A vector drawing data annotation method, system, and terminal based on GAI

By using a GAI-based vector drawing data annotation method, the problem of low annotation efficiency in architectural design has been solved, achieving high-quality, accurate automated annotation and improving design efficiency and intelligence.

CN120997864BActive Publication Date: 2026-11-10SHENZHEN CAPOL INT & ASSOC CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202511084671.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-08-04
Publication Date
2026-11-10
Estimated Expiration
2045-08-04

AI Technical Summary

Technical Problem

Existing technologies in architectural design are inefficient and prone to errors in marking building plan dimensions, making it difficult to meet the needs of efficient and accurate design. Furthermore, multi-disciplinary collaboration leads to low data production efficiency and hinders rapid iteration.

Method used

A vector drawing data annotation method based on GAI is adopted. The original architectural floor plan is preprocessed, structural components are annotated using the GAI model, and the annotation results are optimized to achieve automated annotation.

Benefits of technology

It has achieved high-quality and accurate annotation of architectural floor plans, improved annotation efficiency, and realized the intelligent upgrade of architectural design and analysis.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120997864B_ABST
    Figure CN120997864B_ABST
Patent Text Reader

Abstract

This invention discloses a method, system, and terminal for annotating vector drawing data based on GAI (Graphical Image Processing). The method includes: acquiring original architectural floor plans; preprocessing the original architectural floor plans to obtain a high-quality base map; acquiring a GAI model; using the GAI model to annotate the structural components in the high-quality base map according to annotation rules; and performing data optimization processing on the annotated high-quality base map to obtain an optimized architectural floor plan. This invention transforms architectural floor plans into recognizable, high-quality, and accurate annotated data, achieving automated annotation of architectural floor plans, data optimization of architectural floor plans, and intelligent upgrading of architectural design and analysis.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a vector graphics data annotation method, system, and terminal based on GAI. Background Technology

[0002] In the architectural design industry, architectural plan dimensioning serves as a bridge connecting design concepts and engineering practice. Traditional methods rely on manual dimensioning using CAD software. While the software offers some automatic dimensioning functions, complex images or specific dimensioning requirements still necessitate manual adjustments. This approach is inefficient, error-prone, and fails to meet the demands of efficient and precise design. Furthermore, because dimensioning rules must consider multiple professional standards, including architectural, structural, and mechanical engineering, and rely on collaboration among experts from various fields, current dimensioning methods result in low data production efficiency, making it difficult to meet the needs of rapidly iterating projects.

[0003] Therefore, existing technologies still have shortcomings. Summary of the Invention

[0004] The technical problem to be solved by this invention is to provide a method, system, and terminal for vector paper data annotation based on GAI, addressing the aforementioned deficiencies of the prior art. The technical solution adopted by this invention is as follows:

[0005] In a first aspect, the present invention provides a vector drawing data annotation method based on GAI, wherein the method includes:

[0006] Obtain the original building floor plan, and preprocess the original building floor plan to obtain a high-quality base map;

[0007] Obtain the GAI model, and use the GAI model to annotate the structured components in the high-quality base map according to the annotation rules;

[0008] The high-quality base map after annotation is subjected to data optimization processing to obtain the optimized architectural plan drawing, wherein the optimized architectural plan drawing includes the optimized architectural plan base map and the internal dimension annotation location map.

[0009] In one implementation, the preprocessing of the original architectural floor plan to obtain a high-quality base map includes:

[0010] Using a recursive decomposition algorithm, the multiple nested blocks in the original architectural plan are deeply decomposed and transformed into basic geometric elements;

[0011] Locate components and redundant information that are not related to the annotations, and delete them.

[0012] In one implementation, the preprocessing of the original architectural floor plan to obtain a high-quality base map includes:

[0013] Vector data was extracted from the original architectural floor plan based on a CAD parsing library.

[0014] The scale information is extracted from the original architectural floor plan using OCR technology, and the scale information is used for size conversion.

[0015] In one implementation, the step of using the GAI model to annotate the structured components in the high-quality base map according to annotation rules includes:

[0016] Based on the GAI model, the structured components in the high-quality base map are automatically identified, and the positions of the annotation lines are automatically determined.

[0017] Based on the annotation rules and the determined annotation line positions, the annotation process of the structured component is automatically completed. The annotation rules are used to mark the positioning dimensions of the key features of the structured component.

[0018] In one implementation, the training method of the GAI model includes:

[0019] After preprocessing and annotation of several drawings corresponding to several projects, a dataset is obtained;

[0020] The dataset is output in three different resolutions and divided into a training set, a validation set, and a test set, each of which includes an equal number of samples for each resolution.

[0021] The GAI model is obtained by training a preset neural network model based on the training set.

[0022] In one implementation, the step of performing data optimization processing on the high-quality base map after annotation to obtain optimized architectural floor plans includes:

[0023] Set the fill color and edge color of the structured components in the high-quality base map after annotation to black or gray, and set the background to white to complete the data optimization process.

[0024] The high-quality base map after data optimization is exported to obtain the optimized building plan drawing.

[0025] In one implementation, the method further includes:

[0026] A visual editing window is provided during the annotation process, and the annotation results are identified based on a preset rule recognition model;

[0027] If there are errors in the identified and labeled data, the labeling results can be corrected by dragging or adjusting the control points.

[0028] Secondly, embodiments of the present invention also provide a vector paper data annotation system based on GAI, wherein the system is used to implement the steps of the vector paper data annotation method based on GAI as described in any of the above solutions, and the system includes:

[0029] The preprocessing module is used to acquire the original building floor plan drawings, preprocess the original building floor plan drawings, and obtain a high-quality base map.

[0030] An automatic annotation module is used to obtain a GAI model and use the GAI model to annotate the structured components in the high-quality base map according to annotation rules;

[0031] The data optimization module is used to optimize the high-quality base map after annotation processing to obtain optimized architectural plan drawings, wherein the optimized architectural plan base map and the internal dimension annotation location map are included.

[0032] Thirdly, embodiments of the present invention also provide a terminal, wherein the terminal includes a memory, a processor, and a GAI-based vector paper data annotation program stored in the memory and executable on the processor. When the processor executes the GAI-based vector paper data annotation program, it implements the steps of the GAI-based vector paper data annotation method of any of the above-described solutions.

[0033] Fourthly, embodiments of the present invention also provide a computer-readable storage medium, wherein the computer-readable storage medium stores a vector paper data annotation program based on GAI, and the vector paper data annotation program based on GAI implements the steps of the vector paper data annotation method based on GAI as described in any of the above schemes on the computer-readable storage medium.

[0034] Beneficial Effects: Compared with existing technologies, this invention provides a vector drawing data annotation method based on GAI. First, the invention acquires the original architectural floor plan and preprocesses it to obtain a high-quality base map. Then, it acquires a GAI model and uses the GAI model to annotate the structured components in the high-quality base map according to annotation rules. Finally, it performs data optimization processing on the annotated high-quality base map to obtain an optimized architectural floor plan. This invention automatically annotates architectural floor plans using a GAI model, thereby transforming architectural floor plans into recognizable, high-quality, and accurate annotated data. This achieves automated annotation and data optimization of architectural floor plans, enabling an intelligent upgrade of architectural design and analysis. Attached Figure Description

[0035] Figure 1 A flowchart of a preferred embodiment of the vector paper data annotation method based on GAI provided in this invention.

[0036] Figure 2 This is a schematic diagram of the architecture of a vector paper data annotation system based on GAI provided in an embodiment of the present invention.

[0037] Figure 3 A schematic diagram of a terminal provided in an embodiment of the present invention. Detailed Implementation

[0038] To make the objectives, technical solutions, and effects of this invention clearer and more explicit, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0039] The flowchart shown in the attached diagram is for illustrative purposes only and does not necessarily include all content, operations, or steps, nor does it require execution in the described order. For example, some operations or steps can be broken down, combined, or partially merged, so the actual execution order may change depending on the actual situation.

[0040] It should be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0041] It should be understood that, in order to clearly describe the technical solutions of the embodiments of the present invention, the terms "first" and "second" are used in the embodiments of the present invention to distinguish identical or similar items with essentially the same function and effect. For example, "first control information" and "second control information" are only used to distinguish different control information and do not limit their order.

[0042] Those skilled in the art will understand that the words "first" and "second" do not limit the quantity or the order of execution, and that the words "first" and "second" do not necessarily imply that they are different.

[0043] It should also be understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0044] To address the problems of existing technologies, this invention provides a vector drawing data annotation method based on GAI (Generative Artificial Intelligence). This method enables automated annotation of architectural floor plans. In specific applications, this embodiment first acquires the original architectural floor plan and preprocesses it to obtain a high-quality base map. Then, a GAI model is acquired, and the structured components in the high-quality base map are annotated according to annotation rules using the GAI model. Finally, the annotated high-quality base map is optimized to obtain an optimized architectural floor plan. This embodiment utilizes a GAI model (Generative Artificial Intelligence model) to automatically annotate architectural floor plans, thereby transforming them into recognizable, high-quality, and accurate annotated data. This achieves data optimization of architectural floor plans and enables intelligent upgrades to architectural design and analysis.

[0045] The vector drawing data annotation method based on GAI in this embodiment can be applied to terminals, including intelligent product terminals such as computers. In this embodiment, for example... Figure 1 As shown, the GAI-based vector drawing data annotation method includes the following steps:

[0046] Step S100: Obtain the original building plan drawings and preprocess the original building plan drawings to obtain a high-quality base map.

[0047] The current labeling requirements for project sample drawings specifically include: indicating the location dimensions of water drainage points such as sanitary ware and floor drains, and marking the floor elevation, drainage slope, and direction; indicating the location dimensions of air conditioners, electrical wiring (both high and low voltage), exhaust fans, kitchen exhaust ducts, etc., and ensuring that the openings do not conflict with adjacent risers. Special attention should be paid to ensuring that condensate pipe openings have sufficient height for condensate drainage; otherwise, separate condensate drainage should be considered; indicating the location dimensions of all interior door and window openings, and marking and positioning all equipment pipes; if the balcony facade is not... Miscellaneous details can be combined with the floor plan (in this case, balconies do not need to be numbered) and their location on the plan should be marked. The location dimensions of the balcony drain and various risers on the balcony, the location dimensions of vertical load-bearing components such as railings, and the partition dimensions of the glass should be marked. If the balcony elevation is complex, a separate balcony detail should be drawn (in this case, the balconies should be numbered). In addition to the location on the plan as described above, the partition dimensions of the balustrade (railing) and other components should be located on the balcony elevation and the material should be marked. The construction method, dimensions, materials, colors, etc. of each component should be determined in the detail drawing.

[0048] As can be seen, the current project sample images have significant limitations, specifically including: the data exhibits highly unstructured characteristics, integrating diverse information such as text annotations, symbolic markings, and 3D spatial relationships; the annotation task involves multiple dimensions such as geometric recognition, semantic understanding, and rule validation, resulting in a complex model parameter system and lengthy training cycles. Furthermore, due to significant differences in drawing styles, annotation habits, and professional focuses across different projects, existing models struggle to break through the stylistic constraints of a single project, exhibiting weak cross-project adaptability. In addition, annotation rules must consider multiple professional standards such as architecture, structure, and equipment, relying on multi-domain expert collaboration, leading to low data production efficiency and difficulty in meeting the needs of rapidly iterating projects. Therefore, this embodiment aims to optimize the aforementioned complex annotation requirements into annotation rules that meet lightweight and structured requirements. To this end, this embodiment reduces data complexity by clearly defining the scope and rules of the annotation objects, thereby improving the training efficiency and generalization performance of the GAI model in this embodiment. The annotation rule in this embodiment is to annotate the positioning dimensions of the key features of the structured components, specifically including: for interior walls, annotating the positioning dimensions of each partition wall not expressed in the axial dimensions; for exterior walls and windows, annotating the positioning dimensions; for interior doors and windows, annotating the positioning dimensions of various interior door and window openings; for balconies and cantilever components, annotating the positioning dimensions of the outer edges.

[0049] Furthermore, in actual projects, there are various apartment layouts. Given the significant differences in the representation of architectural apartment layouts between projects, and the limited ability of current models to handle complex topological relationships and diverse representational habits, when selecting architectural apartment layouts as samples, it is necessary to minimize interference. This requires certain processing of the original architectural floor plans and narrowing the scope of annotation objects. Specifically, this embodiment preprocesses the original architectural floor plans to obtain a high-quality base map. The original architectural floor plans in this embodiment can be vector graphics.

[0050] In one implementation, the preprocessing of the original architectural floor plan in this embodiment includes the following steps:

[0051] Step S101: Using a recursive decomposition algorithm, the multiple nested blocks in the original architectural plan are deeply decomposed and transformed into basic geometric elements;

[0052] Step S102: Locate components and redundant information that are not related to the annotation, and delete the components and redundant information that are not related to the annotation.

[0053] Preprocessing original architectural floor plans is a core technical step in improving the quality of automatically labeled data. Its core objective is to construct a high-quality base map with low interference and high structure through structural simplification and rule alignment. Original architectural floor plans contain a dense concentration of unstructured elements, severely interfering with core labeled objects. Furthermore, these plans include various unlabeled components such as sofas, dining tables, toilets, and air conditioner units, whose outlines intersect and overlap with wall axes and door / window openings. In addition, the nesting of blocks in the original floor plans leads to geometric distortion; elements such as doors, windows, and cabinets exist as multiple nested blocks. Directly deleting blocks or incomplete decomposition will result in the loss of geometric anchor points for the labels. Therefore, to address the interference from unstructured elements and the nesting of blocks in original architectural floor plans, this embodiment develops a GAI model to automate the preprocessing of original architectural floor plans. Specifically, a recursive decomposition algorithm is first used to deeply decompose the multiple nested blocks (such as doors, windows, and cabinets) in the original architectural floor plans, transforming them into basic geometric elements such as lines and arcs. Then, components unrelated to the annotation (such as sofas, toilets, etc.) and redundant information (such as decorative patterns and repeating lines) are located and deleted. This embodiment utilizes the GAI model's layer and element closing functions throughout the preprocessing process, reducing the preprocessing time of a single original architectural floor plan to minutes. The processed drawings conform to the AI ​​data standards for architectural floor plan dimension annotation, providing a high-quality base map for subsequent intelligent annotation tasks.

[0054] Furthermore, in other implementations, the preprocessing in this embodiment also includes data type compatibility handling. Specifically, the original architectural floor plan in this embodiment can be a CAD drawing, a PDF drawing, or an image in .jpg / .png format. This embodiment can extract vector data from original architectural floor plans of different formats based on a CAD parsing library, and structural components can be identified based on this vector data. In addition, this embodiment can also extract scale information from the original architectural floor plan based on OCR (Optical Character Recognition) technology, such as scale data located in the corner of the drawing. Size conversion can be performed based on this scale information, ensuring that original architectural floor plans of different sizes are converted to the same size, facilitating subsequent processing. Furthermore, this embodiment can also uniformly convert drawings from different sources into a Cartesian coordinate system, ensuring consistency in the coordinate calculation of elements such as walls and axes, and avoiding errors caused by drawing orientation or scaling.

[0055] Step S200: Obtain the GAI model and use the GAI model to annotate the structured components in the high-quality base map according to the annotation rules.

[0056] In this embodiment, after obtaining a high-quality base map, a GAI model is acquired. Based on this GAI model, the high-quality base map is annotated. During the annotation process, this embodiment uses standard rules for annotation, specifically annotating the positioning dimensions of the key features of the structured components. Specifically, the steps include:

[0057] Step S201: Automatically identify the structured components in the high-quality base map based on the GAI model, and automatically determine the position of the annotation lines;

[0058] Step S202: Based on the annotation rules and the determined annotation line positions, automatically complete the annotation process of the structured components.

[0059] In practical application, this embodiment first trains the GAI model. During the training process, this embodiment first performs preprocessing and annotation on several drawings corresponding to several projects to obtain a dataset. The preprocessing at this time is the same as described above, and the annotation is performed according to the annotation requirements of this embodiment. For example, in the data collection stage, for 362 projects and 2014 CAD drawings, including thousands of floor plans, after the above preprocessing and annotation, a dataset of 600 image samples is generated. Next, the dataset is output in three different resolutions: low resolution (256×256 pixels), medium resolution (640×640 pixels), and high resolution (1024×1024 pixels), with an equal number of samples generated at each resolution. Then, the dataset is divided into a training set (480 images), a validation set (60 images), and a test set (60 images), with an equal number of samples at each resolution. This embodiment allows the trained GAI model to adapt to different input size requirements, ensuring the model's generalization ability across multiple scales. Finally, based on the training set, a preset neural network model (such as the generative model GAI) is trained to obtain the GAI model. The trained GAI model can automatically annotate architectural floor plans, thereby converting architectural floor plans into recognizable, high-quality, and accurate annotation data.

[0060] During the annotation process, this embodiment can automatically identify structured components in a high-quality base map based on the GAI model trained above. For example, it can identify internal and external walls, balconies, stairs, and other structured components, and determine the key features of these components. For instance, for internal walls, the key feature is the various partition walls; for internal doors and windows, the key feature is the openings of the doors and windows. After identifying the key features, this embodiment can automatically determine the position of the annotation lines. Then, based on the annotation rules of this embodiment and the determined annotation line positions, the annotation processing of the structured components is automatically completed.

[0061] Furthermore, in other implementations, this embodiment can also identify structured components based on the YOLO model (an object recognition and localization algorithm based on deep neural networks) or the Segment Anything Model (SAM). The SAM model can also be used to segment the wall outline and obtain the coordinates of closed polygons, providing a basis for size conversion when needed. In addition, this embodiment can extract key features (such as openings in interior doors and windows, and partitions in interior walls) from the identified structured components. This embodiment uses the SIFT / SURF algorithm or a deep learning feature extractor (such as ResNet) to generate feature vectors, identifying these key features and providing a good foundation for automatic annotation using the GAI model in this embodiment. In other words, during the entire automatic annotation process, this embodiment can identify structured components and key features based on the aforementioned other models, and then use the GAI model trained in this embodiment for annotation. This reduces the computational load of the GAI model in this embodiment and improves its annotation efficiency.

[0062] Furthermore, in determining the position of the annotation lines, this embodiment calculates the distance between each structured component based on the coordinate data of the key features of the identified structured components. Then, combined with the extracted scale information, a size conversion is performed to determine the most suitable annotation line position. In this embodiment, the annotation lines do not cover the structured components, do not intersect each other, and do not obscure the annotation data. In practical applications, this embodiment can plan the annotation layout based on graph theory algorithms (such as Dijkstra's shortest path algorithm) to determine the optimal annotation area and thus set the annotation line position, ensuring the rationality and readability of the annotation data.

[0063] Furthermore, the GAI model in this embodiment is a model trained based on GAI (Generative Model). When generating annotation results, this embodiment can use the Transformer architecture or the Seq2Seq model (sequence-to-sequence model) to generate annotation results that conform to the specifications. In addition, for the annotation lines, arrows, and text boxes in the annotation results, this embodiment can combine vector graphics libraries (such as Matplotlib, SVGwrite) to generate appropriate annotation lines, arrows, and text boxes.

[0064] Furthermore, this embodiment provides a visual editing window during the annotation process and identifies the annotation results based on a preset rule recognition model. If the identified annotation data contains errors, such as contradictory annotations, annotation occlusion, or blurry fonts, the annotation position can be adjusted by dragging or adjusting control points to correct the annotation results. This embodiment also provides an interactive interface, allowing designers to manually adjust the annotation position or modify the text, and learn user preferences through reinforcement learning technology to optimize subsequent annotation strategies. For example, after recognizing that a user has moved a certain type of annotation multiple times, the optimal annotation position for that type of element can be automatically memorized. This embodiment can also convert the annotation results into CAD editable formats (.dwg), PDF, or images, and retain vector information for subsequent modifications. Therefore, this embodiment achieves a closed-loop human-machine collaborative quality inspection system. Automatic annotation can reduce most of the workload of manual annotation, which is beneficial for automatically annotating complex irregular components. In actual annotation tests, based on the annotation method of this embodiment, the annotation time for a single drawing was reduced from 1 hour to 20 minutes, with an annotation accuracy rate of 96%, significantly improving the annotation efficiency of architectural floor plans.

[0065] Step S300: Perform data optimization processing on the high-quality base map after annotation processing to obtain the optimized architectural plan drawing.

[0066] After annotation is completed, this embodiment allows users to select the export folder location and automatically export the images. To reduce interference, this embodiment sets the fill color and edge color of the structured components in the high-quality base map after annotation to black or gray, and sets the background to white to complete data optimization processing, thereby enhancing image contrast and improving the accuracy of subsequent AI recognition. Next, the high-quality base map after data optimization is exported to obtain the optimized architectural plan drawing, which includes not only the architectural plan base map but also the dimensioned location diagram of the internal structured components.

[0067] In other implementations, for architectural floor plans of irregularly shaped apartments or multi-story buildings, due to the complex dimensional relationships, this embodiment can introduce Graph Neural Networks (GNNs) to model the topological relationships of elements. For example, the apartment layout can be viewed as a graph structure, with nodes representing wall intersections and edges representing dimensional relationships, ensuring consistent annotation logic. Then, the GAI model of this embodiment is used for intelligent annotation, improving the annotation efficiency of irregularly shaped apartments or multi-story buildings.

[0068] In summary, this embodiment first acquires the original architectural floor plans and preprocesses them to obtain a high-quality base map. Then, it acquires a GAI model and uses the GAI model to annotate the structural components in the high-quality base map according to annotation rules. Finally, it performs data optimization processing on the annotated high-quality base map to obtain optimized architectural floor plans. This embodiment utilizes a GAI model to automatically annotate architectural floor plans, thereby transforming them into recognizable, high-quality, and accurate annotated data. This achieves data optimization of architectural floor plans and, through deep integration with the GAI model, realizes an intelligent upgrade of architectural design and analysis.

[0069] Based on the above embodiments, the present invention also provides a vector paper data annotation system based on GAI, which is used to implement the steps of the vector paper data annotation method based on GAI in the above method embodiments. Specifically, as Figure 2 As shown, the system in this embodiment includes a preprocessing module 10, an automatic annotation module 20, and a data optimization module 30. Specifically, the preprocessing module 10 is used to acquire the original architectural floor plan, preprocess the original architectural floor plan, and obtain a high-quality base map. The automatic annotation module 20 is used to acquire a GAI model, and use the GAI model to annotate the structural components in the high-quality base map according to annotation rules. The data optimization module 30 is used to perform data optimization processing on the annotated high-quality base map to obtain an optimized architectural floor plan.

[0070] The working principle of each module in the GAI-based vector paper data annotation system of this embodiment is the same as that of each step in the above method embodiment, and will not be repeated here.

[0071] The modules in the aforementioned GAI-based vector drawing data annotation system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of the terminal in hardware form or independent of it, or stored in the terminal's memory in software form, so that the processor can call and execute the corresponding operations of each module.

[0072] Based on the above embodiments, the present invention also provides a terminal, the principle block diagram of which can be as follows: Figure 3 As shown. The terminal may include one or more processors 100 ( Figure 3(Only one is shown in the diagram), a memory 101, and a computer program 102 stored in the memory 101 and executable on one or more processors 100. For example, a GAI-based vector paper data annotation program. When one or more processors 100 execute the computer program 102, they can implement the various steps in the GAI-based vector paper data annotation method embodiment. Alternatively, when one or more processors 100 execute the computer program 102, they can implement the functions of various modules / units in the GAI-based vector paper data annotation system embodiment, without limitation here.

[0073] In one embodiment, the processor 100 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0074] In one embodiment, memory 101 may be an internal storage unit of an electronic device, such as a hard drive or RAM. Memory 101 may also be an external storage device of the electronic device, such as a plug-in hard drive, smart media card (SMC), secure digital (SD) card, flash card, etc. Furthermore, memory 101 may include both internal and external storage units. Memory 101 is used to store computer programs and other programs and data required by the terminal. Memory 101 can also be used to temporarily store data that has been output or will be output.

[0075] Those skilled in the art will understand that Figure 3 The block diagram shown is merely a partial structural diagram related to the present invention and does not constitute a limitation on the terminal to which the present invention is applied. A specific terminal may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0076] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, operational databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual operating data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAM bus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A vector drawing data annotation method based on GAI, characterized in that, The method includes: Obtain the original building floor plan, and preprocess the original building floor plan to obtain a high-quality base map; Obtain the GAI model, and use the GAI model to annotate the structured components in the high-quality base map according to the annotation rules; The high-quality base map after annotation is optimized to obtain optimized architectural floor plans. The preprocessing of the original architectural floor plan to obtain a high-quality base map includes: Using a recursive decomposition algorithm, the multiple nested blocks in the original architectural plan are deeply decomposed and transformed into basic geometric elements; Locate components and redundant information that are not related to the annotations, and delete them. Vector data was extracted from the original architectural floor plan based on a CAD parsing library. The scale information is extracted from the original architectural floor plan using OCR technology, and the scale information is used for size conversion. The step of using the GAI model to annotate the structured components in the high-quality base map according to the annotation rules includes: Based on the GAI model, the structured components in the high-quality base map are automatically identified, and the positions of the annotation lines are automatically determined. Based on the annotation rules and the determined annotation line positions, the annotation process of the structured component is automatically completed. The annotation rules are used to mark the positioning dimensions of the key features of the structured component.

2. The vector drawing data annotation method based on GAI according to claim 1, characterized in that, The training methods for the GAI model include: After preprocessing and annotation of several drawings corresponding to several projects, a dataset is obtained; The dataset is output in three different resolutions and divided into a training set, a validation set, and a test set, each of which includes an equal number of samples for each resolution. The GAI model is obtained by training a preset neural network model based on the training set.

3. The vector drawing data annotation method based on GAI according to claim 1, characterized in that, The process of optimizing the high-quality base map after annotation to obtain optimized architectural floor plans includes: Set the fill color and edge color of the structured components in the high-quality base map after annotation to black or gray, and set the background to white to complete the data optimization process. The high-quality base map after data optimization is exported to obtain the optimized building plan drawing.

4. The vector drawing data annotation method based on GAI according to claim 1, characterized in that, The method further includes: A visual editing window is provided during the annotation process, and the annotation results are identified based on a preset rule recognition model; If there are errors in the identified and labeled data, the labeling results can be corrected by dragging or adjusting the control points.

5. A vector drawing data annotation system based on GAI, characterized in that, The system is used to implement the steps of the vector paper data annotation method based on GAI according to any one of claims 1-4, and the system includes: The preprocessing module is used to acquire the original building floor plan drawings, preprocess the original building floor plan drawings, and obtain a high-quality base map. An automatic annotation module is used to obtain a GAI model and use the GAI model to annotate the structured components in the high-quality base map according to annotation rules; The data optimization module is used to optimize the data of the high-quality base map after annotation processing to obtain optimized architectural floor plans.

6. A terminal, characterized in that, The terminal includes a memory, a processor, and a GAI-based vector paper data annotation program stored in the memory and executable on the processor. When the processor executes the GAI-based vector paper data annotation program, it implements the steps of the GAI-based vector paper data annotation method as described in any one of claims 1-4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a vector drawing data annotation program based on GAI, and the vector drawing data annotation program based on GAI implements the steps of the vector drawing data annotation method based on GAI as described in any one of claims 1-4 on the computer-readable storage medium.

Citation Information

Patent Citations

  • Automatic generation method and device for building electrical design drawing and storage device

    CN110059358A

  • Design drawing conversion method and apparatus and related device

    WO2023088087A1