Automatic recognition method for building dwg drawing based on vector graph feature analysis

Abstract

This application discloses an automatic parsing and recognition method for architectural DWG drawings based on vector graphic feature analysis, belonging to the field of architectural engineering design technology. The method includes extracting data from the DWG drawings according to the requirements for drawing recognition; grouping the extracted data according to data type and attributes; analyzing each group of data; identifying door features; combining extracted text to identify doors and their corresponding numbers in the drawing; analyzing wall features; identifying walls through two recognition methods; identifying windows and railings based on window numbers and parallel line features; and defining building rooms based on components: doors, walls, windows, and railings. This method can assist designers in making modifications, thereby achieving automatic checking of drawing consistency and can be widely applied to various design and construction scenarios.

Description

Technical Field

[0001] This application belongs to the field of architectural engineering design technology, specifically, it relates to an automatic parsing and recognition method for architectural DWG drawings based on vector graphic feature analysis. Background Technology

[0002] In the field of architectural design, the application of BIM (Building Information Modeling) technology is still in its initial stage of development. To promote the application and development of BIM-based design, many provincial and municipal administrative regions in China have gradually implemented BIM review systems. These systems require construction projects to submit corresponding three-dimensional models when submitting their applications for approval.

[0003] However, there is currently a lack of a suitable review mechanism for uploaded 3D models, making it impossible to ensure the correspondence between the 3D model and the construction drawings. At present, PDF format construction drawings are still the only legally valid construction drawing design deliverables. This leads many design projects to focus only on ensuring the quality of construction drawings, while paying insufficient attention to 3D models. Consequently, discrepancies between 3D models and construction drawings frequently occur, and there are even cases where the 3D model is merely an empty shell, lacking substantial content.

[0004] Currently, apart from manually verifying the matching degree between construction drawings and 3D models, there is no technical means to automatically check the consistency between drawings and models. Since construction drawings and 3D models of building projects are usually quite complex, manual verification is not only extremely labor-intensive but also has a high error rate, making it impossible to cope with the large number of building projects and therefore not practically feasible.

[0005] As the core carrier of project progress, drawings have long been limited by the differences in the technical systems of design units and the personalized drawing habits of designers, resulting in the industry status quo of "different layers for the same drawing, different colors for the same structure". While this difference provides flexibility for design innovation, it brings huge obstacles to the subsequent digital recognition of drawings and extraction of component information. Traditional recognition technology often relies on preset layer naming rules or fixed color coding systems. Once non-standard drawing formats are encountered, problems such as recognition gaps and missed component identification will occur, which seriously affects the efficiency of design review, engineering quantity calculation and other links. Summary of the Invention

[0006] To address the aforementioned problems and technical deficiencies, this application adopts the following technical solution: an automatic parsing and recognition method for architectural DWG drawings based on vector graphic feature analysis, characterized by comprising the following steps: Based on the requirements for identifying drawings, data is extracted from DWG drawings using vector primitive analysis and attribute association. Based on the extracted data, three-dimensional features are obtained, and the data is grouped into three levels according to the three-dimensional features. Each group of data is then analyzed. A dual recognition mechanism is established to identify the drawing features of the door and, in combination with the extracted text, to identify the door in the drawing and its corresponding number. Analyze the double-line features of the wall drawing, design two recognition methods, and use the two recognition methods to identify the wall. Based on window numbering and parallel line features, combined with spatial constraints and numbering semantics, windows and railings are identified. The identified doors, walls, windows, and railings are magnified, enclosed, calculated, and restored to form the building rooms.

[0007] Preferably, the process of extracting data from the DWG drawing includes: First, extract the vector primitives from the DWG drawing, then read the underlying vector data of the drawing; Extract all text objects from the DWG drawing and establish a text-graphic mapping foundation based on the spatial relationship between text coordinates and graphic element coordinates; Extract the boundary vector information, fill pattern code, and fill ratio of the filled area to distinguish between structural fill and decorative fill. Determine the type of auxiliary component by the spatial range of the filled area and its connection with surrounding graphic elements.

[0008] Furthermore, the three-dimensional features are data type, geometric attributes, and spatial correlation, and the grouping rules for the three-level grouping include: First, divide the elements into three main categories based on data type: graphic elements, text elements, and fill elements, thus achieving the first level of grouping. Then, the graphic element groups are clustered according to their geometric attributes to achieve the second level of grouping; Finally, the distance and angle relationships between different primitives are calculated, and primitives that are spatially adjacent and have matching attributes are grouped into the same analysis group to achieve the third level of grouping.

[0009] Furthermore, the identification of doors and numbers in the drawings involves first initially identifying and traversing groups of arcs and line segments, then matching them using a door geometric feature library to exclude similar geometric combinations that are not door components, and finally using quantitative criteria to select groups of primitives that meet the combination features and mark them as candidate sets of door components. Then, using the spatial coordinates of the candidate set elements as the center, a search range with a radius of a preset drawing unit is set. Text elements within the search range are retrieved, the text content in the text elements is parsed, and the door numbering rules are matched. The successfully matched text numbers are bound to the candidate set elements to confirm the identity and unique number of the door components.

[0010] Furthermore, the two identification methods for identifying walls include: When the target group data is a double-line group, first determine whether the wall feature is a parallel double-line group; If so, use the first identification method: Traverse the parallel line segments in the line segment group, filter out double line groups with a parallelism error of less than the threshold, a line group width that meets the thickness range of the building wall, and a double line group length that is greater than the threshold, and mark them as walls. If not, use the second identification method: Based on common wall layers, calculate the connectivity and enclosure between walls, determine whether the target group data is a wall, and if it meets the wall characteristics, mark it as a wall.

[0011] Furthermore, the identification of the windows and railings includes: Traverse the line segment group, filter out the parallel line group with parallelism error less than the threshold and spacing within the width range of window or railing, and mark it as window or railing candidate set; Select parallel line groups from the candidate set, determine whether their start and end points are connected to the identified wall components, retain parallel line groups whose start and end points are both connected to the wall, and exclude independent decorative parallel lines. Centered on the preserved parallel line group, set the search range of the preset drawing unit, retrieve text elements and parse the text content, and match the window numbering rules; Mark the successfully matched parallel line groups as window components and bind them with the corresponding window numbers; Parallel line groups that do not match window numbers are marked as railings after confirmation based on building design codes and the surrounding spatial environment.

[0012] Furthermore, the enclosure of a building's rooms includes: The vector primitives of all identified components are enlarged proportionally according to a preset ratio, with the enlargement center being the geometric center of the component. The enlargement process fills the tiny gaps between components. Based on the magnified component vector data, the boundary tracking algorithm is used to traverse the outer boundaries of walls, windows and railings, identify the closed areas enclosed by these components, filter out areas with a closed area greater than a threshold, and mark these areas as candidate areas for the room. The candidate room area is scaled down proportionally to the inverse of the magnification ratio to restore the actual size of the component. Combined with the text elements in the room, the room's functional information is bound to complete the room recognition.

[0013] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the content of the automatic parsing and recognition method for architectural DWG drawings based on vector graphic feature analysis as described above.

[0014] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the content of the automatic parsing and recognition method for architectural DWG drawings based on vector graphic feature analysis as described above.

[0015] Compared to existing technologies, the beneficial effects of this application are as follows: (1) This application uses vector feature analysis algorithm to identify two-dimensional construction drawings, and then compares the identification results with three-dimensional models. By grouping and analyzing the vector graphics, it accurately captures the drawing rules and vector features of walls in specific drawings, thereby efficiently identifying planar walls. It further expands the identification of other components such as windows, railings, and rooms to form a complete component identification system. It can also prompt designers of the specific locations where the construction drawings and three-dimensional models do not correspond, so as to assist designers in making modifications. This achieves automatic checking of the consistency between drawings and models and can be widely applied to multiple design and construction scenarios. (2) This application does not require any additional adjustments to the drawing format. As long as the drawings conform to the national standard construction drawing specifications, customized analysis can be carried out for the uniqueness of each drawing, accurately identify various components, effectively solve the identification problem caused by different labels for the same drawing in the industry, adapt to the architectural plan vector image recognition solution of the entire industry, and provide efficient and universal technical support for the digital process of building construction; (3) This application starts with the essential vector features of building components. No matter what color layer the designer uses to draw the wall, the core vector parameters of the component can be accurately captured through the self-developed feature extraction algorithm. It has low hardware requirements and can run smoothly even on the designer's personal computer, thus realizing the automatic recognition of construction drawings. Attached Figure Description

[0016] In the attached diagram: Figure 1 This is a schematic diagram of the method steps in an embodiment of this application; Figure 2 This is a schematic diagram of the overall DWG drawing of an embodiment of this application; Figure 3 This is a schematic diagram of wall line extraction after grouping processing according to an embodiment of this application; Figure 4 This is a schematic diagram of a door combination of a 90-degree arc and a line segment according to an embodiment of this application; Figure 5 This is a schematic diagram of the consistency check in an embodiment of this application. Figure 1 ; Figure 6 This is a schematic diagram of the consistency check in an embodiment of this application. Figure 2 . Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of this application, but not all embodiments. Generally, the components of the embodiments of this application described and shown in the accompanying drawings can be arranged and designed in various different configurations. Example

[0018] like Figure 1 As shown, the automatic parsing and recognition method for architectural DWG drawings based on vector graphic feature analysis includes the following steps: Based on the requirements for identifying drawings, data is extracted from DWG drawings using vector primitive analysis and attribute association. Data in DWG drawings includes: Elements used for drawing components such as walls, doors, and windows; Text used to identify room function and door / window numbers; Used to assist in identifying the filling of shear wall columns.

[0019] The process of extracting data from DWG drawings includes: First, extract the vector primitives from the DWG drawing, then read the underlying vector data of the drawing; Extract all text objects from the DWG drawing and establish a text-graphic mapping foundation based on the spatial relationship between text coordinates and graphic element coordinates; Extract the boundary vector information, fill pattern code, and fill ratio of the filled area to distinguish between structural fill and decorative fill. Determine the type of auxiliary component by the spatial range of the filled area and its connection with surrounding graphic elements.

[0020] Based on the extracted data, three-dimensional features are obtained, and the data is grouped into three levels according to the three-dimensional features. Each group of data is then analyzed. The three-dimensional features are data type, geometric attributes, and spatial correlation. The grouping rules for the three-level grouping include: First, divide the elements into three main categories based on data type: graphic elements, text elements, and fill elements, thus achieving the first level of grouping. Then, the graphic element groups are clustered according to their geometric attributes to achieve the second level of grouping; Finally, the distance and angle relationships between different primitives are calculated, and primitives that are spatially adjacent and have matching attributes are grouped into the same analysis group to achieve the third level of grouping.

[0021] A dual recognition mechanism is established to identify the drawing features of the door and, in combination with the extracted text, to identify the door in the drawing and its corresponding number. The characteristic of drawing a door is that it is presented as a combination of 90-degree arcs and line segments on the drawing.

[0022] The dual recognition mechanism includes a gate geometry feature library and a quantization standard; The door geometry feature library defines the drawing features of door components as a combination of a 90° arc and two line segments; The quantitative standard is to determine that the central angle of the arc is 90°±5°, the radius of the arc is the radius of the building's doorway, the two line segments are tangent to the two endpoints of the arc, the length of the line segments is greater than or equal to the diameter of the arc, and the angle between the extensions of the two line segments is 90°.

[0023] The identification of doors and numbers in drawings involves first initially identifying and traversing groups of arcs and line segments, then matching them using a door geometric feature library to exclude similar geometric combinations that are not door components, and finally using quantitative criteria to select groups of elements that meet the combination features and mark them as candidate sets of door components. Then, using the spatial coordinates of the candidate set elements as the center, a search range with a radius of a preset drawing unit is set. Text elements within the search range are retrieved, the text content in the text elements is parsed, and the door numbering rules are matched. The successfully matched text numbers are bound to the candidate set elements to confirm the identity and unique number of the door components.

[0024] Analyze the double-line features of the wall drawing, design two recognition methods, and use the two recognition methods to identify the wall. The first identification method is double-line group identification. When the target group data is a double-line group, first determine whether the wall feature is a parallel double-line group. If so, then iterate through the parallel segments in the segment group, filter out double-line groups whose parallelism error is less than the threshold, whose line group width conforms to the thickness range of the building wall, and whose double-line group length is greater than the threshold, and mark them as walls.

[0025] When the first identification method fails to identify the wall, the second identification method is used: Based on common wall layers, calculate the connectivity and enclosure between walls, determine whether the target group data is a wall, and if it meets the wall characteristics, mark it as a wall.

[0026] Based on window numbering and parallel line features, combined with spatial constraints and numbering semantics, windows and railings are identified. The starting and ending points of the target group of parallel lines can only be used to determine whether the group of parallel lines is a window or a railing when the identified walls are connected.

[0027] Windows and railings are distinguished by numbers. High-quality architectural construction drawings will mark the window number next to each window. Window numbers indicate that they are windows, and the rest are railings.

[0028] Window and railing identification includes: Traverse the line segment group, filter out the parallel line group with parallelism error less than the threshold and spacing within the width range of window or railing, and mark it as window or railing candidate set; Select parallel line groups from the candidate set, determine whether their start and end points are connected to the identified wall components, retain parallel line groups whose start and end points are both connected to the wall, and exclude independent decorative parallel lines. Centered on the preserved parallel line group, set the search range of the preset drawing unit, retrieve text elements and parse the text content, and match the window numbering rules; Mark the successfully matched parallel line groups as window components and bind them with the corresponding window numbers; Parallel line groups that do not match window numbers are marked as railings after confirmation based on building design codes and the surrounding spatial environment.

[0029] The identified doors, walls, windows, and railings are magnified, enclosed, calculated, and restored to form the building rooms.

[0030] The enclosure of a building's rooms includes: The vector primitives of all identified components are enlarged proportionally according to a preset ratio, with the enlargement center being the geometric center of the component. The enlargement process fills the tiny gaps between components. Based on the magnified component vector data, the boundary tracking algorithm is used to traverse the outer boundaries of walls, windows and railings, identify the closed areas enclosed by these components, filter out areas with a closed area greater than a threshold, and mark these areas as candidate areas for the room. The candidate room area is scaled down proportionally to the inverse of the magnification ratio to restore the actual size of the component. Combined with the text elements in the room, the room's functional information is bound to complete the room recognition. Example

[0031] First, the system automatically recognizes the CAD drawings of the building construction. In the CAD drawings that need to be recognized, use a rectangular frame to select the floor plan that needs to be recognized, enter the command to call the recognition function, select the drawn rectangular frame, and after confirmation, you will get the recognition result, which recognizes all rooms.

[0032] like Figure 2 As shown, Figure 2 To create a DWG drawing, the first step is to extract the data. Based on the requirements for drawing recognition, all elements that could potentially be used to draw components such as walls, doors, and windows need to be extracted. Text and fill elements also need to be extracted. Text is used to identify room functions, door and window numbers, etc., while fill elements are used to assist in identifying shear wall columns. Figure 2 The excessive number of door and window numbers in the system would severely impact display quality. Figure 2 Other content is being displayed, so the door and window numbers are hidden, but... Figure 4 , 5 Door and window numbers can be seen in both numbers 6 and 7.

[0033] After obtaining all the necessary data, it needs to be grouped based on data type and attributes. After grouping, schematic diagrams of walls, doors, and windows are obtained. The walls are shown in the diagram. Figure 3 As shown, the door is as follows Figure 4 As shown.

[0034] After the data is grouped, each group is analyzed.

[0035] Figure 4 The features of the central door are relatively obvious; it appears as a combination of a 90-degree arc and line segments in the drawing, which can generally be considered a door. Combined with the extracted text, most of the doors in the drawing and their corresponding numbers can be identified.

[0036] A combination of a 90-degree arc and a line segment is considered a door.

[0037] Figure 3 Identifying the central wall is the most difficult part because the wall here is a wall in a broad sense. It may be represented in many ways in the drawings, including brick walls, infill walls, lightweight partition walls, and so on. Architects will use multiple ways to represent these walls. After analyzing the characteristics of the walls, we mainly identify the walls through two identification methods.

[0038] According to the working habits and standards of architectural construction drawings, the width of most walls is 50, 100, 200, 250, etc. When a certain set of data is mostly double lines and the line width meets the above width, this set of data is considered to be a wall line. After testing a large number of drawings, the accuracy rate of obtaining wall lines in this way is over 98%.

[0039] The judgment of the second identification method in the above embodiment 1 includes: determining the endpoints of the two line segments of the identified door component, taking the endpoints of the two line segments as the starting point, extending the line along the perpendicular direction of the door opening direction, and searching for line segments on the extension line that meet the characteristics of parallel double line groups; Connect the line segment group that meets the characteristics of parallel double line group with the door component to confirm that it is an extension of the wall. Complete the wall recognition result according to the spatial correlation characteristics of the door component.

[0040] Data sets that meet the above conditions can be identified as wall lines.

[0041] In addition, some walls in the diagram do not meet the above requirements. The second method is used for identification. In process three, the door has already been identified. In architectural construction drawings, a door cannot exist alone; it must be connected to a wall. Following this logic, the remaining walls can be identified through the double-line group features on both sides of the door. This method is mainly used for walls with non-standard widths or curtain walls, which cannot be identified by the first method.

[0042] The identification of windows and railings is based on two conditions: window number and parallel line features.

[0043] Windows and railings are also represented by parallel lines in the drawings, for... Figure 2 The parallel lines in the window and railing are blue, light blue, and dark brown, but unlike the walls, the parallel lines of the window and railing are spaced closer together, usually less than 50mm. In addition, the beginning and end points of this set of parallel lines must connect with the wall identified in process four, so that it can be determined whether the set of parallel lines is a window or a railing.

[0044] Windows and railings are distinguished by numbering. A high-quality architectural drawing will mark the window number next to each window. Based on this logic, windows with the corresponding window number are considered windows, and the rest are railings.

[0045] Windows with windows are numbered – windows; windows without windows are numbered – railings.

[0046] With walls, doors, windows, and railings, we can use these components to enclose building rooms, such as bedrooms, balconies, and corridors. This step of the algorithm is relatively simple, and we only need to pay attention to the tolerance issue: the construction drawings may have some small gaps in the components that are not closed. In order to prevent the enclosure of the space from failing, we need to enlarge all the components first. After the room is enclosed, we can restore it according to the enlarged size to get the correct room.

[0047] This technological logic makes it a truly universal architectural plan vector graphics recognition solution for the industry. It can not only adapt to the drawing systems of design units of different sizes in China—whether it is the standardized drawing process of large state-owned design institutes or the flexible drawing mode of small and medium-sized private design institutes; it can also be compatible with the drawing habits of designers of different ages—it conforms to the personalized annotation methods of senior designers based on experience, and also adapts to the standardized drawing style of young designers relying on software templates.

[0048] In practical applications, this solution can be widely used in multiple stages of building construction: in the design review stage, it can quickly identify component conflicts and dimensional deviations in drawings; in the quantity calculation stage, it can accurately count the quantity and specifications of components such as walls, doors, windows, beams and columns; in the digital modeling stage, it can automatically convert two-dimensional drawing components into three-dimensional model elements, greatly reducing manual operation costs and improving the digital efficiency of the entire life cycle of building construction. Example

[0049] After the building rooms are enclosed and identified, a consistency check between the drawings and the model is required. The specific process includes: First, read the IFC model and DWG file of the corresponding individual building, and then open the corresponding file in CAD. Then read the architectural floor plan scope and name from the DWG drawing and automatically match them with the floors in the IFC; Then, a DWG / IFC consistency comparison is performed. The consistency comparison includes: DWG plane recognition, automatic alignment of DWG and IFC, and automatic comparison of DWG / IFC, and the corresponding results are displayed. Users can click on the DWG / IFC comparison results to view them. The colored fill in the image is the planar projection of the corresponding floor in the model. The projection will automatically align with the DWG plane. If the error is due to a mistake, a "locate" button is provided after each question to view the location of the error. Figure 5 and Figure 6 As shown, Figure 5 and Figure 6 The model is missing fire doors, which makes the fire door numbers mismatched and prevents it from passing the DWG / IFC consistency check. The user needs to modify the model to make it consistent with the DWG drawings in order to pass the DWG / IFC consistency check. The appeal function is also available. If the appeal is successful, the appealed issues and related materials will be uploaded to the backend through the client for verification.

[0050] After DWG / IFC conformance comparison, DWG / PDF comparison is then performed. The drawing / model consistency check is considered passed only if both the DWG / IFC consistency comparison and the DWG / PDF consistency comparison pass.

[0051] The consistency of the three documents was ensured by comparing them twice, once with DWG / IFC and once with DWG / PDF. Example

[0052] From a hardware perspective, this application provides an embodiment of an electronic device containing all or part of an automatic parsing and recognition method for architectural DWG drawings based on vector graphic feature analysis. The electronic device includes a service processor and a distributed memory. The service processor is connected to the memory, and the distributed memory stores a service self-management program configured to store machine-readable instructions. The service processor executes the service self-management program, and the instructions, when executed by the processor, implement the automatic parsing and recognition method for architectural DWG drawings based on vector graphic feature analysis as described above. Example

[0053] The embodiments of this application also provide a computer-readable storage medium capable of implementing the automatic parsing and recognition method for architectural DWG drawings based on vector graphic feature analysis, in which the execution subject is a server or client as described in the above embodiments. The computer-readable storage medium stores a computer program that, when executed by a processor, implements all the contents of the automatic parsing and recognition method for architectural DWG drawings based on vector graphic feature analysis, in which the execution subject is a server or client as described in the above embodiments.

[0054] The embodiments described above are merely preferred embodiments of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications, improvements, and substitutions without departing from the concept of this application, and these all fall within the protection scope of this application.

Claims

1. An automatic parsing and recognition method for architectural DWG drawings based on vector graphic feature analysis, characterized in that, Includes the following steps: Based on the requirements for identifying drawings, data is extracted from DWG drawings using vector primitive analysis and attribute association. The process of extracting data from DWG drawings includes: First, extract the vector primitives from the DWG drawing, then read the underlying vector data of the drawing; Extract all text objects from the DWG drawing and establish a text-graphic mapping foundation based on the spatial relationship between text coordinates and graphic element coordinates; Extract the boundary vector information, fill pattern code, and fill ratio of the filled area to distinguish between structural fill and decorative fill. Determine the type of auxiliary component by the spatial range of the filled area and its connection with surrounding graphic elements. Based on the extracted data, three-dimensional features are obtained, and the data is grouped into three levels according to the three-dimensional features. Each group of data is then analyzed. The three-dimensional features are data type, geometric attributes, and spatial correlation. The grouping rules for the three-level grouping include: First, divide the elements into three main categories based on data type: graphic elements, text elements, and fill elements, thus achieving the first level of grouping. Then, the graphic element groups are clustered according to their geometric attributes to achieve the second level of grouping; Finally, the distance and angle relationship between different primitives are calculated, and primitives that are spatially adjacent and have matching attributes are grouped into the same analysis group to achieve the third level of grouping; A dual recognition mechanism is established to identify the drawing features of the door and, in combination with the extracted text, to identify the door in the drawing and its corresponding number. Analyze the double-line features of the wall drawing, design two recognition methods, and use the two recognition methods to identify the wall. Based on window numbering and parallel line features, combined with spatial constraints and numbering semantics, windows and railings are identified. The identified doors, walls, windows, and railings are magnified, enclosed, calculated, and restored to form the building rooms.

2. The automatic parsing and recognition method for architectural DWG drawings based on vector graphic feature analysis according to claim 1, characterized in that, The identification of doors and numbers in the drawings involves first initially identifying and traversing groups of arcs and line segments, then matching them using a door geometric feature library to exclude similar geometric combinations that are not door components, and finally using quantitative criteria to select primitive groups that meet the combination features and mark them as candidate sets of door components. Then, using the spatial coordinates of the candidate set elements as the center, a search range with a radius of a preset drawing unit is set. Text elements within the search range are retrieved, the text content in the text elements is parsed, and the door numbering rules are matched. The successfully matched text numbers are bound to the candidate set elements to confirm the identity and unique number of the door components.

3. The automatic parsing and recognition method for architectural DWG drawings based on vector graphic feature analysis according to claim 1, characterized in that, The two identification methods for identifying walls include: When the target group data is a double-line group, first determine whether the wall feature is a parallel double-line group; If so, use the first identification method: Traverse the parallel line segments in the line segment group, filter out double line groups with a parallelism error of less than the threshold, a line group width that meets the thickness range of the building wall, and a double line group length that is greater than the threshold, and mark them as walls. If not, use the second identification method: Based on common wall layers, calculate the connectivity and enclosure between walls, determine whether the target group data is a wall, and if it meets the wall characteristics, mark it as a wall.

4. The automatic parsing and recognition method for architectural DWG drawings based on vector graphic feature analysis according to claim 1, characterized in that, The identification of the windows and railings includes: Traverse the line segment group, filter out the parallel line group with parallelism error less than the threshold and spacing within the width range of window or railing, and mark it as window or railing candidate set; Select parallel line groups from the candidate set, determine whether their start and end points are connected to the identified wall components, retain parallel line groups whose start and end points are both connected to the wall, and exclude independent decorative parallel lines. Centered on the preserved parallel line group, set the search range of the preset drawing unit, retrieve text elements and parse the text content, and match the window numbering rules; Mark the successfully matched parallel line groups as window components and bind them with the corresponding window numbers; Parallel line groups that do not match window numbers are marked as railings after confirmation based on building design specifications and the surrounding spatial environment.

5. The automatic parsing and recognition method for architectural DWG drawings based on vector graphic feature analysis according to claim 1, characterized in that, The enclosure of a building's rooms includes: The vector primitives of all identified components are enlarged proportionally according to a preset ratio, with the center of the enlargement being the geometric center of the component. The enlargement process fills the tiny gaps between components. Based on the magnified component vector data, the boundary tracking algorithm is used to traverse the outer boundaries of walls, windows and railings, identify the closed areas enclosed by these components, filter out areas with a closed area greater than a threshold, and mark these areas as candidate areas for the room. The candidate room area is scaled down proportionally to the inverse of the magnification ratio to restore the actual size of the component. Combined with the text elements in the room, the room's functional information is bound to complete the room recognition.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the content of the automatic parsing and recognition method for architectural DWG drawings based on vector graphic feature analysis as described in claim 1.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the automatic parsing and recognition method for architectural DWG drawings based on vector graphic feature analysis as described in claim 1.

Images