atlas name of each block atlas through text recognition technology, and to extract key text information by understanding the process structure construction schematic diagram of each block atlas through a large language model.

[0042] The block atlas knowledge graph construction subunit 13 is used to reconstruct the text recognition content and key text information of each block atlas in the form of a structure tree to establish a knowledge graph of the block atlas content.

[0043] The construction atlas large language model processing unit 2 includes a subunit 21 for analyzing the advantages and disadvantages of the process practices of the block atlas, a subunit 22 for analyzing the differences in the process practices of the block atlas, and a subunit 23 for evaluating the optimal construction plan of the block atlas, wherein:

[0044] The sub-unit 21 for analyzing the advantages and disadvantages of process practices in the block atlas is used to compare and analyze the process annotation contents in each knowledge graph through a large language model, and analyze the advantages and disadvantages of various process practices.

[0045] The sub-unit 22 for analyzing differences in process practices in the block atlas is used to compare the process structure construction schematic diagrams of each block atlas for the same construction process in each knowledge graph through a large language model, find out the differences between the two and analyze the impact of the differences on the process practices.

[0046] The sub-unit 23 for evaluating the optimal construction plan for the block atlas is used for the large language model to generate the optimal implementation plan for the same block atlas based on the two dimensions of the advantages and disadvantages of the process practices and the impact of the differences in the process practices.

[0047] The method and system for evaluating the optimal solution of construction drawings based on a large language model provided in an embodiment of the present invention utilize construction drawing detection, segmentation, analysis, and knowledge graph construction to convert the structured characteristics of the construction drawings into unstructured text content to form a language and format that can be understood by the large language model. The large language model is used to analyze the advantages and disadvantages of process practices and the differences in process practices of the text content of the unstructured drawings to understand and evaluate the optimal construction solution of the construction drawings, thereby realizing intelligent analysis of the optimal solution of the construction drawings for guiding on-site construction, improving the evaluation efficiency and accuracy of determining the optimal solution, and avoiding the problem of low evaluation accuracy caused by omissions of construction drawings in manual evaluation.

[0048] The present invention is not limited to the above-mentioned specific implementation modes. Obviously, the above-mentioned embodiments are only some embodiments of the embodiments of the present invention, but not all embodiments. Based on the described embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention. Those skilled in the art can make other levels of modifications and changes to the present invention. In this way, if these modifications and changes of the present invention fall within the scope of the claims of the present invention, the present invention is also intended to include these changes and changes.

Claims

1. A method for evaluating the optimal solution of construction drawings based on a large language model, characterized in that: include: Construction drawing detection and segmentation: The large language model is used to detect each unstructured construction drawing using a multi-scale page detection algorithm, extract the process structure construction schematic diagram, process annotation content, and the drawing name and drawing number to form a segmented drawing, and label the segmented drawing. Block atlas analysis: Use text recognition technology to perform text recognition on the process annotation content and atlas name of each block atlas, and use a large language model to understand the process structure construction schematic diagram of each block atlas to extract key text information; Constructing a knowledge graph of the block atlas: Reconstruct the text recognition content and key text information of each block atlas in the form of a structure tree to establish a knowledge graph of the block atlas content; Analyze the advantages and disadvantages of the process practices in the block atlas: Use the large language model to compare and analyze the process annotation content in each knowledge graph to analyze the advantages and disadvantages of various process practices; Analyze the differences in process practices of block atlases: Use the large language model to compare the process structure construction schematics of the same construction process in each knowledge graph in each block atlas, find out the differences between the two, and analyze the impact of the differences on the process practices; Evaluate the optimal construction plan for a block atlas: The large language model generates the optimal implementation plan for the same block atlas based on the two dimensions of the advantages and disadvantages of the process practices and the impact of differences in process practices.

2. The method for evaluating the optimal solution of construction drawings based on a large language model according to claim 1 is characterized in that: Text recognition is performed through optical character recognition technology.

3. The method for evaluating the optimal solution of construction drawings based on a large language model according to claim 1 is characterized in that: The constructed block atlas knowledge graph is the global content of the construction atlas, partial construction atlas of more than one construction node and / or the related content of the same process practices in different versions of construction atlas.

4. The method for evaluating the optimal solution of construction drawings based on a large language model according to claim 1 is characterized in that: The large language model understands the meaning of the construction drawing by using the block atlas knowledge graph to understand the name, atlas number, and annotation content of the construction drawing, and thus gives answers and explanations that combine pictures and text.

5. A construction drawing set optimal solution evaluation system based on a large language model, characterized in that: include: The construction atlas parsing unit includes a construction atlas detection block sub-unit, a block atlas parsing sub-unit, and a block atlas knowledge graph building sub-unit, wherein: The construction atlas detection block subunit is used to detect each unstructured construction atlas using a large language model and a multi-scale page detection algorithm, extract the process structure construction schematic diagram, process annotation content, atlas name, atlas number and other contents to form a block atlas, and assign labels to the block atlas; The block atlas parsing subunit is used to perform text recognition on the process annotation content and atlas name of each block atlas through text recognition technology, and to understand the process structure construction schematic diagram of each block atlas through a large language model to extract key text information; Construct a block atlas knowledge graph subunit, which is used to reconstruct the text recognition content and key text information of each block atlas in the form of a structure tree to establish a knowledge graph of the block atlas content; The large language model processing unit of the construction drawing set includes a sub-unit for analyzing the advantages and disadvantages of the process practices of the block drawing set, a sub-unit for analyzing the differences in the process practices of the block drawing set, and a sub-unit for evaluating the optimal construction plan of the block drawing set, among which: Analyze the advantages and disadvantages of process practices in the block atlas sub-unit, which is used to compare and analyze the process annotation content in each knowledge graph through a large language model to analyze the advantages and disadvantages of various process practices; The sub-unit for analyzing differences in process practices in block atlases is used to compare the process structure construction schematic diagrams of the same construction process in each knowledge graph through a large language model, find out the differences between the two, and analyze the impact of the differences on the process practices; The sub-unit for evaluating the optimal construction plan of the block atlas is used for the large language model to generate the optimal implementation plan for the same block atlas based on the two dimensions of the advantages and disadvantages of the process practices and the impact of differences in process practices.

Images