Vector drawing data labeling method and system based on GAI and terminal

Abstract

The invention discloses a GAI-based vector drawing data labeling method, system and terminal, and the method comprises the steps: obtaining an original building plane drawing, and carrying out the preprocessing of the original building plane drawing, and obtaining a high-quality base drawing; a GAI model is obtained, and the GAI model is used for conducting labeling processing on the structured components in the high-quality base map according to labeling rules; and performing data optimization processing on the marked high-quality base drawing to obtain a building plane optimization drawing. According to the method, the building plane drawing is converted into recognizable high-quality and accurate annotation data, automatic annotation of the building plane drawing is achieved, data optimization of the building plane drawing is achieved, and intelligent upgrading of building design and analysis is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, and in particular to a vector drawing data labeling method, system and terminal based on GAI. BACKGROUND

[0002] In the building design industry, building plan size labeling is a bridge connecting design concepts and engineering practice. The traditional method uses CAD software for manual labeling. Although the software has some automatic labeling functions, for complex images or special labeling requirements, manual adjustment is still required. This way is inefficient, prone to errors, and difficult to meet the needs of efficient and accurate design. In addition, since the labeling rules need to take into account building, structure, equipment and other multi-specialty specifications, relying on multi-field expert collaboration, the labeling method of the prior art makes the data production efficiency low, and it is difficult to meet the rapid iteration project requirements.

[0003] Therefore, the prior art also has defects. SUMMARY

[0004] The technical problem to be solved by the present application is to provide a vector drawing data labeling method, system and terminal based on GAI to solve the above defects of the prior art. The technical solution adopted by the present application is as follows:

[0005] In a first aspect, the present application provides a vector drawing data labeling method based on GAI, wherein the method comprises:

[0006] Obtain an original building plan drawing, preprocess the original building plan drawing to obtain a high-quality base drawing;

[0007] Obtain a GAI model, and use the GAI model to label and process the structured components in the high-quality base drawing according to the labeling rules;

[0008] Optimize the high-quality base drawing after labeling and processing to obtain a building plan optimization drawing, wherein the building plan optimization drawing includes an optimized building plan base drawing and an internal size labeling position drawing.

[0009] In an implementation manner, the preprocessing of the original building plan drawing to obtain a high-quality base drawing comprises:

[0010] Using a recursive decomposition algorithm, the multiple nested blocks in the original building plan drawing are deeply disassembled into basic geometric elements;

[0011] Positioning and labeling irrelevant components and redundant information, and deleting irrelevant components and redundant information.

[0012] In an implementation manner, the preprocessing of the original architectural plan paper to obtain a high-quality base drawing comprises:

[0013] extracting vector data in the original architectural plan paper based on a CAD analysis library;

[0014] extracting scale information from the original architectural plan paper based on an OCR technology, the scale information being used for size conversion.

[0015] In an implementation manner, the labeling processing of the structured component in the high-quality base drawing according to the labeling rule by using the GAI model comprises:

[0016] automatically identifying the structured component in the high-quality base drawing based on the GAI model, and automatically determining the position of a labeling line;

[0017] automatically completing the labeling processing of the structured component based on the labeling rule and the determined position of the labeling line, wherein the labeling rule is labeling of the key features of the structured component.

[0018] In an implementation manner, the training manner of the GAI model comprises:

[0019] obtaining a data set after preprocessing and labeling processing of a plurality of drawings corresponding to a plurality of projects;

[0020] outputting the data set into three different resolutions, and dividing the data set into a training set, a validation set and a test set, wherein each of the training set, the validation set and the test set comprises an equal amount of samples of each resolution;

[0021] training a preset neural network model based on the training set to obtain the GAI model.

[0022] In an implementation manner, the data optimization processing of the high-quality base drawing after the labeling processing to obtain an architectural plan optimization drawing comprises:

[0023] setting the fill color of the structured component and the line color of the structured component in the high-quality base drawing after the labeling processing to be black or gray, and setting the background to be white to complete the data optimization processing;

[0024] performing image export on the high-quality base drawing after the data optimization processing to obtain the architectural plan optimization drawing.

[0025] In an implementation manner, the method further comprises:

[0026] providing a visual editing window in the process of the labeling processing, and identifying the labeling result based on a preset rule identification model.

[0027] If the identified and labeled data has errors, the labeling result is corrected by dragging or adjusting the control points.

[0028] In a second aspect, the embodiments of the present application further provide a GAI-based vector drawing data labeling system, wherein the system is configured to implement the steps of the GAI-based vector drawing data labeling method according to any one of the above solutions, and the system comprises:

[0029] a preprocessing module configured to obtain an original architectural plan drawing, and perform preprocessing on the original architectural plan drawing to obtain a high-quality base drawing;

[0030] an automatic labeling module configured to obtain a GAI model, and use the GAI model to perform labeling processing on a structured component in the high-quality base drawing according to a labeling rule;

[0031] a data optimization module configured to perform data optimization processing on the high-quality base drawing after the labeling processing to obtain an architectural plan optimization drawing, wherein the architectural plan optimization drawing comprises an optimized architectural plan base drawing and an internal dimension labeling position drawing.

[0032] In a third aspect, the embodiments of the present application further provide a terminal, wherein the terminal comprises a memory, a processor, and a GAI-based vector drawing data labeling program stored in the memory and executable on the processor, and the processor implements the steps of the GAI-based vector drawing data labeling method according to any one of the above solutions when executing the GAI-based vector drawing data labeling program.

[0033] In a fourth aspect, the embodiments of the present application further provide a computer-readable storage medium, wherein the computer-readable storage medium stores a GAI-based vector drawing data labeling program, and the GAI-based vector drawing data labeling program implements the steps of the GAI-based vector drawing data labeling method according to any one of the above solutions on the computer-readable storage medium.

[0034] Beneficial effects: compared with the prior art, the application provides a GAI-based vector drawing data labeling method, the original architectural plan drawing is first acquired, the original architectural plan drawing is pretreated to obtain a high-quality base drawing. Then, a GAI model is acquired, and the GAI model is used to label and process the structured components in the high-quality base drawing according to labeling rules. Finally, the high-quality base drawing after the labeling processing is subjected to data optimization processing to obtain an architectural plan optimization drawing. The GAI model is used to automatically realize labeling of the architectural plan drawing, so as to convert the architectural plan drawing into identifiable high-quality and accurate labeling data, realize automatic labeling of the architectural plan drawing, realize data optimization of the architectural plan drawing, and realize intelligent upgrading of architectural design and analysis. BRIEF DESCRIPTION OF DRAWINGS

[0035] Figure 1 The flowchart of the preferred embodiment of the GAI-based vector drawing data labeling method provided for the embodiment of the application.

[0036] Figure 2 The architecture schematic diagram of the GAI-based vector drawing data labeling system provided for the embodiment of the application.

[0037] Figure 3 The principle block diagram of the terminal provided for the embodiment of the application. DETAILED DESCRIPTION

[0038] To make the objectives, technical solutions and effects of the application clearer and more explicit, the application is further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the application and do not limit the application.

[0039] The flowchart shown in the drawings is only an example and does not necessarily include all contents and operations or steps, nor does it necessarily be executed in the described order. For example, some operations or steps can be decomposed, combined or partially combined, so that the actual execution order can be changed according to the actual situation.

[0040] It should be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in the application 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 facilitate clear description of the technical solutions of the embodiments of the present application, in the embodiments of the present application, the terms "first", "second", etc. are used to distinguish the same or similar items with basically the same function and role. For example, the first control information and the second control information are only used to distinguish different control information, and do not limit the order.

[0042] It can be understood by those skilled in the art that the terms "first", "second", etc. do not limit the quantity and execution order, and the terms "first", "second", etc. also do not necessarily mean different.

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

[0044] Based on the problems of the prior art, the present application provides a GAI-based vector drawing data labeling method, which can realize automatic labeling of architectural plan drawings based on the method of the present embodiment. In specific applications, the present embodiment first acquires an original architectural plan drawing, pre-processes the original architectural plan drawing to obtain a high-quality base drawing. Then, a GAI model is acquired, and the GAI model is used to label and process the structured components in the high-quality base drawing according to labeling rules. Finally, the high-quality base drawing after labeling and processing is subjected to data optimization processing to obtain an architectural plan optimization drawing. The present embodiment automatically realizes labeling of architectural plan drawings by using a GAI model (Generative AI model), thereby converting the architectural plan drawings into identifiable high-quality and accurate labeling data, realizing data optimization of the architectural plan drawings, and realizing intelligent upgrading of architectural design and analysis.

[0045] The GAI-based vector drawing data labeling method of the present embodiment can be applied to a terminal, which includes intelligent product terminals such as computers. As shown in Figure 1 the GAI-based vector drawing data labeling method includes the following steps:

[0046] Step S100, acquire an original architectural plan drawing, pre-process the original architectural plan drawing to obtain a high-quality base drawing.

[0047] The current labeling requirements for project sample drawings include: indicating the positioning size of water and drainage points such as sanitary wares and floor drains, and labeling the floor elevation and drainage slope; labeling the positioning size of air conditioners, strong and weak electricity, exhaust fans, kitchen flues, etc., and paying attention to the fact that the hole does not conflict with the adjacent vertical pipes, in addition, special attention should be paid to the fact that the condensate pipe hole should have enough height for the drainage of condensate water, otherwise separate drainage of condensate water should be considered; labeling the positioning size of various interior door and window hole sizes, and labeling all equipment pipes and positioning; if the balcony elevation is not complex, it can be combined with the house type sample (at this time the balcony can not be numbered), and the positioning on the plane should be performed, the positioning size of the balcony floor drain and various vertical pipes on the balcony, the positioning size of the vertical force member such as the railing, and the partition size of the glass should be labeled; if the balcony elevation is complex, a separate balcony sample should be drawn (at this time the balcony should be numbered), in addition to the above-mentioned positioning on the plan, the partition size of the rail plate (railing) should be positioned and labeled on the balcony elevation, and the construction method, size, material, color, etc. of each component should be determined in the node diagram.

[0048] It can be seen that the current project sample drawings have significant limitations, including: the data presents highly unstructured characteristics, integrating text annotations, symbolic identifiers, and three-dimensional spatial relationships; the labeling task involves multiple dimensions such as geometric recognition, semantic understanding, and rule checking, resulting in complex model parameter systems and long training cycles. At the same time, due to the significant differences in drawing expression style, labeling habits, and professional focus of different projects, the existing model is difficult to break through the shackles of a single project style, and the cross-project migration and adaptation ability is weak. In addition, the labeling rules need to consider multiple professional specifications such as architecture, structure, and equipment, and rely on multi-disciplinary expert collaboration, making data production inefficient and difficult to meet the rapid iteration needs of projects. Therefore, the present embodiment needs to optimize the above complex labeling requirements to meet the lightweight and structured labeling rules. To this end, the present embodiment reduces data complexity by clearly defining the scope and rules of the labeling objects, in order to improve the training efficiency and generalization performance of the GAI model. The labeling rules of the present embodiment are to label the positioning size of the key features of the structured components, including: for interior walls, labeling the positioning size of each partition wall that is not expressed in the axial size; for exterior walls and windows, labeling the positioning size; for interior doors and windows, labeling the positioning size of various interior door and window holes; for balconies and cantilever components, labeling the positioning size of the outer edge.

[0049] In addition, in actual projects, there are various house types. Given that there are significant differences between projects in the expression of architectural house type layouts, and the current model has limited ability to handle complex topological relationships and diverse expression habits, when selecting an architectural house type layout as a sample, it is necessary to consider reducing interference as much as possible, and the original architectural plan should be processed and the range of labeled objects should be reduced. Specifically, the original architectural plan is preprocessed to obtain a high-quality base map. The original architectural plan of the embodiment can be a vector drawing.

[0050] In an implementation manner, the original architectural plan is preprocessed by the embodiment, including the following steps:

[0051] Step S101, using a recursive decomposition algorithm to deeply disassemble multiple nested blocks in the original architectural plan, and converting them into basic geometric elements;

[0052] Step S102, locating and labeling irrelevant components and redundant information, and deleting irrelevant components and redundant information.

[0053] The preprocessing of the original architectural plan is a core technical link to improve the quality of automatic labeling data. The core goal is to construct a low-interference, highly structured high-quality base map through structure simplification and rule alignment. Because the original architectural plan contains a large number of unstructured elements, which seriously interfere with the core labeling objects, and the original architectural plan contains sofas, dining tables, toilets, air conditioning positions and other types of non-labeling components, the outline of which overlaps with the wall axis, door and window hole line. In addition, the block nesting in the original architectural plan will cause geometric distortion, and the door, window, and cabinet are in the form of multiple nested blocks. Directly deleting the block or incomplete disassembly will cause the labeling to lose the geometric anchor point. Therefore, to solve the problem of unstructured elements in the original architectural plan and block nesting, the embodiment can develop a GAI model to realize the automatic preprocessing of the original architectural plan. Specifically, first, using a recursive decomposition algorithm to deeply disassemble multiple nested blocks (such as doors, windows, and cabinets) in the original architectural plan, and converting them into basic geometric elements such as straight lines and arcs. Then locate the components irrelevant to labeling (such as sofas and toilets) and redundant information (such as decorative patterns and repeated lines), and delete the components irrelevant to labeling and redundant information. The entire preprocessing process of the embodiment uses the closed layer and closed graphic element function of the GAI model, shortens the preprocessing time of a single original architectural plan to minutes, and the processed drawing meets the AI data standard for architectural plan size labeling, providing a high-quality base map for subsequent intelligent labeling tasks.

[0054] In addition, in other implementations, the preprocessing of the present embodiment also includes compatibility processing of data types. Specifically, the original architectural plan paper of the present embodiment can be a CAD paper, a PDF paper, or a picture in.jpg / .png format, and the present embodiment can extract vector data in the original architectural plan paper in different formats based on a CAD analysis library, and based on these vector data, structured components can be identified. In addition, the present embodiment can also extract scale information from the original architectural plan paper based on an OCR (Optical Character Recognition) technology, such as scale data located in the corner of the paper, and based on the scale information, size conversion can be performed, so that the original architectural plan paper of different sizes can be converted into the same size, facilitating subsequent processing. In addition, the present embodiment can also convert papers of different sources into a Cartesian coordinate system, ensuring the consistency of coordinate calculation of elements such as walls and axes, and avoiding errors caused by paper direction or scaling.

[0055] Step S200, obtaining a GAI model, and using the GAI model to perform annotation processing on the structured components in the high-quality base map according to annotation rules.

[0056] In the present embodiment, after obtaining the high-quality base map, a GAI model is obtained, and the high-quality base map is annotated based on the GAI model. In the annotation process, the present embodiment performs annotation based on standard rules, that is, annotation of the positioning size of the key features of the structured components. Specifically, the following steps are included:

[0057] Step S201, automatically identifying the structured components in the high-quality base map based on the GAI model, and automatically determining the position of the annotation line;

[0058] Step S202, automatically completing the annotation processing of the structured components based on the annotation rules and the determined position of the annotation line.

[0059] In a specific application, the embodiment first trains a GAI model. In the training process, the embodiment first pre-processes and labels a number of drawings corresponding to a number of projects to obtain a data set. The pre-processing is the same as described above, and the labeling is performed according to the labeling requirements of the embodiment. For example, in the data collection stage, 2014 CAD drawings corresponding to 362 projects, including thousands of house type drawings, are pre-processed and labeled to generate a data set of 600 image samples. Then, the data set is output in three different resolutions, specifically low resolution (256x256 pixels), medium resolution (640x640 pixels), and high resolution (1024x1024 pixels), and an equal number of samples are generated for each resolution. Then, the data set is divided into a training set (480 samples), a validation set (60 samples), and a test set (60 samples), and each set includes an equal number of samples for each resolution. The trained GAI model can adapt to different input size requirements and has good generalization ability at multiple scales. Finally, the pre-set neural network model (such as the generative model GAI) is trained based on the training set to obtain the GAI model. The trained GAI model can automatically label architectural plan drawings to convert them into high-quality and accurate labeled data.

[0060] In the labeling process, the embodiment can automatically identify structured components in high-quality base drawings based on the trained GAI model described above. For example, it can identify structured components such as interior and exterior walls, balconies, and staircases, and determine the key features of these structured components. For example, for interior walls, the key features are the individual partitions; for interior doors and windows, the key features are the openings of the interior doors and windows. After identifying the key features, the embodiment can automatically determine the positions of the labeling lines. Then, based on the labeling rules of the embodiment and the determined labeling line positions, the labeling of the structured components is automatically completed.

[0061] In addition, in other implementations, the embodiment can also identify the structured component based on a YOLO model (an object recognition and positioning algorithm based on a deep neural network) or a Segment Anything Model (SAM) when identifying the structured component, and can also segment the wall contour through the SAM model to obtain closed polygon coordinates to provide a basis when size conversion is needed. In addition, the embodiment can also extract key features (such as the openings of interior doors and windows, and partition walls in interior walls) of the identified structured component, and the embodiment uses a SIFT / SURF algorithm or a deep learning feature extractor (such as a ResNet) to generate a feature vector, which provides a good basis for automatic labeling using a GAI model in the embodiment. That is, in the entire automatic labeling process, the embodiment can identify the structured component and key features based on the above-mentioned other models, and then use the GAI model trained by the embodiment to perform labeling, which can reduce the computational amount of the GAI model of the embodiment and improve the labeling efficiency of the GAI model of the embodiment.

[0062] In addition, when determining the position of the labeling line, the embodiment can calculate the distance between each structured component according to the coordinate data of the key features of the identified structured component, and then combine the extracted scale information to perform size conversion to determine the most appropriate labeling line position. The labeling line position of the embodiment will not cover the structured component, the labeling lines will not intersect each other, and will not cause the labeling data to be obscured. In actual application, the embodiment can plan the labeling layout based on a graph theory algorithm (such as a Dijkstra shortest path algorithm) to determine the optimal labeling area to set the labeling line position, which can ensure the rationality and readability of the labeling data.

[0063] Further, the GAI model of the embodiment is a model trained based on a GAI (generative model), and when generating the labeling result, the embodiment can use a Transformer architecture or a Seq2Seq model (sequence-to-sequence model) to generate a standardized labeling result. In addition, for the labeling line, arrow, and text box in the labeling result, the embodiment can combine a vector graphics library (such as Matplotlib, SVGwrite) to generate a suitable labeling line, arrow, and text box.

[0064] Further, the embodiment provides a visual editing window in the process of the labeling processing, and identifies the labeling result based on a preset rule recognition model; if the identified labeled data has errors, such as contradictory labeling, occluded labeling, and fuzzy font labeling, the labeling position is adjusted by dragging or adjusting the control point to correct the labeling result. The embodiment can also provide an interactive interface to allow designers to manually adjust the labeling position or modify the text, and learn user preferences through reinforcement learning technology to optimize subsequent labeling strategies. For example, after identifying that the user has moved a certain type of label multiple times, the best labeling position for this type of element can be automatically remembered. The embodiment can also convert the labeling result into a CAD editable format (.dwg), PDF or picture, and retain the vector information for subsequent modification. As can be seen, the embodiment realizes a human-computer collaborative quality inspection closed loop, and automatic labeling can reduce most of the workload of manual labeling, which is conducive to automatically labeling complex special-shaped components, and in actual labeling tests, based on the labeling method of the embodiment, the time required for labeling a single drawing is reduced from 1 hour to 20 minutes, and the labeling accuracy is 96%, which significantly improves the labeling efficiency of architectural plan drawings.

[0065] Step S300, performing data optimization processing on the high-quality base drawing after labeling processing to obtain an architectural plan optimization drawing.

[0066] After completing the labeling, the embodiment can select the folder position for export and automatically complete the export of the image. In order to reduce interference, the embodiment can set the fill color of the structural component and the line color of the structural component in the high-quality base drawing after labeling processing to black or gray, and set the background to white, complete the data optimization processing, to enhance the image contrast and improve the accuracy of subsequent AI recognition. Then, the high-quality base drawing after data optimization processing is subjected to image export to obtain the architectural plan optimization drawing, which includes not only the architectural plan base drawing but also the size labeling position drawing of the internal structural component.

[0067] In other implementations, for architectural plan drawings of special-shaped house types or multi-story buildings, since the size correlation is complex, the embodiment can introduce a graph neural network (GNN) to model the topological relationship of elements, such as regarding the house type as a graph structure, the node as the intersection of the wall, and the edge as the size correlation, to ensure the consistency of the labeling logic. Then, the GAI model of the embodiment is used for intelligent labeling to improve the labeling efficiency of special-shaped house types or multi-story buildings.

[0068] In summary, the embodiment first acquires an original architectural plan paper, pre-processes the original architectural plan paper to obtain a high-quality base map. Then, a GAI model is acquired, and the GAI model is used to label the structured components in the high-quality base map according to labeling rules. Finally, the high-quality base map after labeling is processed to obtain an optimized architectural plan paper. The embodiment automatically labels the architectural plan paper by using the GAI model, so as to convert the architectural plan paper into identifiable high-quality and accurate labeled data, realizes data optimization of the architectural plan paper, and realizes intelligent upgrading of architectural design and analysis through deep fusion with the GAI model.

[0069] Based on the above embodiment, the application further provides a GAI-based vector drawing data labeling system, which is used to realize the steps of the GAI-based vector drawing data labeling method in the above method embodiment. Specifically, as shown in Figure 2 The system of the embodiment includes a preprocessing module 10, an automatic labeling module 20, and a data optimization module 30. Specifically, the preprocessing module 10 is used to acquire an original architectural plan paper, pre-process the original architectural plan paper to obtain a high-quality base map. The automatic labeling module 20 is used to acquire a GAI model, and use the GAI model to label the structured components in the high-quality base map according to labeling rules. The data optimization module 30 is used to process the high-quality base map after labeling to obtain an optimized architectural plan paper.

[0070] The working principles of the various modules in the GAI-based vector drawing data labeling system of the embodiment are the same as the principles of the various steps in the above method embodiment, and will not be repeated here.

[0071] The various modules in the above GAI-based vector drawing data labeling system can be realized by software, hardware, and combinations thereof, in whole or in part. The above modules can be embedded in or independent of the processor in the terminal in hardware form, or can be stored in the memory in the terminal in software form, so as to be called and executed by the processor to perform the operations corresponding to the above various modules.

[0072] Based on the above embodiment, the application further provides a terminal, and the principle block diagram of the terminal can be as shown in Figure 3 The terminal can include one or more processors 100 Figure 3The memory 101 and the computer program 102 stored in the memory 101 and executable on the one or more processors 100 are shown in FIG. 1. For example, a GAI-based vector drawing data labeling program. The one or more processors 100 can implement each step in the GAI-based vector drawing data labeling method embodiment when executing the computer program 102. Alternatively, the one or more processors 100 can implement the functions of each module / unit in the GAI-based vector drawing data labeling system embodiment when executing the computer program 102, which is not limited here.

[0073] In an embodiment, the processor 100 can be a central processing unit (CPU), and can also be other general-purpose processors, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

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

[0075] Those skilled in the art can understand that, Figure 3 The block diagram shown in FIG. 1 is only a block diagram of part of the structure related to the present application scheme, and does not constitute a limitation on the terminal to which the present application scheme is applied. The specific terminal can include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0076] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by a computer program instructing relevant hardware. The computer program can be stored in a non-volatile computer readable storage medium, and when executed, can include the processes of the above-mentioned embodiment methods. Any reference to memory, storage, operating database or other medium used in the embodiments of the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchl ink) DRAM (SLDRAM), memory bus (Rambus) 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 application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A GAI-based vector drawing data labeling method, characterized in that, The method comprises: obtaining an original architectural plan paper, preprocessing the original architectural plan paper to obtain a high-quality base map; obtaining a GAI model, and labeling the structured components in the high-quality base map according to a labeling rule by using the GAI model; performing data optimization processing on the high-quality base map after labeling to obtain an optimized architectural plan paper.

2. The GAI-based vector drawing data annotation method of claim 1, wherein, The preprocessing of the original architectural plan paper to obtain a high-quality base map comprises: using a recursive decomposition algorithm to deeply disassemble multiple nested blocks in the original architectural plan paper into basic geometric elements; locating components irrelevant to labeling and redundant information, and deleting the components irrelevant to labeling and the redundant information. 3.The GAI-based vector drawing data labeling method of claim 1, wherein, The preprocessing of the original architectural plan paper to obtain a high-quality base map comprises: extracting vector data in the original architectural plan paper based on a CAD analysis library; extracting scale information from the original architectural plan paper based on an OCR technology, the scale information being used for size conversion.

4. The GAI-based vector drawing data annotation method of claim 1, wherein, The labeling processing of the structured components in the high-quality base map according to a labeling rule by using the GAI model comprises: automatically identifying the structured components in the high-quality base map based on the GAI model, and automatically determining the positions of labeling lines; automatically completing the labeling processing of the structured components based on the labeling rule and the determined positions of the labeling lines, wherein the labeling rule is to label the key features of the structured components.

5. The GAI-based vector drawing data annotation method of claim 4, wherein, The training method of the GAI model comprises: performing preprocessing and labeling processing on a plurality of papers corresponding to a plurality of projects to obtain a data set; outputting the data set into three different resolutions, and dividing the data set into a training set, a validation set and a test set, each of the training set, the validation set and the test set comprising an equal amount of samples of each resolution; training a preset neural network model based on the training set to obtain the GAI model.

6. The GAI-based vector drawing data annotation method of claim 1, wherein, The data optimization processing of the high-quality base map after labeling to obtain an optimized architectural plan paper comprises: setting the fill color of the structured components and the line color of the structured components in the high-quality base map after labeling to black or gray, and setting the background to white to complete the data optimization processing; performing image export on the high-quality base map after data optimization to obtain the optimized architectural plan paper.

7. The GAI-based vector drawing data annotation method of claim 4, wherein, The method further comprises: providing a visual editing window during the labeling processing, and identifying the labeling result based on a preset rule recognition model; if the labeled data contains errors, correcting the labeling result by dragging or adjusting the control points.

8. A GAI-based vector drawing data labeling system, characterized by, The system is used to implement the steps of the GAI-based vector paper data labeling method according to any one of claims 1-7, and comprises: a preprocessing module configured to obtain an original architectural plan paper, and preprocess the original architectural plan paper to obtain a high-quality base map; an automatic labeling module configured to obtain a GAI model, and label the structured components in the high-quality base map according to a labeling rule by using the GAI model. A data optimization module is configured to perform data optimization on the high-quality base drawing after the annotation to obtain an optimized building plan drawing.

9. A terminal, characterized by comprising: The terminal comprises a memory, a processor, and a GAI-based vector drawing data annotation program stored in the memory and executable on the processor; when the processor executes the GAI-based vector drawing data annotation program, the steps of the GAI-based vector drawing data annotation method according to any one of claims 1-7 are implemented.

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

Images