Structural CAD planar graph analysis method and device based on geometric topological constraints

Abstract

The present application relates to the technical field of intelligent identification, in particular to a structure CAD planar graph analysis method and device based on geometric topological constraint, the method comprising: obtaining original CAD graph data; determining a to-be-identified structure planar graph according to the original CAD graph data, and performing constraint analysis on the to-be-identified structure planar graph based on the spatial correspondence relationship between axis numbers and axis lines and the topological relationship of axis nets to obtain an axis net identification result, and determining a single planar graph effective range according to the axis net identification result; obtaining column identification results and beam identification results based on the axis net intersection position priori according to the single planar graph effective range and the axis net identification result, and obtaining a structure CAD planar graph analysis result based on the single planar graph effective range, the column identification results and the beam identification results. The problems that the drawing analysis is limited by drawing habits and layer rules and the complex drawing recognition effect is poor are solved, and the adaptability and efficiency of the structure planar graph analysis are improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent recognition technology, and in particular to a method and apparatus for analyzing structural CAD plan drawings based on geometric topological constraints. Background Technology

[0002] In the field of architectural engineering, CAD structural plans are crucial for design, construction, and cost estimation. To overcome the inefficiencies and errors of manual drawing interpretation, the industry is gradually developing research on intelligent CAD drawing analysis technology. This technology aims to automatically extract structural component information from drawings and promote the digitalization of the entire architectural engineering process.

[0003] The relevant technologies mainly include three types: layer and element style filtering, local geometric feature discrimination, and deep learning model recognition. These technologies can complete basic component recognition work under conditions of standardized drawings and simple layouts.

[0004] However, the various recognition methods of related technologies have strict requirements on drafting standards and drawing formats, making them unsuitable for the different drawing habits of different units; they lack geometric topological constraints between axes and beams and columns, resulting in insufficient accuracy and completeness in complex scenarios such as diverse component shapes, line occlusion, and mixed layout of multiple plans; at the same time, the recognition results are only scattered element information and cannot generate structured calculable data containing component connection relationships, dimensions, and reinforcement parameters, which cannot directly serve subsequent work such as structural modeling and quantity statistics, thus having significant application limitations that urgently need to be addressed. Summary of the Invention

[0005] This invention provides a method and apparatus for analyzing structural CAD plan views based on geometric topological constraints, in order to solve the problem that drawing analysis is limited by drawing habits and layer rules, resulting in poor recognition of complex drawings, and to improve the adaptability and efficiency of structural plan view analysis.

[0006] To achieve the above objectives, a first aspect of the present invention provides a method for analyzing structural CAD plan views based on geometric topological constraints, comprising the following steps: acquiring original CAD drawing data; determining the structural plan view to be identified based on the original CAD drawing data, and performing constraint analysis on the structural plan view to be identified based on the spatial correspondence between axis numbers and axes and the axis grid topological relationship to obtain axis grid identification results, and determining the effective range of a single plan view based on the axis grid identification results; obtaining column identification results and beam identification results based on the prior knowledge of the axis grid intersection positions based on the effective range of the single plan view and the axis grid identification results, and obtaining the structural CAD plan view analysis result based on the effective range of the single plan view, the column identification results, and the beam identification results.

[0007] Furthermore, in some embodiments, determining the structural plan view to be identified based on the original CAD drawing data includes: standardizing the original CAD drawing data to obtain a standardized CAD element set; identifying the frame area and component attribute information in the standardized CAD element set, and performing region discrimination on the frame area to obtain a structural plan view region discrimination result; and determining the structural plan view to be identified based on the structural plan view region discrimination result.

[0008] Furthermore, in some embodiments, the standardization process of the original CAD drawing data to obtain a standardized CAD primitive set includes: obtaining original primitive data in the original CAD drawing data; expanding all block references in the original primitive data to obtain a block virtual expansion result; and aligning the coordinates of the internal entities of the blocks in the original primitive data according to the block virtual expansion result to obtain a standardized CAD primitive set.

[0009] Further, in some embodiments, the step of identifying the frame region and component attribute information in the standardized CAD element set, and performing region discrimination on the frame region to obtain the structural plan view region discrimination result, includes: determining the candidate frame spatial boundary of each standardized CAD element data based on the standardized CAD element set; dividing each standardized CAD element data into the corresponding target frame region according to a preset element inclusion relationship; taking the target frame region that satisfies the preset plan view characteristics as the valid structural plan view region, and deleting the target frame region that does not satisfy the preset plan view characteristics, thereby obtaining the structural plan view region discrimination result.

[0010] Further, in some embodiments, obtaining the grid identification result based on constraint analysis of the planar diagram of the structure to be identified includes: acquiring all candidate circular primitives and candidate linear primitives in the planar diagram of the structure to be identified; determining whether there is text content that conforms to a preset grid number rule within each candidate circular primitive; if there is no text content that conforms to the preset grid number rule, skipping the candidate circular primitive; otherwise, generating a grid number object based on the candidate circular primitive and the text content corresponding to the candidate circular primitive; merging the candidate linear primitives that satisfy a preset merging rule to generate candidate long lines; determining the spatial correspondence between each grid number object and each candidate long line; taking the candidate long line that satisfies the preset grid number object correspondence condition as the axis line corresponding to the grid number object; and obtaining the grid identification result based on the grid number object and the axis line corresponding to the grid number object.

[0011] Furthermore, in some embodiments, determining the effective range of a single planar map based on the grid identification result includes: constructing a grid topology map based on the grid identification result, calculating the connected components of the grid topology map, obtaining the connected component calculation result, generating multiple grid units based on the connected component calculation result; obtaining the center coordinates of all grid objects in each grid unit, generating the minimum envelope region corresponding to each grid object based on the center coordinates of the grid objects; expanding the minimum envelope region according to a preset expansion distance to obtain the effective range of a single planar map.

[0012] Further, in some embodiments, obtaining column identification results and beam identification results based on the effective range of the single plan view includes: calculating the coordinates of the grid intersection points based on the effective range of the single plan view; determining multiple initial column candidate elements within a preset range of the grid intersection point coordinates; selecting column candidate elements that satisfy preset column identification constraints as target column candidate elements; obtaining the column identification results based on the target column candidate elements; acquiring all initial candidate beam lines in the candidate linear elements within the effective range of the single plan view after excluding frame lines, axis lines, axis number elements, and identified column elements; using the column objects in the column identification results as beam identification seeds; searching for pairs of parallel candidate lines around the column objects; filtering the initial candidate beam lines based on directional consistency, line spacing, projection overlap, and proximity, crossing, or connection with column boundaries; and obtaining the beam identification results based on beam line endpoint propagation, collinear continuation, nearby parallel support, missing beam edge filling, and pseudo-beam filtering.

[0013] Further, in some embodiments, obtaining the structural CAD plan view parsing result based on the effective range of the single plan view, the column identification result, and the beam identification result includes: mapping each column object in the column identification result to a target structural diagram node; calculating the beam centerline based on the beam edge line in the beam identification result, matching both ends of the beam centerline to the corresponding target structural diagram node to obtain the target structural diagram edge; identifying the beam number, beam cross-sectional dimensions, and reinforcement annotation text based on the effective range of the single plan view, and matching the beam number, beam width, beam height, and reinforcement information to the corresponding target structural diagram edge based on the distance between the text bounding box and the beam centerline, the projection position of the text in the beam direction, the spatial relationship between the text and the beam edge buffer, and the consistency of the beam number; matching the column number and column dimension information to the corresponding target structural diagram edge; and matching the column number and column dimension information to the corresponding target structural diagram node; obtaining the structural CAD plan view parsing result based on the target structural diagram node, the target structural diagram edge, and the effective range of the single plan view.

[0014] The structural CAD plan view parsing method provided by the present invention first reads the original CAD drawing data and determines the structural plan view to be identified from it. It then obtains the grid information through the spatial correspondence between axis numbers and axes, and the grid topology, and determines the effective recognition area of ​​a single plan view accordingly. Next, within the defined area, it identifies column members based on the prior location of grid intersections. Using column members as seeds, it identifies beam members by combining beam-column connection relationships and beam line propagation constraints. After integrating the plan range, beam-column identification information, and member attribute information, it finally obtains the complete parsing result of the structural CAD plan view. This method solves the problem of drawing parsing being limited by drawing habits and layer rules, resulting in poor recognition of complex drawings, and improves the adaptability and efficiency of structural plan view parsing.

[0015] To achieve the above objectives, a second aspect of the present invention provides a structural CAD plan view parsing device based on geometric topological constraints, comprising: an acquisition module for acquiring original CAD drawing data; a determination module for determining a structural plan view to be identified based on the original CAD drawing data, and performing constraint analysis on the structural plan view to be identified based on the spatial correspondence between axis numbers and axes and the axis grid topology relationship to obtain axis grid identification results, and determining the effective range of a single plan view based on the axis grid identification results; and a parsing module for obtaining column identification results based on the effective range of the single plan view and the axis grid identification results, based on the prior knowledge of the axis grid intersection positions, and using the column objects in the column identification results as beam identification seeds, combining paired parallel beam edges, beam-column connection relationships, beam line endpoint propagation, collinear continuation, missing beam edge filling, and pseudo-beam filtering constraints to obtain beam identification results, and obtaining structural CAD plan view parsing results based on the effective range of the single plan view, the column identification results, and the beam identification results.

[0016] Furthermore, in some embodiments, the determining module is specifically used for: standardizing the original CAD drawing data to obtain a standardized CAD element set; identifying the frame area and component attribute information in the standardized CAD element set, and performing region discrimination on the frame area to obtain a structural plan area discrimination result; and determining the structural plan to be identified based on the structural plan area discrimination result.

[0017] Furthermore, in some embodiments, the determining module is also used to: obtain original graphic element data in the original CAD drawing data; expand all block references in the original graphic element data to obtain a block virtual expansion result; and perform coordinate alignment of the internal entities of the blocks in the original graphic element data according to the block virtual expansion result to obtain a standardized CAD element set.

[0018] Furthermore, in some embodiments, the determining module is further configured to: determine the candidate frame spatial boundary of each standardized CAD drawing element based on the standardized CAD drawing element set; divide each standardized CAD drawing element into a corresponding target frame region according to a preset element inclusion relationship; take the target frame region that satisfies the preset planar diagram features as a valid structural planar diagram region, and delete the target frame region that does not satisfy the preset planar diagram features, thereby obtaining the structural planar diagram region discrimination result.

[0019] Furthermore, in some embodiments, the determining module is further configured to: acquire all candidate circular primitives and candidate linear primitives in the planar diagram of the structure to be identified; determine whether there is text content that conforms to a preset axis number rule within each candidate circular primitive; if there is no text content that conforms to the preset axis number rule, skip the candidate circular primitive; otherwise, generate an axis number object based on the candidate circular primitive and the text content corresponding to the candidate circular primitive; merge the candidate linear primitives that satisfy a preset merging rule to generate candidate long lines; determine the spatial correspondence between each axis number object and each candidate long line; take the candidate long line that satisfies the preset axis number object correspondence condition as the axis line corresponding to the axis number object; and obtain the axis grid recognition result based on the axis number object and the axis line corresponding to the axis number object.

[0020] Furthermore, in some embodiments, the determining module is also used to: construct an axis grid topology map based on the axis grid identification result, calculate the connected components of the axis grid topology map, obtain the connected component calculation result, generate multiple axis grid units according to the connected component calculation result; obtain the center coordinates of all axis number objects in each axis grid unit, generate the minimum envelope region corresponding to each axis number object according to the center coordinates of the axis number objects; expand the minimum envelope region according to a preset expansion distance to obtain the effective range of a single planar map.

[0021] Further, in some embodiments, the parsing module is specifically used for: calculating the coordinates of the grid intersection points based on the effective range of the single plan view; determining multiple initial column candidate elements within a preset range of the grid intersection point coordinates; selecting column candidate elements that satisfy preset column recognition constraints as target column candidate elements; obtaining the column recognition result based on the target column candidate elements; acquiring all initial candidate beam lines in the candidate linear elements within the effective range of the single plan view after excluding frame lines, axis lines, axis number elements, and already recognized column elements; using the column objects in the column recognition result as beam recognition seeds; searching for pairs of parallel candidate lines around the column objects; filtering the initial candidate beam lines based on directional consistency, line spacing, projection overlap, and proximity, crossing, or connection with column boundaries; and obtaining the beam recognition result based on beam line endpoint propagation, collinear continuation, nearby parallel support, missing beam edge filling, and pseudo-beam filtering.

[0022] Furthermore, in some embodiments, the parsing module is also used to: map each column object in the column identification result to a target structural diagram node; calculate the beam centerline based on the beam edge line in the beam identification result, and match both ends of the beam centerline to the corresponding target structural diagram node to obtain the target structural diagram edge; based on the effective range of the single plan view, identify the beam number, beam cross-sectional dimensions, and reinforcement annotation text, and match the beam number, beam width, beam height, and reinforcement information to the corresponding target structural diagram edge based on the distance between the text bounding box and the beam centerline, the projection position of the text in the beam direction, the spatial relationship between the text and the beam edge buffer, and the consistency of the beam number, and match the column number and column dimension information to the corresponding target structural diagram node; and obtain the structural CAD plan view parsing result based on the target structural diagram node, the target structural diagram edge, and the effective range of the single plan view.

[0023] The structural CAD plan view parsing device based on geometric topological constraints provided in this embodiment of the invention first reads the original CAD drawing data and determines the structural plan view to be identified. It obtains the grid information through the spatial correspondence between axis numbers and axes and the grid topology relationship, and determines the effective recognition area of ​​a single plan view accordingly. Then, within the defined area, it identifies column members based on the a priori position of grid intersections. Using column members as seeds, it identifies beam members by combining beam-column connection relationships and beam line propagation constraints. After integrating the plan range, beam-column identification information, and member attribute information, it finally obtains the complete parsing result of the structural CAD plan view. This solves the problem that drawing parsing is limited by drawing habits and layer rules, resulting in poor recognition of complex drawings, and improves the adaptability and efficiency of structural plan view parsing.

[0024] To achieve the above objectives, a third aspect of the present invention provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the structural CAD plan view analysis method based on geometric topological constraints as described in the above embodiments.

[0025] Therefore, the present invention has the following beneficial effects: (1) The present invention identifies the geometric shape, text features, spatial position relationship and structural topological constraints of the original CAD drawing elements, reduces the dependence on CAD drawing standards, reduces the identification failure caused by non-standard layer naming, inconsistent color settings or mixed line types, and improves the adaptability under different drawing sources.

[0026] (2) The present invention uses the axis grid topology to distinguish different plan views, reduces the workload of manual selection of the range, and reduces the risk of multiple plan views being mistakenly merged or components being incorrectly assigned, thereby improving the ability of automatic positioning and splitting of structural plan views in complex CAD drawings.

[0027] (3) The present invention uses the identified column components as seeds and combines them with constraints such as paired parallel beam edge lines, beam-column connection relationship, beam line endpoint propagation, collinear continuation, nearby parallel support and missing beam edge replenishment to form a beam line network, thereby improving the completeness of beam component identification and reducing non-beam line interference.

[0028] (4) The present invention converts the identified column components into structural diagram nodes and the beam components into structural diagram edges to form structural diagram data. In other words, it converts discrete data into computable data with the semantics of beam and column components and connection relationships, which facilitates structural modeling, component verification and engineering quantity calculation. Attached Figure Description

[0029] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart of a structural CAD plan view analysis method based on geometric topological constraints provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of axis number constraint axis and axis grid connection component division according to a specific embodiment of the present invention; Figure 3 This is a schematic diagram of the grid, column, and beam identification process according to a specific embodiment of the present invention; Figure 4 This is a schematic diagram of a graph data structure provided according to a specific embodiment of the present invention; Figure 5 This is a schematic diagram of the overall process of the intelligent analysis method for structural CAD plan drawings provided according to a specific embodiment of the present invention; Figure 6 This is a block diagram of a structural CAD plan view analysis device based on geometric topological constraints provided in an embodiment of the present invention; Figure 7 This is a schematic diagram of the structure of an electronic device provided according to an embodiment of the present invention. Detailed Implementation

[0030] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0031] The following describes the method and apparatus for analyzing structural CAD plan views based on geometric topological constraints according to embodiments of the present invention with reference to the accompanying drawings. First, the method for analyzing structural CAD plan views based on geometric topological constraints according to embodiments of the present invention will be described with reference to the accompanying drawings.

[0032] Figure 1 This is a flowchart of a structural CAD plan view analysis method based on geometric topological constraints provided according to an embodiment of the present invention.

[0033] like Figure 1 As shown, the method for analyzing structural CAD plan views based on geometric topological constraints includes the following steps: In step S101, the original CAD drawing data is obtained.

[0034] In step S102, the plan view of the structure to be identified is determined based on the original CAD drawing data, and constraint analysis is performed on the plan view of the structure to be identified based on the spatial correspondence between axis numbers and axes and the axis grid topology relationship to obtain the axis grid identification result. The effective range of a single plan view is determined based on the axis grid identification result.

[0035] Among them, the structural plan view to be identified refers to the effective drawing area containing the architectural structure layout obtained from all CAD elements and drawing frame areas; constraint analysis refers to the analysis method of identification and judgment based on element shape, relative position, and combination relationship; grid recognition result refers to the overall information of the grid composed of axis number and axis line; and the effective range of a single plan view refers to the boundary of the polygon area determined by the grid recognition result.

[0036] In some embodiments, determining the structural plan view to be identified based on the original CAD drawing data includes: standardizing the original CAD drawing data to obtain a standardized CAD element set; identifying the frame area and component attribute information in the standardized CAD element set, and performing region discrimination on the frame area to obtain the structural plan view region discrimination result; and determining the structural plan view to be identified based on the structural plan view region discrimination result.

[0037] Specifically, this invention takes a DWG or DXF format structural construction drawing file (i.e., the original CAD drawing data) to be identified as input, performs standardization processing on it, converts or reads it into a unified set of CAD elements, identifies the drawing frame area in the standardized CAD element set, and splits the original CAD file into one or more independent drawing frame areas. Thus, it can automatically separate different drawing areas in CAD files containing multiple drawing frames, multiple drawing contents, or multiple construction drawings, avoiding the mixing of axes, beams, columns, or text in different drawing frames during subsequent identification.

[0038] In some embodiments, the original CAD drawing data is standardized to obtain a standardized CAD primitive set, including: obtaining the original primitive data in the original CAD drawing data; expanding all block references in the original primitive data to obtain a block virtual expansion result; and aligning the internal entities of the blocks in the original primitive data according to the block virtual expansion result to obtain a standardized CAD primitive set.

[0039] As one possible implementation, embodiments of the present invention expand block references, perform coordinate transformations on entities such as line segments, text, and fills within the block, and preserve their spatial positional relationships in the original drawing. For each CAD element, information such as its geometric type, bounding box, handle, insertion point or control point, text content, layer, color, and line type is uniformly recorded to obtain a standardized set of CAD elements. The standardized set of CAD elements includes line segments, polylines, circles, text, multiline text, attribute text, fills, block references, and virtual elements after the block references are expanded.

[0040] It should be noted that layers, colors, and line types are only used as auxiliary information or output debugging information, and are not necessary for the identification of structural objects such as beams, columns, and axes.

[0041] Furthermore, in some embodiments, identifying the frame region and component attribute information in the standardized CAD element set, and performing region discrimination on the frame region to obtain the structural plan view region discrimination result, includes: determining the candidate frame spatial boundary of each standardized CAD element data based on the standardized CAD element set; dividing each standardized CAD element data into the corresponding target frame region according to the preset element inclusion relationship; taking the target frame region that meets the preset characteristics as the valid structural plan view region, and deleting the target frame region that does not meet the preset characteristics, to obtain the structural plan view region discrimination result.

[0042] As one possible implementation method, embodiments of the present invention detect large-size rectangular closed outlines, inner and outer frame combinations, title areas, and frame lines in drawings to determine the spatial range of candidate frames. When multiple frames exist in the same CAD file, the elements are assigned to the corresponding frames based on the center point of each element's bounding box, the element's inclusion relationship with the frame, and the frame's area. If the same element falls into multiple candidate frames, it is preferentially assigned to the frame with the smaller area and clearer inclusion relationship. Then, the text, number of elements, number of candidate axis numbers, line segment distribution characteristics, and title keywords within each frame area are analyzed to determine whether the frame belongs to a structural plan, a construction plan, or a layout plan. This allows for the filtering of non-structural plan areas from complex CAD drawing content, reducing interference from details, tables, explanatory text, and partial construction drawings on the identification of grid lines, columns, and beams, and obtaining structural plan area discrimination results.

[0043] It should be noted that drawing areas with extensive structural specifications, beam reinforcement details, column reinforcement details, joint details, stair details, foundation details, or tables or explanatory text lacking planar grid features will not be included in the beam-column identification process. Drawing areas with features such as axis numbers, axes, and beam-column layouts will be treated as structural plan views to be identified and processed further. Further, in some embodiments, the grid identification result is obtained by constraining the planar diagram of the structure to be identified, including: acquiring all candidate circular primitives and candidate linear primitives in the planar diagram of the structure to be identified; determining whether there is text content that conforms to the preset grid number rules in each candidate circular primitive; if there is no text content that conforms to the preset grid number rules, the candidate circular primitive is skipped; otherwise, a grid number object is generated based on the candidate circular primitive and the text content corresponding to the candidate circular primitive; merging the candidate linear primitives that meet the preset merging rules to generate candidate long lines; determining the spatial correspondence between each grid number object and each candidate long line; and taking the candidate long line that meets the preset grid number object correspondence conditions as the axis line corresponding to the grid number object; and obtaining the grid identification result based on the grid number object and the axis line corresponding to the grid number object.

[0044] As one possible implementation, this embodiment of the invention first traverses the circular primitives, extracts their center coordinates and radius, and searches for text, multi-line text, or attribute text near the center. The searched text content is then standardized, including removing spaces, unifying full-width and half-width characters, unifying capitalization, and merging scattered text. If short text conforming to axis number rules exists inside or near the circular primitive, the circular primitive and the short text are combined and identified as an axis number object. The axis number rules include numeric axis numbers, letter axis numbers, and composite axis numbers composed of numbers, letters, slashes, etc. Circular primitives containing descriptive text, excessively long text, multi-line complex text, or those that do not conform to axis number rules are not considered axis number objects. Thus, the axis number is identified through the combination relationship between the circular boundary and the short text inside the circle.

[0045] Furthermore, after identifying the axis number, candidate linear primitives are extracted from line segments, polylines, and block reference unfolded line segments. Based on the line segment direction, endpoint distance, collinearity, and gap size, line segments with similar directions and spatial continuity are merged into candidate long lines. For each axis number object, the distance between its center and the endpoint or extension of the candidate long line is calculated. The candidate long line direction, length, and spatial correspondence with the axis number are then used for judgment. If the endpoint of the candidate long line is close to the axis number (e.g., the distance between the endpoint of the candidate long line and the axis number is less than or equal to 1 cm), or the extension of the candidate long line passes near the axis number, and the candidate long line meets the length and direction requirements, it is identified as the axis corresponding to that axis number. When the same or corresponding axis numbers exist at both ends of the same axis, the axis is further confirmed using the axis number relationship at both ends. When multiple candidate line segments are collinear and have the same direction as the confirmed axis, they are merged into the same axis object. Thus, using the axis number as a semantic constraint to identify the axis can avoid misidentifying ordinary long lines, annotation lines, drawing frame lines, or component edges as axes.

[0046] Furthermore, the identified axis numbers and axes are constructed into a grid diagram data. In this grid diagram, the axis number, axis endpoints, and axis intersections serve as nodes, the axes as edges, and the axis number text, coordinates, direction, and length as attributes of the nodes or edges. If two axis numbers are located at opposite ends of the same axis, or if multiple axes have intersecting connections, corresponding connections are established in the grid diagram. Connected components are calculated based on the connectivity of the grid diagram. For example, a complete structural plan view corresponds to a larger-scale grid connected component with complete horizontal and vertical axes; detailed views, local annotations, or noisy axes correspond to smaller-scale or structurally incomplete connected components. When multiple structural plan views exist within the same frame, different structural plan views typically correspond to different grid connected components. By comparing the number of axis numbers, the number of axes, the completeness of horizontal and vertical axes, and the spatial range in the connected components, multiple structural plan views within the same frame can be automatically distinguished. Thus, axis numbers and axes are elevated from isolated primitives to a grid topology, and grid connectivity is used to split multiple plan views. Grid diagram data, i.e., the grid recognition result, is obtained based on the axis number object and the corresponding axis.

[0047] Furthermore, in some embodiments, determining the effective range of a single planar map based on the grid identification results includes: constructing a grid topology map based on the grid identification results, calculating the connected components of the grid topology map, obtaining the connected component calculation results, generating multiple grid units based on the connected component calculation results; obtaining the center coordinates of all grid objects in each grid unit, generating the minimum envelope region corresponding to each grid object based on the center coordinates of the grid objects; expanding the minimum envelope region according to a preset expansion distance to obtain the effective range of a single planar map.

[0048] As one possible approach, for each valid grid connected component, the coordinates of all the center points of the included axis numbers are extracted, and the outer set of axis numbers of the structural plan is determined based on the spatial distribution of the axis numbers. The outer set of axis numbers is used to construct the envelope region of the plan. The envelope region can be represented by the minimum bounding rectangle of the axis number center, the convex hull, or other geometric regions that can enclose the main grid. After obtaining the initial envelope region, the envelope region is expanded outward based on the axis number radius, the possible overhang distance of beams and columns, and the preset expansion distance to obtain the structural plan identification range for element attribution. This avoids missing beams, columns, and local components located outside the outermost axis when only using the axis bounding box to determine the plan range.

[0049] Among them, the effective grid connection component refers to the complete grid unit of a single independent structural plan view, the outer grid number set refers to all the grid numbers that are spatially located on the outermost edge in a single grid unit, the envelope region refers to the closed geometric region generated based on the position of the outer grid number, and the region expansion refers to expanding the initial envelope region outward by combining the grid number size, beam and column overhang characteristics and preset values.

[0050] For example, Figure 2 This is a schematic diagram of axis number constraint axis and axis grid connectivity component division according to a specific embodiment of the present invention, wherein, Figure 2 (a) is a schematic diagram of the candidate long line extension under axis number constraint. Figure 2 (b) is a schematic diagram of the confirmed orthogonal grid identification results. Figure 2 (c) A schematic diagram of dividing the network of axes into multiple planar diagrams, such as Figure 2 As shown in (a), using the circular axis number (A) as a reference, and considering spatial distance and positional relationships, determine whether the candidate long line / extension line (B) is a valid axis (C), as follows. Figure 2 As shown in (b), the axis numbers are distributed at both ends of the axis, and the horizontal and vertical axes intersect to form a complete orthogonal grid, demonstrating the basic framework of the structural plan, as follows. Figure 2 As shown in (c), there are two sets of unconnected grid lines within the same frame. The system distinguishes two independent grid line units based on the grid line connectivity, corresponding to two different structural planar diagrams.

[0051] In step S103, the column identification results and beam identification results are obtained based on the effective range of a single plan view, and the structural CAD plan view analysis results are obtained based on the effective range of the single plan view, the column identification results, and the beam identification results.

[0052] Further, in some embodiments, obtaining column and beam identification results based on the effective range of a single plan view includes: calculating the coordinates of the grid intersection points based on the effective range of the single plan view; determining multiple initial candidate column elements within a preset range of the grid intersection point coordinates; selecting candidate column elements that satisfy preset column identification constraints as target candidate column elements; obtaining column identification results based on the target candidate column elements; acquiring all initial candidate beam lines within the effective range of the single plan view after excluding frame lines, axis lines, axis number elements, and identified column elements; using the column objects in the column identification results as beam identification seeds; searching for paired parallel candidate lines around the column objects; filtering the initial candidate beam lines based on directional consistency, line spacing, projection overlap, and proximity, crossing, or connection with the column boundary; and obtaining beam identification results based on beam line endpoint propagation, collinear continuation, nearby parallel support, missing beam edge filling, and pseudo-beam filtering.

[0053] It should be noted that since structural columns are usually located near the intersection of the grid lines, the grid line intersection is taken as the a priori location where column members may appear.

[0054] As one possible implementation, after determining the scope of the structural plan, this embodiment of the invention calculates the intersection points of the grid based on the identified transverse and longitudinal axes, and removes duplicates from the intersection points that are close to each other. For each grid intersection point, a local search window is constructed around it, and the primitives that may represent columns are searched within the window. The candidate primitives for columns include closed polylines, rectangular wireframes, filled areas, closed loops formed by ordinary line segments, and closed contours after block references are expanded. For each candidate column element, we determine whether it has a closed boundary, whether its size is within a reasonable range, whether its area meets the characteristics of a column component, and whether it contains or is close to the intersection of the grid. When there are multiple candidate column elements near the same grid intersection, we filter them based on the distance between the candidate element and the grid intersection, the size of its area, the degree of closure, and the overlap relationship. For the same column component that is simultaneously expressed by a fill, outer contour, polyline, or block reference, we merge multiple candidate elements that are spatially overlapping or mutually contained into a unified column object to obtain the column recognition result. Thus, column recognition can be transformed from simple rectangle or fill recognition to local closed region recognition under grid intersection constraints. Column components are comprehensively identified through grid position, closed geometry, size, and area constraints.

[0055] Furthermore, after identifying column components, this embodiment of the invention uses column objects as seeds for beam component identification, identifying beam edge lines from the remaining linear primitives. First, the frame lines, axes, axis numbers, and already identified column primitives are added to an exclusion set to prevent them from participating in beam line identification. Then, candidate beam lines are extracted from the remaining line segments and polylines, and categorized into horizontal, vertical, and diagonal candidate lines according to direction. For each column object, a beam line search area is constructed around it, searching for pairs of parallel candidate lines that are adjacent to the column boundary, cross the column boundary, or are located in the extension direction of the column edge. If two candidate lines have approximately the same direction, a vertical spacing within a reasonable beam width range, sufficient overlap in their projections in the beam direction, and a close, crossing, or connected relationship with the column boundary, then the line pair is identified as the initial beam edge line pair. After obtaining the initial beam edge lines, the beam line endpoints are used as the propagation front, and the search continues for candidate lines near the endpoints that are collinear with, connected to, or continuous in direction with the already identified beam lines. If a candidate line satisfies the same conditions as an identified beam line in terms of endpoint distance, direction consistency, collinear continuity, or a fracture relationship that can be explained by column occlusion, it is added to the beam line set. The endpoints of the newly added beam lines continue to serve as the propagation front, forming a recursive identification process for the beam line network. In cases where one side of the beam edge is missing or is not identified, candidate lines that are parallel to the target beam line, reasonably spaced, and have sufficient projection overlap are searched from the non-beam candidate lines based on the nearby parallel support relationship. These candidate lines are then used to supplement the missing beam edge, resulting in the beam identification result. Thus, beam identification can be established on the relationship between column members and beam-column connections. The completeness of the beam line network identification is improved by combining constraints of column seed—parallel beam edge pair—endpoint propagation—collinear continuity—missing beam edge replenishment, and the misidentification of dimension lines, annotation lines, axis fragments, and other non-beam lines is reduced.

[0056] For example, Figure 3 This is a schematic diagram of the grid, column, and beam identification process according to a specific embodiment of the present invention, such as... Figure 3 As shown, the search range is first defined by the intersection points of the grid lines, and the column members are identified using these as prior locations. Then, the identified columns are used as identification seeds to search for parallel lines in the surrounding area that meet the requirements in terms of spacing and projection relationship, thus obtaining the initial beam edge line pairs. Next, the beam line endpoints are used as propagation base points, and beam lines that are broken due to occlusion are completed according to collinearity and connection rules, while missing single-sided beam edge lines are also added. Finally, the overall beam line network is sorted and optimized, and invalid lines and pseudo-beam structures such as column edge lines, isolated line segments, and annotation lines are deleted. Line merging and breakpoint correction are completed, and finally a complete and accurate beam and column member network is formed.

[0057] Further, in some embodiments, the structural CAD plan view parsing result is obtained based on the effective range of a single plan view, column identification results, and beam identification results, including: mapping each column object in the column identification results to a target structural diagram node; calculating the beam centerline based on the beam edge lines in the beam identification results, matching both ends of the beam centerline to the corresponding target structural diagram nodes to obtain the target structural diagram edge; identifying the beam number, beam cross-sectional dimensions, and reinforcement annotation text based on the effective range of the single plan view, and matching the beam number, beam width, beam height, and reinforcement information to the corresponding target structural diagram edge based on the distance between the text bounding box and the beam centerline, the projection position of the text in the beam direction, the spatial relationship between the text and the beam edge buffer, and the consistency of the beam number, and matching the column number and column dimension information to the corresponding target structural diagram node; obtaining the structural CAD plan view parsing result based on the target structural diagram node, the target structural diagram edge, and the effective range of the single plan view.

[0058] Specifically, after obtaining the beam identification results, the initially identified beam line set is cleaned and corrected. Line segments that clearly belong to column edges rather than beam edges are deleted; beam lines with connected endpoints and consistent directions are merged; completely overlapping or nearly duplicated line segments are removed; the beam line network is re-broken at intersections; isolated line segments or branches lacking nearby parallel supports are deleted; and beam edges that were mistakenly deleted or missed are restored. The nearby parallel support determination is used to distinguish between true beam edges and ordinary auxiliary lines. For a target beam line, a set of line segments parallel to it is searched along its vertical direction within a reasonable beam width. If the projection coverage length of this set of line segments in the direction of the target beam line reaches a preset threshold, the target beam line is considered to have parallel beam edge support; if the corresponding support is lacking, the target beam line may be a label line, axis fragment, or other non-beam line, and it is deleted or temporarily not considered a beam edge line. Thus, the structural constraint check of the initially identified beam line network is performed, ensuring that the final beam identification result not only meets local geometric conditions but also satisfies the structural plan view representation rule that beams are usually expressed by pairs of parallel edges and connected to columns or other beams.

[0059] Furthermore, the output beam-column geometric recognition results are provided. Each column object includes at least the column outline, column fill, column bounding box, column center point, source element handle, and associated grid intersection. Each beam object includes at least the beam edge set, beam direction, candidate beam width value, beam endpoint, source element handle, and the structural plan number. This further constructs the structural diagram data. G =( V , E ), Figure 4 This is a schematic diagram of a graph data structure provided according to a specific embodiment of the present invention, wherein, V For a set of structural nodes, E This is a set of structural edges. Each column object is converted into a structural node, and each beam object is converted into one or more structural edges. For column objects, the representative point near its center point or grid intersection is taken as the node position, and node attributes such as column bounding box, column size, source drawing, drawing frame, plan number, and source element handle are recorded. For beam objects, the beam centerline is calculated based on its paired beam edges, and the two ends of the beam centerline are matched with the nearest column node, beam-beam intersection node, or virtual end node. When the two ends of the beam centerline are matched with two different nodes, a structural edge is established between the two nodes. Beam edges that are consistent in direction, spatially overlapping, or belong to the same continuous beam can be merged into the same structural edge or segmented by span. Each structural drawing edge records attributes such as beam centerline, beam direction, beam span, beam width, end nodes, source beam edge handle, plan number, and source drawing. This enables the conversion of CAD geometric recognition results into computable structural drawing data, so that the recognition results are no longer limited to line segments and fill levels, but have structural semantics such as beams, columns, and their connection relationships.

[0060] Furthermore, text information such as beam number, beam cross-sectional dimensions, and reinforcement annotations are extracted from the structural plan view and attached to the corresponding structural drawing edge. For beam cross-sectional dimension text, beam numbers and cross-sectional dimension expressions such as "KL1(2) 300×600" and "L1 250×500" are identified, and the beam number, beam width, and beam height are parsed. Based on the distance between the text insertion point or text enclosing box and the beam centerline, the projection position of the text in the beam direction, and the spatial relationship between the text and the beam drawing edge, the parsed beam width, beam height, and beam number are written into the corresponding drawing edge attribute. For reinforcement annotation text, information such as top reinforcement, bottom reinforcement, stirrups, web reinforcement, support reinforcement, and mid-span reinforcement are identified. Based on the distance from the text coordinates to the beam centerline, the projection range of the text in the beam direction, the beam edge buffer where the text is located, and the consistency with the beam number, the reinforcement information is written into the corresponding beam drawing edge. For centralized annotations applied to multi-span continuous beams, the corresponding attributes can be propagated to multiple corresponding edges based on beam number or continuous edge relationships. This transforms the text annotations originally scattered in CAD drawings into structural edge attributes, ensuring that the structural drawing data not only includes beam-column geometry and topology but also engineering semantic information such as beam width, beam height, beam number, and reinforcement. The final output is structural drawing data and its attribute table. The structural drawing data includes node sets, edge sets, node attributes, edge attributes, and source CAD element index relationships. Node attributes include column location, column size, drawing location, and source element; edge attributes include beam centerline, beam span, beam width, beam height, beam number, reinforcement information, end nodes, and source beam edge line.

[0061] The structural CAD plan view parsing method based on geometric topological constraints provided by the present invention first reads the original CAD drawing data and determines the structural plan view to be identified. It then obtains the grid information through the spatial correspondence between axis numbers and axes, and the grid topology, and determines the effective recognition area of ​​a single plan view accordingly. Next, within the defined area, column members are identified a priori based on the grid intersection positions. Using column members as seeds, beam members are identified by combining beam-column connection relationships and beam line propagation constraints. After integrating the planar range, beam-column identification information, and member attribute information, a complete parsing result of the structural CAD plan view is finally obtained. This method solves the problem of drawing parsing being limited by drawing habits and layer rules, resulting in poor recognition of complex drawings, and improves the adaptability and efficiency of structural plan view parsing.

[0062] To enable those skilled in the art to better understand the structural CAD plan view analysis method of the present invention, the following explanation will be provided in conjunction with specific embodiments.

[0063] Figure 5 This is a schematic diagram of the overall process of the intelligent analysis method for structural CAD plan drawings provided according to a specific embodiment of the present invention, as follows: Figure 5As shown, the analysis process is divided into three main stages: Stage 1 is drawing preprocessing and planar positioning, which sequentially completes the input of original CAD elements, element standardization processing, drawing frame detection and drawing area splitting, and structural plan area discrimination, completing the initial screening and positioning of the drawings to be analyzed; Stage 2 is the progressive identification of grid lines, columns, and beams, which sequentially performs axis number identification, axis line identification, multi-planar division and range determination, column member identification based on grid intersection priors, and initial beam edge pair identification based on column seeds, progressively identifying beam and column members based on the grid skeleton; Stage 3 is beam line network cleaning and structured output, which sequentially performs beam line propagation and collinear continuation completion, beam line network cleaning and pseudo-beam filtering, beam and column geometric result output, structural drawing data generation and attribute attachment, and finally outputs calculable structural drawing data.

[0064] It should be noted that the embodiments of the present invention are based on geometric shape, text features, spatial position relationship and topological constraints for parsing, without relying on layer name, color and line type, so as to realize the automated and universal parsing of CAD structural plan drawings.

[0065] The structural CAD plan view parsing method based on geometric topological constraints provided by the present invention first reads the original CAD drawing data and determines the structural plan view to be identified. It then obtains the grid information through the spatial correspondence between axis numbers and axes, and the grid topology, and determines the effective recognition area of ​​a single plan view accordingly. Next, within the defined area, column members are identified a priori based on the grid intersection positions. Using column members as seeds, beam members are identified by combining beam-column connection relationships and beam line propagation constraints. After integrating the planar range, beam-column identification information, and member attribute information, a complete parsing result of the structural CAD plan view is finally obtained. This method solves the problem of drawing parsing being limited by drawing habits and layer rules, resulting in poor recognition of complex drawings, and improves the adaptability and efficiency of structural plan view parsing.

[0066] Next, the structural CAD plan view analysis device based on geometric topological constraints according to an embodiment of the present invention is described with reference to the accompanying drawings.

[0067] Figure 6 A block diagram of a structural CAD plan view analysis device based on geometric topological constraints provided in an embodiment of the present invention.

[0068] like Figure 6 As shown, the structural CAD plan view parsing device 10 based on geometric topological constraints includes: an acquisition module 100, a determination module 200, and a parsing module 300.

[0069] The acquisition module 100 is used to acquire the original CAD drawing data; the determination module 200 is used to determine the structural plan view to be identified based on the original CAD drawing data, and to perform constraint analysis on the structural plan view to be identified based on the spatial correspondence between axis numbers and axes and the axis grid topology to obtain the axis grid identification result, and to determine the effective range of a single plan view based on the axis grid identification result; the parsing module 300 is used to obtain the column identification result based on the prior knowledge of the axis grid intersection position and the axis grid identification result, and to obtain the beam identification result based on the pseudo-beam filtering constraint, and to obtain the structural CAD plan view parsing result based on the effective range of a single plan view, the column identification result, and the beam identification result.

[0070] Furthermore, in some embodiments, the determining module 200 is specifically used for: standardizing the original CAD drawing data to obtain a standardized CAD element set; identifying the frame area and component attribute information in the standardized CAD element set, and performing region discrimination on the frame area to obtain the structural plan area discrimination result; and determining the structural plan to be identified based on the structural plan area discrimination result.

[0071] Furthermore, in some embodiments, the determining module 200 is also used to: obtain original drawing data in the original CAD drawing data; expand all block references in the original drawing data to obtain a block virtual expansion result; and perform coordinate alignment of the internal entities of the blocks in the original drawing data according to the block virtual expansion result to obtain a standardized CAD drawing data set.

[0072] Furthermore, in some embodiments, the determining module 200 is also used to: determine the candidate drawing frame spatial boundary of each standardized CAD drawing element based on the standardized CAD drawing element set; divide each standardized CAD drawing element into the corresponding target drawing frame region according to the preset element inclusion relationship; take the target drawing frame region that meets the preset plan view features as the effective structural plan view region, and delete the target drawing frame region that does not meet the preset plan view features to obtain the structural plan view region discrimination result, wherein the preset plan view features include at least one of the following: the number of candidate axis numbers, line segment distribution features, title keywords, and beam and column layout features.

[0073] Furthermore, in some embodiments, the determining module 200 is further configured to: acquire all candidate circular primitives and candidate linear primitives in the planar diagram of the structure to be identified; determine whether there is text content that conforms to the preset axis number rules within each candidate circular primitive; if there is no text content that conforms to the preset axis number rules, skip the candidate circular primitive; otherwise, generate an axis number object based on the candidate circular primitive and the text content corresponding to the candidate circular primitive; merge the candidate linear primitives that satisfy the preset merging rules to generate candidate long lines; determine the spatial correspondence between each axis number object and each candidate long line; take the candidate long line that satisfies the preset axis number object correspondence conditions as the axis line corresponding to the axis number object; and obtain the axis grid recognition result based on the axis number object and the axis line corresponding to the axis number object.

[0074] Furthermore, in some embodiments, the determining module 200 is also used to: construct an axis grid topology map based on the axis grid recognition results, calculate the connected components of the axis grid topology map, obtain the connected component calculation results, generate multiple axis grid units based on the connected component calculation results; obtain the center coordinates of all axis number objects in each axis grid unit, generate the minimum envelope region corresponding to each axis number object based on the center coordinates of the axis number objects; expand the minimum envelope region according to a preset expansion distance to obtain the effective range of a single planar map.

[0075] Further, in some embodiments, the parsing module 300 is specifically used for: calculating the coordinates of the intersection points of the grid lines based on the effective range of a single plan view; determining multiple initial candidate column elements within a preset range of the grid line intersection coordinates; selecting candidate column elements that satisfy preset column recognition constraints as target candidate column elements; obtaining column recognition results based on the target candidate column elements; acquiring all initial candidate beam lines within the effective range of a single plan view after excluding frame lines, axis lines, axis number elements, and already recognized column elements; using the column objects in the column recognition results as beam recognition seeds; searching for pairs of parallel candidate lines around the column objects; filtering the initial candidate beam lines based on directional consistency, line spacing, projection overlap, and proximity, crossing, or connection with the column boundary; and obtaining beam recognition results based on beam line endpoint propagation, collinear continuation, nearby parallel support, missing beam edge filling, and pseudo-beam filtering.

[0076] Furthermore, in some embodiments, the parsing module 300 is also used to: map each column object in the column identification result to a target structural diagram node; calculate the beam centerline based on the beam edge line in the beam identification result, match both ends of the beam centerline to the corresponding target structural diagram node to obtain the target structural diagram edge; based on the effective range of a single plan view, identify the beam number, beam cross-sectional dimensions, and reinforcement annotation text, and match the beam number, beam width, beam height, and reinforcement information to the corresponding target structural diagram edge based on the distance between the text bounding box and the beam centerline, the projection position of the text in the beam direction, the spatial relationship between the text and the beam edge buffer, and the consistency of the beam number, and match the column number and column dimension information to the corresponding target structural diagram node; obtain the structural CAD plan view parsing result based on the target structural diagram node, the target structural diagram edge, and the effective range of a single plan view.

[0077] It should be noted that the foregoing explanation of the embodiment of the structural CAD plan view parsing method based on geometric topology constraints also applies to the structural CAD plan view parsing device based on geometric topology constraints in this embodiment, and will not be repeated here.

[0078] The structural CAD plan view parsing device based on geometric topological constraints provided in this embodiment of the invention first reads the original CAD drawing data and determines the structural plan view to be identified. It obtains the grid information through the spatial correspondence between axis numbers and axes and the grid topology relationship, and determines the effective recognition area of ​​a single plan view accordingly. Then, within the defined area, it identifies column members based on the a priori position of grid intersections. Using column members as seeds, it identifies beam members by combining beam-column connection relationships and beam line propagation constraints. After integrating the plan range, beam-column identification information, and member attribute information, it finally obtains the complete parsing result of the structural CAD plan view. This solves the problem that drawing parsing is limited by drawing habits and layer rules, resulting in poor recognition of complex drawings, and improves the adaptability and efficiency of structural plan view parsing.

[0079] Figure 7 This is a schematic diagram of an electronic device provided according to an embodiment of the present invention. The electronic device may include: The memory 701, the processor 702, and the computer program stored on the memory 701 and capable of running on the processor 702.

[0080] When the processor 702 executes the program, it implements the structural CAD plan view parsing method based on geometric topological constraints provided in the above embodiments.

[0081] Furthermore, electronic devices also include: Communication interface 703 is used for communication between memory 701 and processor 702.

[0082] The memory 701 is used to store computer programs that can run on the processor 702.

[0083] The memory 701 may include high-speed RAM (Random Access Memory) memory, and may also include non-volatile memory, such as at least one disk storage.

[0084] If the memory 701, processor 702, and communication interface 703 are implemented independently, then the communication interface 703, memory 701, and processor 702 can be interconnected via a bus to complete communication between them. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 7 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0085] Optionally, in a specific implementation, if the memory 701, processor 702, and communication interface 703 are integrated on a single chip, then the memory 701, processor 702, and communication interface 703 can communicate with each other through an internal interface.

[0086] The processor 702 may be a CPU (Central Processing Unit), an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement embodiments of the present invention.

[0087] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0088] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0089] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for analyzing structural CAD plan views based on geometric topological constraints, characterized in that, Includes the following steps: Obtain the original CAD drawing data; The plan view of the structure to be identified is determined based on the original CAD drawing data, and the plan view of the structure to be identified is subjected to constraint analysis based on the spatial correspondence between axis numbers and axes and the axis grid topology to obtain the axis grid identification result. The effective range of a single plan view is determined based on the axis grid identification result. Based on the effective range of the single plan view and the grid identification results, the column identification results and beam identification results are obtained based on the prior location of the grid intersections. Based on the effective range of the single plan view, the column identification results, and the beam identification results, the structural CAD plan view analysis results are obtained.

2. The method according to claim 1, characterized in that, The step of determining the plan view of the structure to be identified based on the original CAD drawing data includes: The original CAD drawing data is standardized to obtain a standardized set of CAD elements; Identify the frame area and component attribute information in the standardized CAD drawing element set, and perform region discrimination on the frame area to obtain the structural plan view region discrimination result; The structural planar diagram to be identified is determined based on the regional discrimination result of the structural planar diagram.

3. The method according to claim 2, characterized in that, The standardization process of the original CAD drawing data to obtain a standardized CAD element set includes: Obtain the original drawing data from the original CAD drawing data; All block references in the original graphic data are expanded to obtain a block virtual expansion result. Based on the block virtual expansion result, the internal entities of the original graphic data are aligned with coordinates to obtain a standardized CAD primitive set.

4. The method according to claim 2, characterized in that, The process of identifying the frame region and component attribute information in the standardized CAD element set, and performing region discrimination on the frame region to obtain the structural plan view region discrimination result, includes: Based on the standardized CAD primitive set, the candidate drawing frame space boundary of each standardized CAD primitive is determined; Based on the preset element inclusion relationship, each standardized CAD element is divided into the corresponding target drawing frame area; The target frame regions that satisfy the preset planar diagram features are taken as valid structural planar diagram regions, and the target frame regions that do not satisfy the preset planar diagram features are deleted, thus obtaining the structural planar diagram region discrimination result.

5. The method according to claim 1, characterized in that, The step of obtaining the grid identification result based on the constraint analysis of the plan view of the structure to be identified includes: Obtain all candidate circular primitives and candidate linear primitives in the planar diagram of the structure to be identified; Determine whether there is text content that conforms to the preset axis number rule within each candidate circular graphic element. If there is no text content that conforms to the preset axis number rule, skip the candidate circular graphic element; otherwise, generate an axis number object based on the candidate circular graphic element and the text content corresponding to the candidate circular graphic element. The candidate linear primitives that meet the preset merging rules are merged to generate candidate long lines. The spatial correspondence between each axis number object and each candidate long line is determined. The candidate long lines that meet the preset axis number object correspondence conditions are taken as the axis lines corresponding to the axis number objects. The grid identification result is obtained based on the axis number object and the axis line corresponding to the axis number object.

6. The method according to claim 1, characterized in that, Determining the effective range of a single plan view based on the grid recognition result includes: Based on the grid identification results, a grid topology map is constructed, and the connected components of the grid topology map are calculated to obtain the connected component calculation results. Multiple grid units are generated based on the connected component calculation results. Obtain the center coordinates of all axis number objects in each axis grid unit, and generate the minimum envelope region corresponding to each axis number object based on the center coordinates of the axis number objects; The minimum envelope region is expanded according to a preset expansion distance to obtain the effective range of a single planar image.

7. The method according to claim 1, characterized in that, The process of obtaining column and beam identification results based on the effective range of the single plan view includes: The coordinates of the grid intersection points are calculated based on the effective range of the single plan view. Multiple initial candidate column elements are determined within a preset range of the coordinates of the intersection points of the grid lines. The candidate column elements that satisfy the preset column recognition constraints are taken as target candidate column elements. The column recognition result is obtained based on the target candidate column elements. All initial candidate beam lines are obtained from the candidate linear elements within the effective range of the single plan view after excluding the frame lines, axis lines, axis number elements, and identified column elements. Using the column objects in the column identification results as beam identification seeds, pairs of parallel candidate lines are searched around the column objects. The initial candidate beam lines are filtered by combining directional consistency, line spacing, projection overlap relationship, and proximity, crossing, or connection relationship with the column boundary. The beam identification result is obtained based on beam line endpoint propagation, collinear continuation, nearby parallel support, missing beam edge filling, and pseudo-beam filtering.

8. The method according to claim 2, characterized in that, The structural CAD plan view analysis result obtained based on the effective range of the single plan view, the column identification result, and the beam identification result includes: Each column object in the column identification result is mapped to a node in the target structure graph. Calculate the beam centerline based on the beam edge lines in the beam identification results, and match both ends of the beam centerline to the corresponding target structure diagram nodes to obtain the target structure diagram edge; Based on the effective range of the single plan view, the beam number, beam cross-sectional dimensions, and reinforcement annotation text are identified. According to the distance between the text bounding box and the beam centerline, the projection position of the text in the beam direction, the spatial relationship between the text and the beam edge buffer, and the consistency of the beam number, the beam number, beam width, beam height, and reinforcement information are matched to the corresponding target structural diagram edge according to the preset spatial position relationship, and the column number and column dimension information are matched to the corresponding target structural diagram node. The analytical result of the structural CAD plan view is obtained based on the target structure diagram nodes, the target structure diagram edges, and the effective range of the single plan view.

9. A structural CAD plan view analysis device based on geometric topological constraints, characterized in that, include: The acquisition module is used to acquire the original CAD drawing data; The determination module is used to determine the plan view of the structure to be identified based on the original CAD drawing data, and to perform constraint analysis on the plan view of the structure to be identified based on the spatial correspondence between axis numbers and axes and the axis grid topology relationship to obtain the axis grid identification result, and to determine the effective range of a single plan view based on the axis grid identification result; The parsing module is used to obtain column and beam identification results based on the effective range of the single plan view and the grid identification results, and to obtain the structural CAD plan view parsing results based on the prior position of the grid intersection.

10. An electronic device, characterized in that, include: The memory, the processor, and the computer program stored in the memory and capable of running on the processor, wherein the processor executes the program to implement the structural CAD plan view analysis method based on geometric topological constraints as described in any one of claims 1-8.

Images