A method and system for identifying and resource networking of communication engineering CAD drawings

By utilizing path constraints and local defect area identification and correction methods in communication engineering CAD drawings, the problem of identification error under fixed parameters was solved, and a more efficient and accurate resource network access process was achieved.

CN123123183APending Publication Date: 2026-11-17WHALE CLOUD TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611620831.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-20
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In the process of automatic recognition of CAD drawings for communication engineering, existing technologies are prone to missing point entities and line entities when using fixed similarity thresholds and fixed spatial connection tolerances. However, directly relaxing the recognition parameters for the entire drawing can easily increase the problems of incorrect recognition and incorrect connections.

Method used

By using the engineering records in the occupancy map and the optical cable distribution map, the identification results of the routing map are used to form path constraints. When the path constraints are not closed, local defect areas are identified, and the identification parameters are relaxed in these areas. Through re-identification and closure judgment, a closed-loop correction process is formed.

Benefits of technology

It achieves accurate identification of point entities and line entities, reduces the risk of incorrect identification and connection caused by uniformly relaxing the identification parameters of the whole map, and improves the reliability and accuracy of the identification results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123123183A_ABST
    Figure CN123123183A_ABST
Patent Text Reader

Abstract

The application discloses a kind of communication engineering CAD drawing identification and resource network access method and system. Including: in the route map with first threshold value and first buffer distance, point entity and line entity are identified, and route topology graph is constructed;Each record in pipe hole occupation map, cable distribution map is regarded as path constraint, and it is required that there is connected path between the corresponding nodes of start point name and end point name, and closed determination is carried out;For the path constraint that is not closed, according to its missing part in the route map, determine the defect area, start from the vector element intersected with the defect area to carry out local re-identification with second threshold value or second buffer distance, when candidate entity makes the path constraint re-close, it is retained, otherwise discarded;The identification result is written into resource management system, and the topology relationship is established along the closed path constraint.The application can correct the missed identification, while avoiding the misidentification caused by the relaxation of identification parameters.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of communication network resource management technology, specifically to a method and system for identifying and integrating communication engineering CAD drawings into the network. Background Technology

[0002] In the design and completion phases of communication pipelines and optical cables, CAD drawings serve as the primary carrier of engineering information. These drawings are typically in DWG or DXF format. A set of CAD drawings for a communication project generally includes multiple sheets, such as routing diagrams, conduit occupancy diagrams, and optical cable reel layout diagrams. Routing diagrams use graphic symbols to represent routing nodes such as manholes, handholes, and poles, lines to represent pipeline segments or pole sections between these nodes, and text labels indicating the node numbers. Conduit occupancy diagrams record the conduit numbers, occupancy status, and names of the optical cables laid within each pipeline segment. Optical cable reel layout diagrams record the name, model, number of cores, start and end points, and reel length of each optical cable segment. After project completion, this information needs to be entered into a resource management system to establish the topological relationships between pipelines, conduits, and optical cables for resource scheduling and maintenance.

[0003] Currently, resource entry is mainly completed by staff manually comparing drawings to items in the resource management system. A single drawing typically contains hundreds to thousands of graphic elements, making manual entry time-consuming and prone to errors. To reduce manual entry, some automated processing methods have emerged in existing technologies. One method uses data conversion tools between CAD and Geographic Information Systems (GIS) to batch convert CAD graphic elements into point, line, and polygon features according to their geometric types, then imports them into a spatial database after attaching annotation attributes. This method only completes the format conversion; the converted features do not have resource types and resource codes, resource determination still requires manual completion, and it cannot handle the association between different map sheets. Another method uses feature values ​​such as the distance between feature points as identifiers, searches for graphic elements with the same feature values ​​within the layer range, combines them into a graphic, and then compares it with the feature values ​​of the source graphic to find similar graphics. This method requires precise equality of feature values ​​and has low tolerance for cartographic deviations.

[0004] Using a similarity threshold for template matching can tolerate some cartographic deviations, but the recognition results are still difficult to directly use for resource integration into the network. Actual engineering drawings are often produced by different design units, resulting in inconsistent drawing scales and line types for symbols such as manholes, and it's common for pipeline endpoints not to accurately connect to node symbols. When using a fixed similarity threshold and buffer distance for recognition, these locations may experience missed identifications. If the similarity threshold is lowered or the buffer distance is increased for the entire drawing to improve recall, symbols with similar shapes, such as manholes and solid round holes, may be misidentified as manholes, or unconnected pipelines may be linked together. Missed identifications are usually difficult to detect during the recognition phase, often only manifesting when the routing map and the pipe / hole occupancy map, or the fiber optic cable distribution map are correlated, or during topology consistency verification before data integration. In such cases, problems like pipe / hole occupancy relationships not matching pipe segments or fiber optic cable route breaks usually occur. At this point, manual processing is generally the only option, as information already recorded in other maps is not used to correct the recognition results. Summary of the Invention

[0005] To address the problem that existing automatic recognition processes for CAD drawings in communication engineering are prone to missing point and line entities when using fixed similarity thresholds and fixed spatial connection tolerances, while relaxing the recognition parameters for the entire drawing can easily increase misidentification and incorrect connections, this invention provides a method and system for recognizing and integrating CAD drawings for communication engineering into the network.

[0006] This scheme uses engineering records in the occupancy map and optical cable distribution map to form path constraints based on the identification results of the routing map. When the path constraints are not closed, the missing parts are used to determine the local defect area. Only for this local area, the parameters are relaxed and re-identified. The same path constraints are used to re-identify the results and make a closure judgment again, thus forming an identification correction closed loop driven by the associated map information.

[0007] Includes the following steps: The vector elements obtained from the parsed CAD drawings of the communication engineering project are obtained. The vector elements include line elements and text elements. The routing map sheets and associated map sheets are determined from the vector elements. The associated map sheets include at least one of the following: the pipe hole occupancy map sheet and the optical cable reel distribution map sheet.

[0008] In the routing map, vector elements are matched for graphical similarity based on point entity templates to identify point entities, and text elements are associated with the point entities to obtain their names. Line elements are spatially connected to the point entities with a first buffer distance as tolerance, and line entities connecting different point entities are identified based on the spatial connection results. A routing topology map is constructed using the point entities as nodes and the line entities as edges. The initial identification of the point entities uses a first threshold as the similarity judgment threshold.

[0009] Records are extracted from the associated map, each record including a start-point name and an end-point name; a path constraint is generated from each record, the path constraint requiring that there be a connected path between the node corresponding to the start-point name and the node corresponding to the end-point name in the routing topology map.

[0010] For each path constraint, a closure determination is performed. The path constraint is considered closed when both the origin name and the destination name have corresponding nodes in the routing topology graph, and there is a connected path between the two corresponding nodes; otherwise, the path constraint is considered unclosed.

[0011] For unclosed path constraints, the defective region is determined based on the location of its missing portion in the routing map. Vector elements intersecting with the defective region are used as local re-identification objects, and spatial connections are initiated. During re-identification, at least one of the following relaxation methods is adopted: the similarity judgment threshold is reduced from the first threshold to the second threshold; the buffer distance of spatial connections is increased from the first buffer distance to the second buffer distance, thereby obtaining candidate entities.

[0012] After adding the candidate entity to the routing topology graph, the corresponding path constraints are checked for closure again. If the path constraint is closed, the candidate entity is retained and the routing topology graph is updated accordingly; if the path constraint is not closed, the candidate entity is discarded.

[0013] Write the point entities and line entities in the updated routing topology graph into the resource management system, and associate the resources corresponding to their source records with the line entities on the corresponding connected paths according to the closed path constraints.

[0014] Furthermore, a nearest neighbor matching method is used between text elements and point entities. In each round, the nearest text element to an unpaired point entity is determined, and then it is determined whether the nearest point entity of that text element is the current point entity. If the two are the closest objects to each other, a pairing is established, and the above process is repeated until no new pairings are generated.

[0015] Furthermore, when the missing part of the unclosed path constraint is manifested as the endpoint name not having a corresponding node in the routing topology map, a text element that is not paired with a point entity is selected from the text elements whose content is consistent with the endpoint name as an anchor point. The defect area is determined with the anchor point as the center and a preset multiple of the average feature size of the identified point entities in the routing map as the radius.

[0016] Furthermore, when the missing part of the unclosed path constraint manifests as the lack of a connected path between two corresponding nodes, the point entities, line entities, and line features that intersect with the point entities in the connected component but have not yet been identified as line entities in the two corresponding nodes are taken as candidate features. The pair of features closest to each other between the candidate features of the two connected components is determined, and the shortest connection between the pair of features is taken as the center line, and the defect area is determined according to the preset width.

[0017] Furthermore, fidelity analysis is performed on the CAD drawings of the communication engineering project. This fidelity analysis includes disassembling the block references and filtering out text elements generated by invisible attributes, as well as decoding multi-byte encoded text represented by MIF escape sequences back to the original text.

[0018] Before comparing the start-point and end-point names with the point entity names and text element content in the associated map sheet records, name normalization is performed on both sides of the comparison. The name normalization includes performing character replacement according to the mapping table and performing truncation according to the truncation rules. Each side of the comparison adopts its own configured truncation rules.

[0019] Furthermore, the second threshold is greater than the maximum similarity between the point entity template and other categories of graphics in the template library, but less than the first threshold. The other categories of graphics include point entity templates of other resource types and pre-registered non-resource symbols. This ensures that the reduced similarity threshold remains within the distinguishing range between the target template and registered easily confused graphics.

[0020] Furthermore, the second buffer distance is greater than the first buffer distance and less than half of the minimum distance between any two identified point entities in the routing map, thereby increasing the spatial connectivity tolerance while limiting the possibility of overlap between the buffer ranges of different point entities.

[0021] Furthermore, the point entity template includes a base graphic and auxiliary graphics. The graphic to be identified and the template graphic are geometrically normalized, and a total similarity is calculated based on the normalized geometric features. Candidate base graphics are determined based on the total similarity, and auxiliary graphics are then matched within the neighborhood of the candidate base graphics. A combined similarity is formed based on the candidate base graphics and the successfully matched auxiliary graphics. The first threshold and the second threshold are used as the determination thresholds for the combined similarity. When re-identifying within a defective region, the similarity threshold used for matching the base graphics and auxiliary graphics is reduced, while maintaining consistency in scaling ratio, relative position, and primitive attribute conditions.

[0022] Furthermore, when the associated map sheet includes both the conduit occupancy map sheet and the optical cable reel distribution map sheet, the conduit occupancy record also includes the name of the optical cable being laid, and the optical cable segment record also includes the optical cable segment name and the reel distribution length. First, the path constraints generated from the conduit occupancy record are processed, and the conduit occupancy relationship corresponding to the closed path is established; then, the path constraints generated from the optical cable segment record are processed.

[0023] For closed path constraints generated from optical cable segment records, candidate paths without duplicate nodes between two corresponding nodes are enumerated, and the occupancy evidence rate and length deviation of each candidate path are calculated. The occupancy evidence rate is determined based on the number of line entities on the candidate path with a name matching the optical cable segment name, and the length deviation is determined based on the difference between the candidate path length and the reel length. Candidate paths with higher occupancy evidence rates are preferentially selected; if the occupancy evidence rates are the same, candidate paths with smaller length deviations are selected, and the selected candidate paths are used as the connectivity paths for establishing the optical cable segment topology.

[0024] Furthermore, the two corresponding nodes of the closed path constraint generated by the optical cable segment record are used as the endpoints of the optical cable segment, and the optical cable connectors located at the endpoints are identified; the optical cable segment is divided into connected components based on the relationship of sharing the same optical cable connector.

[0025] For each connected component, if at least one reuse condition is met, the connected component is determined to be reused and the corresponding existing network identifier is retained. The first reuse condition is that the connected component contains an optical cable segment with an existing network identifier; the second reuse condition is that the connected component contains a reused optical cable connector, and the reused optical cable connector is associated with an existing optical cable. The reused optical cable connector includes optical cable connectors marked as reused in the drawings and optical cable connectors already existing in the resource management system. When neither the first nor the second reuse condition is met, the connected component is determined to be newly created, and the optical cable segments within it are assigned the same newly generated network identifier.

[0026] Furthermore, a recovery marker is attached to the candidate entity retained after the re-closure determination. The recovery marker records the defect area where the candidate entity is located, the relaxation parameters used during re-identification, and the path constraints that allow the candidate entity to be retained.

[0027] Before writing to the resource management system, graph models are constructed based on the routing topology map and the existing routing data corresponding to the communication project in the resource management system. The two graph models are then compared for topological consistency based on node labels and edge attributes. When inconsistencies are found in the comparison results, the inconsistent parts involving entities with the recovery markers are output separately from the remaining inconsistent parts, allowing for targeted verification of the re-identified recovery entities.

[0028] A system for identifying and integrating CAD drawings for communication engineering into the network, applicable to the above-mentioned method, includes a drawing sheet determination module, a route identification module, a constraint generation module, a closure determination module, a defect correction module, and a network integration module.

[0029] The map sheet determination module is used to obtain vector elements obtained by parsing the CAD drawings of the communication engineering project, and to determine the routing map sheet and associated map sheet.

[0030] The routing identification module is used to identify point entities based on a first threshold, identify the spatial connection of line entities based on a first buffer distance, and construct a routing topology graph with point entities as nodes and line entities as edges.

[0031] The constraint generation module is used to extract records including start-point and end-point names from associated map sheets, and generate path constraints from each record that require a connected path between the corresponding nodes.

[0032] The closure determination module is used to determine the closure status of path constraints based on whether the corresponding node exists and whether there is a connected path between the corresponding nodes.

[0033] The defect correction module is used to determine the defect region based on the missing part of the unclosed path constraint, take the vector elements that intersect with the defect region as the local re-identification object and start spatial connection, and perform re-identification by at least one of the following methods: lowering the similarity judgment threshold and increasing the spatial connection buffer distance, to obtain candidate entities, and process the candidate entities according to the re-closure judgment result.

[0034] The network entry module is used to write the point entities and line entities in the updated routing topology map into the resource management system, and associate the resources corresponding to their source records with the line entities on the corresponding connected paths according to the closed path constraints.

[0035] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention provides a method and system for identifying and integrating resources into CAD drawings used in communication engineering projects. It converts engineering records in associated drawing sheets into path constraints on the routing topology map, enabling the identification results of point and line entities to be verified for closure using engineering information from other drawing sheets. When path constraints are not closed, their missing portions can reflect potential missed identification locations, thereby identifying local defect areas and allowing parameter relaxation to focus on the local area relevant to the problem.

[0036] 2. This invention provides a method and system for identifying and integrating communication engineering CAD drawings into the network. It involves at least one of lowering the similarity threshold and increasing the spatial connection buffer distance within defective areas, and then re-performing a closure determination using the original path constraints after obtaining candidate entities. Only candidate entities that can close the corresponding path constraints are retained. This creates a closed-loop correction process of "defect detection—local re-identification—path constraint verification," reducing the risk of misidentification and incorrect connections caused by uniformly relaxing identification parameters across the entire drawing.

[0037] 3. This invention provides a method and system for identifying and integrating resources into CAD drawings for communication engineering. It identifies defective areas for different missing scenarios, such as missing corresponding nodes for endpoint names and missing connecting paths between corresponding nodes. This establishes a spatial correspondence between the re-identification range and the specific missing location, which helps reduce redundant calculations in irrelevant areas and maintains the stability of existing identification results in other areas.

[0038] 4. This invention provides a method and system for identifying and integrating CAD drawings for communication engineering into the network. It further sets constraints on the second threshold and the second buffer distance, so that the identification parameters are subject to the limitations of template confusion and spatial distance of identified point entities while obtaining greater tolerance. Combining the combination and matching of basic graphics and auxiliary graphics can further improve the reliability of point entity identification results.

[0039] 5. This invention provides a method and system for identifying and integrating resources into the network of CAD drawings for communication engineering. When both duct occupancy information and optical cable reel information are available, this invention first establishes the duct occupancy relationship, and then uses the occupancy evidence formed by the name of the optical cable and the length constraint formed by the reel length to select the connection path corresponding to the optical cable segment, so that the information already confirmed in the duct drawing can participate in the determination of optical cable routing.

[0040] 6. This invention provides a method and system for identifying and integrating resources into the network of CAD drawings for communication engineering projects. It can determine the connected components of optical cable segments based on the sharing relationship of optical cable joints, and determine the network affiliation of the corresponding connected components based on the existing network identifier and the information of the reused optical cable joints. At the same time, it can distinguish the entities obtained by re-identification by restoring the markers. When differences occur in the topology consistency audit, it can output the identified restoration items and other topology differences separately, thereby facilitating subsequent audits and resource integration processing. Attached Figure Description

[0041] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0042] Figure 1 A flowchart illustrating the method for identifying and integrating communication engineering CAD drawings into the network, as provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the path constraint closure determination and defect correction process in an embodiment of the present invention. Figure 3 This is a schematic diagram illustrating the determination of the defect region when the endpoint name has no corresponding node in an embodiment of the present invention; Figure 4 This is a schematic diagram illustrating the determination of the defective region when there is a missing connecting path between two corresponding nodes in an embodiment of the present invention. Figure 5 This is a schematic diagram of the combined template and matching of manholes in an embodiment of the present invention; Figure 6 This is a schematic diagram illustrating the selection of candidate paths for optical cable segments in an embodiment of the present invention; Figure 7 This is a schematic diagram of the optical cable segment affiliation group in an embodiment of the present invention; Figure 8 A schematic diagram of the structure of the communication engineering CAD drawing recognition and resource network access system provided in an embodiment of the present invention; Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0043] Figure reference numerals: 31. Identified point entity; 32. Unidentified manhole symbol; 33. Anchor point; 34. Circular defect area; 35. Line element; 41. First route connected component; 42. Second route connected component; 43. Line element not identified as a line entity; 44. Shortest connection; 45. Strip-shaped defect area; 51. Base graphic; 52. Subsidiary graphic; 53. Buffer zone; 54. Candidate base graphic; 55. Subsidiary graphic candidate; 61. First candidate path; 62. Second candidate path; 71. First optical cable 72. Second optical cable segment connected component; 73. Optical cable joint; 74. Reused optical cable joint; 75. Optical cable segment; 100. Recognition and resource network access system for communication engineering CAD drawings; 101. Drawing sheet determination module; 102. Route identification module; 103. Constraint generation module; 104. Closure judgment module; 105. Defect correction module; 106. Network access module; 200. Resource management system; 300. Electronic equipment; 301. Processor; 302. Memory; 303. Communication interface; 304. Bus. Detailed Implementation

[0044] The specific embodiments of the present invention will be described below with reference to the accompanying drawings. The following embodiments use communication pipeline and optical cable engineering drawings as examples to provide a directly implementable processing procedure. The listed parameters are used to illustrate the value method of the corresponding steps and can be configured according to the drawing scale, drafting specifications and template calibration results.

[0045] Terminology Explanation: In this embodiment of the invention, a drawing sheet refers to an area in a communication engineering CAD drawing that carries a type of engineering information, such as a routing drawing sheet, a conduit occupancy drawing sheet, and an optical cable reel drawing sheet. Multiple drawing sheets can be drawn in the same CAD file and separated by drawing frames, or they can be located in multiple CAD files for the same project.

[0046] Point entities are point-like resources identified by vector features, such as manholes, handholes, utility poles, and fiber optic cable connectors, and have coordinates, resource type, and name. Line entities are line-like resources connecting two point entities, such as pipe segments and pole segments, and have a start point, end point, and geometric path. The routing topology graph is an undirected graph with point entities as nodes and line entities as edges; the name of the node is the name of the corresponding point entity.

[0047] Path constraints are constraints generated from a record in the associated map sheet, requiring that there be a connected path between the node corresponding to the start name and the node corresponding to the end name of that record in the routing topology. A node corresponding to the start name is a node whose name matches the start name, and a node corresponding to the end name is a node whose name matches the end name. Whether the names match is determined by the result of name normalization. A path constraint is said to be closed when it meets the above requirements.

[0048] A defective region is a local area defined in the routing map based on the missing portion of an unclosed path constraint, used to limit the scope of re-identification. Candidate entities are point or line entities that are re-identified within the defective region with relaxed parameters and are yet to be confirmed.

[0049] In this embodiment of the invention, the feature size of a graphic refers to the diagonal length of its bounding box; the feature size of a point entity refers to the feature size of its base graphic; the distance between two point entities refers to the minimum distance between their base graphics; and the intersection of a line element and a point entity means that the line element intersects with the first buffer distance buffer zone of the point entity's base graphic.

[0050] As a specific embodiment 1, such as Figure 1 As shown, this embodiment of the invention provides a method for identifying and integrating CAD drawings for communication engineering into the network. This method can be executed by electronic devices such as servers and personal computers, and includes the following steps S101 to S108.

[0051] Step S101: Obtain the vector elements obtained from the parsed CAD drawings of the communication project, and determine the routing map sheet and associated map sheet from the vector elements.

[0052] In this embodiment, the CAD drawings for communication engineering are parsed according to the following process. First, the file format is determined: For DWG format files, the command-line interface of a format conversion tool (such as ODA File Converter) is called, specifying the input path, output directory, target version (such as AutoCAD 2010), and output format, to convert it to DXF format. The converted DXF file is then audited, and any recoverable errors are fixed. For DXF format files, they are parsed directly. Subsequently, a DXF parsing library (such as ezdxf) is used to read the DXF file. If reading fails, the recovery mode is used to reread the file. All primitives in the model space are obtained and converted into vector features according to their primitive types, as shown in Table 1.

[0053] Table 1

[0054] In a lightweight polyline, arc segments with non-zero convexity *b* are discretized according to the definition of convexity. Convexity *b* is the tangent of one-quarter of the central angle corresponding to the arc segment. Let the chord length between the two endpoints of the arc segment be *c*, then the radius *r* of the circle containing the arc segment and the directed distance *e* of the center of the circle relative to the midpoint of the chord along the unit normal vector direction are respectively: Let the central angle corresponding to the arc segment be . Then convexity Defined as:

[0055] Let the length of the chord between the two endpoints of the arc segment be c, then the radius r of the circle containing the arc segment is:

[0056] The directed distance of the center of the circle relative to the midpoint of the chord along the direction of the unit normal vector. for:

[0057] in, This represents the square of the convexity b. This represents the absolute value of the convexity b.

[0058] Rotate the chord direction from the starting point to the ending point counterclockwise by 90° to obtain the unit normal vector n. Let M be the midpoint of the chord. Then the center O of the circle containing the arc segment is:

[0059] When b > 0, the arc segment extends counterclockwise from the starting point to the ending point; when b < 0, the arc segment extends clockwise from the starting point to the ending point. Based on the center O and radius r, samples are taken on the arc at angular intervals not greater than 10° to obtain a broken line representing the arc segment.

[0060] During parsing, attribute information is added to each vector feature, including feature type, layer, color, line type, line width, text content, block name, and unique identifier. Text features also include width, height, and rotation angle. When the line width identifier is set to "layer-dependent," the line width defined by the layer is used. When the line width identifier is set to "block-dependent" and the feature is within a block reference, the line width is taken by ascending layer by layer along the nested link of the blocks, using the line width explicitly defined in the innermost layer. If no line width is explicitly defined on any link, the line width of the layer-dependent feature is used.

[0061] To ensure that the text elements are consistent with what drafters see in CAD software, this embodiment performs fidelity analysis on the drawings, including: Block references are disassembled. Most resource symbols in routing maps are drawn in the form of block references. Before disassembly, the invisible attributes of the block references are collected. After disassembly, text elements generated from invisible attributes are filtered out, while text elements generated from visible attributes and other elements within the block are retained. Invisible attributes in the block definition are not displayed in CAD software. If not filtered out, actually invisible text will appear in the routing map. The content of this text may be the same as the node name, which will create incorrect anchor points when identifying defective areas. Decoding multibyte-encoded text. Some cartographic environments use MIF escape sequences to represent multibyte characters such as Chinese characters. The format is "\M+" followed by a one-digit encoding page identifier and a four-digit hexadecimal code. For example, the encoding page identifier is 1 for Japanese Shift-JIS encoding, 2 for Traditional Chinese Big5 encoding, and 5 for Simplified Chinese GB2312 encoding. During decoding, the character set is selected according to the encoding page identifier. The four-digit hexadecimal code is split into high-order and low-order bytes and then decoded into a single character according to the character set, replacing the escape sequences in the original string. Undecoded node numbers cannot be compared with their names in the associated map sheet, which may cause path constraints to be misjudged as unclosed. For multi-line text, estimate the number of characters that each line can hold based on the text width and width factor, and perform line break expansion. Consecutive English letters and numbers are treated as an indivisible whole, while Chinese characters are broken into single characters.

[0062] After parsing, line features with a length less than 0.0001 and polygon features with an area less than 0.0001 are filtered out. Continuous duplicate points in the filled boundary are deduplicated, and the boundary is closed. After all primitives have been processed, the vector features are assembled into GeoJSON feature sets and written to the spatial database by batch identifier for subsequent batch-based queries and spatial calculations.

[0063] When multiple drawing sheets are drawn in the same CAD file, the drawing frame is first identified in the vector features. The drawing frame is a rectangle composed of closed lightweight polylines or solid surfaces, and its area is not less than a preset proportion of the area of ​​the rectangle enclosing all vector features, such as 5%. Each vector feature is assigned to the drawing frame where its centroid is located. If the centroid is located in multiple drawing frames, it is assigned to the drawing frame with the smallest area, thus obtaining the vector features contained in each drawing sheet. Then, the drawing sheet type is determined according to whether the drawing sheet contains text elements with preset keywords. For example, a drawing sheet containing "routing diagram" or "pipeline routing" is a routing drawing sheet, a drawing sheet containing "pipe hole occupancy" or "pipe hole cross-section" is a pipe hole occupancy drawing sheet, and a drawing sheet containing "reel" is a fiber optic cable reel drawing sheet. Keywords can be configured according to the design unit's drafting habits. When the routing drawing sheet, pipe hole occupancy drawing sheet, and fiber optic cable reel drawing sheet are located in different CAD files of the same project, each file is treated as a drawing sheet, and its type is determined according to the keywords in the file.

[0064] Step S102: Identify point entities and line entities in the routing map and construct a routing topology map, including: Before identification, vector features, zero-length lines, and duplicate vector features in irrelevant layers of the routing map are removed. This embodiment pre-establishes a template library. The template library includes point entity templates for various resource types, including manhole templates, handhole templates, utility pole templates, and fiber optic cable connector templates. It also registers several non-resource symbols, including solid circular holes, compasses, and other symbols whose shapes are similar to point entity templates but do not represent resources.

[0065] Each point entity template includes a base graphic and zero or more subordinate graphics, and records the position of the subordinate graphics relative to the base graphic, as well as the requirements for primitive attributes such as layer, color, and linetype. Please refer to [link / reference]. Figure 5 In (a) of the example, the basic graphic 51 of the manhole template is a circle, and the auxiliary graphic 52 is two straight lines located inside the circle and perpendicular to each other.

[0066] The similarity between vector elements and template graphics is calculated using the following steps.

[0067] Calculate the scaling factor, which is the feature size of the graphic to be recognized. Feature dimensions of template graphics The ratio, that is .

[0068] Geometric normalization involves sequentially performing translation, rotation, and scaling on the graphic to be identified and the template graphic: translation moves the centroid of all the vertices of the graphic to the origin; Rotation rotates the graphic's main direction to the X-axis direction; scaling scales the length of the graphic's bounding box diagonal to 1.

[0069] The angle θ between the principal direction and the X-axis is calculated using the following formula: (2) in, and These are the variances of the X and Y coordinates of the graph vertices, respectively. Let X be the covariance of the X and Y coordinates, and arctan2 be the two-parameter arctangent function. θ is the direction angle of the eigenvector corresponding to the largest eigenvalue of the vertex coordinate covariance matrix. Since the principal direction has an ambiguity of 180°, the rotated image can be rotated another 180°, and the similarity can be calculated separately, taking the larger value.

[0070] When the difference between the two eigenvalues ​​of the vertex coordinate covariance matrix is ​​less than a preset degradation threshold and a unique principal direction cannot be determined, the orientation angle of equation (2) is not used as the unique rotation direction. For rotationally symmetric figures, the principal direction rotation step is omitted; for non-rotationally symmetric figures, normalization and subsequent similarity calculations are performed according to the candidate rotation angles pre-registered by the template, and the maximum similarity is taken as the matching result between the figure to be identified and the template figure. This avoids relying on an uncertain principal direction when the covariance matrix degenerates.

[0071] Calculate the similarity of the three features, and resample the normalized graphic contour into N contour points at equal intervals according to arc length. For example, N is 64, resulting in the point set A of the graphic to be identified and the point set B of the template graphic. The Hausdorff distance between the two is: (3) in ; In equation (3), This represents the Euclidean distance between contour point a and contour point b. Topological similarity. The Hausdorff distance similarity is obtained through a linear mapping. The result is obtained by the square root decay mapping of Hausdorff distance: (4) The square root operation amplifies the difference when the Hausdorff distance is small, therefore It is more sensitive to subtle differences between graphics. Area and perimeter are compared to similarity. It is obtained from the roundness of the two shapes: (5)

[0072] Where R represents the circularity of the shape, S represents the area of ​​the shape, and L represents the perimeter of the shape. and These are the circularity values ​​of the graphic to be identified and the template graphic, respectively. For non-closed line graphics, the circularity is calculated using the area and perimeter of its convex hull; if the circularity of both graphics is 0, ... Choose 1. The outer contours of a solid circular hole and a hollow circular ring are similar, and Hausdorff distance is insufficient to distinguish between them, but their roundness differs significantly; the roundness of a circular symbol and a regular polygon symbol are similar, but Hausdorff distance is sensitive to the difference in their contours. Using the three feature similarities together can suppress these two types of misclassifications respectively.

[0073] The three feature similarities are weighted and fused together using the following formula to obtain the total similarity: , (6) The weight is set by default. = 0.3、 = 0.4、 = 0.3, or can be configured separately according to the category of point entity templates. During configuration, for each point entity template, several correct samples taken from actual drawings and easily confused graphics of the template are collected. On the weight value grid, the difference between the minimum similarity of the correct samples and the maximum similarity of the easily confused graphics is calculated one by one. The weight combination with the largest difference is taken as the weight of the template. Easily confused graphics refer to graphics that are similar in shape to the point entity template but do not represent the resource type corresponding to the template, including point entity templates of other resource types in the template library and registered non-resource symbols.

[0074] Specifically, the selection of the weight combination can be expressed as:

[0075] To optimize and obtain the optimal weights, To correctly identify the lowest total similarity in the samples, argmax represents the highest total similarity among easily confused graphs, and argmax is the combination of parameters that maximizes the objective function.

[0076] After template calibration, the first threshold used for initial identification is also determined. . The value is set to be higher than the maximum similarity of easily confused graphics but not higher than the minimum similarity of correct samples, so that there is a judgment interval between correct samples and registered easily confused graphics. If a template cannot form the above judgment interval under the existing basic graphics, auxiliary graphics and primitive attributes, then the template is not used directly for automatic recognition. Instead, distinguishing features are added or the template is recalibrated.

[0077] When a point entity template includes attached graphics, perform a combination matching. For example... Figure 5 As shown in (b), the process of combination matching is as follows.

[0078] First, select graphics with a total similarity to the base graphic that is not lower than the base threshold and whose scaling ratio is within the preset range [s]. min ,s max Vector elements within ] are used as candidate base graphics 54. The base threshold is, for example, 0.8, and the preset range is, for example, [0.5, 2.0].

[0079] Then, taking candidate base graphic 54 as the center, a buffer query is performed with the preset ratio of its feature size as the buffer distance to obtain buffer 53. Vector features whose intersection with buffer 53 accounts for a proportion of their own not less than a preset value are taken as subordinate graphic candidates 55. Among them, the proportion of surface features is calculated according to area, and the proportion of line features is calculated according to length. Figure 5 Only a small portion of the pipelines connected to candidate base graphic 54 are located within buffer 53, therefore they are not considered as auxiliary graphic candidates.

[0080] Match each candidate auxiliary graphic (55) with one of the auxiliary graphics in the template. A successful match requires the following conditions to be met simultaneously: the total similarity between the two graphics is not lower than the auxiliary similarity threshold (e.g., 0.8); and the difference between the scaling ratio of the candidate auxiliary graphic and the scaling ratio of the candidate base graphic does not exceed 10% of the latter. The relative positions are consistent, meaning that the area coverage ratio of surface-type auxiliary graphics to the base graphics is not less than 95%, the length ratio of line-type auxiliary graphics covered by the base graphics is not less than 95%, point-type auxiliary graphics are located within the buffer zone of the base graphics, the ratio of the distance between the center of the auxiliary graphics and the center of the base graphics to the feature size of the base graphics differs from the corresponding ratio in the template by no more than 0.1, the offset of the center of the auxiliary graphics in the coordinate system normalized to the base graphics does not exceed 0.1, and the angular error of line-type auxiliary graphics does not exceed 5°; the primitive attributes such as layers, colors, and line types meet the requirements of the template.

[0081] Finally, the combination similarity is calculated using the following formula: (7) in, The overall similarity between the candidate base graphic and the template base graphic. This is the average of the total similarity scores of the successfully matched subordinate graphics. If none of the subordinate graphics in the template match successfully, the candidate base graphic does not constitute a point entity. The combined similarity score must be no lower than the first threshold. At that time, the candidate base graphic and its successfully matched subordinate graphic are identified as a point entity. For example, a value of 0.75 is used. When the point entity template does not include attached graphics, the total similarity is used as the similarity between the graphic to be identified and the point entity template, and the first threshold is used for determination.

[0082] Each identified point entity generates a point entity record, including the point entity identifier, resource type, scaling factor, coordinates, and a list of vector elements constituting that point entity. The coordinates are taken from the center of the bounding box of the base graphic. Alternatively, the coordinates of each point entity can be recalculated using the center of the coordinates of all identified point entities as the origin to eliminate the influence of overall map offset.

[0083] The names of point entities are obtained through nearest neighbor matching. The text elements participating in the matching are those in the routing map whose content conforms to the node naming rules. The node naming rules can be configured according to design units, for example, a number starting with "#", followed by a number, and may include a "-number" suffix. Matching is performed in rounds: in each round, for each unpaired point entity, the text element T closest to it is identified among the unpaired text elements. Then, the point entity closest to text element T is identified among the unpaired point entities. If the latter is the current point entity, the two are paired, and the content of text element T is used as the name of the point entity. After one round, unpaired point entities and text elements enter the next round, until no new pairings are generated in a certain round. The distance is the Euclidean distance between the coordinates of the point entity and the insertion point of the text element. Preferably, text elements whose distance from a point entity exceeds a preset multiple of the point entity's feature size are not included in the point entity's matching. Nearest neighbor matching ensures that a text element is paired with at most one point entity, preventing multiple point entities from being paired with the same text when numbered text is dense. After setting a distance upper limit, if a node symbol is missed, its numbered text will not be paired with other distant point entities and will remain unpaired, which can be used as an anchor point in step S105.

[0084] Line entities are identified through spatial connections between line features and point entities. Line features participating in spatial connections can be filtered by layer, line type, or block name; for example, only line features on the pipe layer can be retained, excluding line segments that constitute node symbols. Let the first buffer distance be... , For example, take 1.0 drawing unit. For each point entity, use... To buffer the distance, perform a buffer intersection query on its base graphic to obtain the intersecting line features. For each intersecting line feature, check its two endpoints: if the endpoints fall on a certain point entity's base graphic... When the endpoint is within the buffer, it is connected to the point entity; when the endpoint is not within the buffer of any point entity, the endpoint is searched for if its distance does not exceed a certain threshold. For other line features, continue checking their other endpoints along the line feature, while recording the intermediate line features passed through and the identifiers of visited line features. Visited line features are not re-entered into the recursion. This recursion continues until a point entity is reached, there are no more line features to search for, or the recursion depth limit is reached (e.g., 20 levels). When the two ends of a line feature and its intermediate line features are connected to two different point entities, these line features are merged into one line entity, with the start and end points of the line entity being the two point entities respectively. Recursive search is used to handle situations where pipe segments are interrupted by fold points or drawn from multiple line features connected end-to-end. When pipe segments of the same project are drawn across drawing sheets, paired separator points are searched according to the configured separator text. Connecting lines are generated based on the closest distance, connecting the pipe segments across drawing sheets.

[0085] The attributes of line entities are extracted through spatial position relationships. For example, the text element that is closest to the line entity and whose content conforms to the length format is used as the length label of the line entity, and the text element whose content conforms to the hole number format is used as the hole number of the line entity.

[0086] The length of a line entity is taken from the length label value first; if there is no length label, the geometric length of the line entity is taken as the product of the drawing scale, which can be read from the scale label in the drawing label.

[0087] A routing topology graph is constructed using point entities as nodes and line entities as edges. Each edge records the identifier and length of the corresponding line entity, and the routing topology graph is stored in the form of an adjacency list.

[0088] Step S103: Extract records from the associated map sheet and generate a path constraint for each record.

[0089] When the associated map sheet includes a pipe occupancy map sheet, the pipe occupancy record is extracted from the pipe occupancy map sheet. Pipe occupancy maps are typically drawn as pipe cross-section diagrams or pipe occupancy tables, with the pipe segment identified by a routing subkey, such as "#006-#007" or "#006-#006-1". Each pipe occupancy record includes the start name, end name, pipe number, occupancy status, and name of the fiber optic cable installed. The start and end names are obtained by parsing the routing subkey: each connector in the routing subkey is used as a split point, and the split results where both sides conform to the node naming rules are used as the start and end names. For example, the routing subkey "#006-#006-1" is split at the first connector to obtain "#006" and "#006-1", both of which conform to the node naming rules; splitting at the second connector yields "#006-#006" and "1", neither of which conform to the node naming rules. Therefore, the starting name corresponding to this route subkey is "#006" and the ending name is "#006-1".

[0090] When the associated map sheet includes an optical cable reel layout map sheet, the optical cable segment record is extracted from the optical cable reel layout map sheet. The optical cable segment name is identified by matching text elements according to a preset optical cable name pattern; the start and end names are determined based on the start and end labels near the optical cable segment name or the start and end columns in the reel layout table; the reel layout length is extracted from the "Reel Length" or "Reel Length" column in the reel layout table, or from labels conforming to the length format near the optical cable segment name, and the value in kilometers is converted to meters. Each optical cable segment record includes the optical cable segment name, model, number of cores, start name, end name, and reel layout length.

[0091] Before comparing the start point name, the end point name and the node name, name normalization is performed on both sides involved in the comparison respectively. Name normalization first performs character replacement according to the mapping table, and then performs truncation according to the truncation rule. The mapping table, for example, replaces full-width characters with half-width characters, and deletes redundant spaces and slashes; the truncation rules include "before:X", "after:X" and "suffix:X", which respectively represent taking the content before the first separator X, taking the content after the first separator X, and taking the content after the last separator X. Different truncation rules can be configured for the two sides. For example, the node text in the routing map sheet is "#006", and the name in the duct occupancy record is "#006 manhole". The mapping table replaces the full-width "#" with the half-width "#", and the side associated with the map sheet is configured with the truncation rule "before:manhole", so that both sides are normalized to "#006".

[0092] After name normalization, the node names used to establish the correspondence of the routing topology within the same project shall be unique. If the same normalized name corresponds to multiple identified nodes, any one of the nodes will not be directly selected to establish a path constraint, instead, the record will be marked as a name conflict and output for manual confirmation; the subsequent closure determination is premised on that both the start point name and the end point name have uniquely corresponding nodes.

[0093] One path constraint is generated from each record, and the path constraint includes a start point name, an end point name, a constraint type and a source record, wherein the constraint type is a duct occupancy constraint or an optical cable segment constraint.

[0094] In step S104, closure determination is performed on each path constraint.

[0095] For each path constraint, a node whose name is consistent with the start point name and a node whose name is consistent with the end point name are searched in the routing topology diagram. When both exist, a breadth-first search is carried out starting from the node corresponding to the start point name. If the node corresponding to the end point name can be reached, the path constraint is determined to be closed, and the connected path with the least number of hops obtained by the search is recorded; if the node cannot be reached, or at least one of the start point name and the end point name has no corresponding node, the path constraint is determined to be unclosed.

[0096] For the duct occupancy constraint, it may also be required that the number of hops of the connected path does not exceed the number of pipeline segments represented by the routing sub-key, and the number is usually 1. When the pipe network forms a loop, after a pipeline segment between two adjacent nodes is missed in identification, the two nodes may still be communicated with each other via other pipeline segments, and the above requirement can avoid misjudging this situation as closure.

[0097] In step S105, for each unclosed path constraint, its missing part is determined, and a defect area is determined according to the position of the element related to the missing part in the routing map sheet.

[0098] Missing parts fall into two categories. The first category consists of endpoint names without corresponding nodes in the routing topology graph; that is, at least one of the start and end names cannot be matched with a node. The second category consists of missing connected paths between two corresponding nodes; that is, both the start and end names have corresponding nodes, but the two nodes are located in different connected components of the routing topology graph. Connected components can be obtained by performing a breadth-first search on the routing topology graph.

[0099] like Figure 3 As shown, for the first type of missing part, the elements related to the endpoint name are text elements whose content is consistent with the endpoint name. The normalized text elements in the routing map that are consistent with the endpoint name and are not paired with any point entity are searched and used as anchor point 33. A circle is drawn with the insertion point of anchor point 33 as the center and... A circular defect region 34 is defined with a radius of 34, where, The average feature size of the identified point entities in the routing map is k, which is a preset multiple. For example, k can be 2, and it can generally be between 1.5 and 3.

[0100] When multiple text elements meet the conditions, the defect areas are determined separately for each. If no anchor point is found, or if there are no identified point entities in the routing map that can be used to calculate the average feature size, the path constraint is not re-identified in this way and is directly output for manual confirmation. If a node symbol has been identified and paired with the text, a corresponding node exists in the routing topology map. Therefore, unpaired text elements with the same name usually have unidentified node symbols nearby, such as... Figure 3 Unidentified manhole symbol 32. Figure 3 In the diagram, 31 represents the identified point entities, and 35 represents the line elements.

[0101] like Figure 4 As shown, for the second type of missing part, the features related to the missing connected path are the features in the connected components where the two corresponding nodes each belong. For each connected component, the point entities, line entities, and line features that intersect with the point entities in that connected component but are not identified as line entities are considered as candidate features of that connected component. The latter type of line feature usually has one end connected to a node symbol and the other end not connected to any node symbol, such as... Figure 4Line features 43 that are not identified as line entities. Between the candidate features of the first route connected component 41 and the candidate features of the second route connected component 42, the closest pair of features is determined, and the shortest connection 44 between them is calculated. The shortest connection can be obtained using functions provided by the spatial database, such as the ST_ShortestLine function in PostGIS; when there are many candidate features, a nearest neighbor search can be performed using a spatial index first. Using the shortest connection 44 as the center line, the area is extended to both sides by half a preset width to obtain a strip-shaped defect region 45. The preset width w is, for example, the first buffer distance. 3 times.

[0102] Step S106: Re-identify and spatially connect the starting point entities of the vector features intersecting with the defect area to obtain candidate entities by relaxing the parameters.

[0103] Vector features intersecting the defect area are selected from the routing map. Vector features that already constitute identified point entities are no longer included in the point entity identification process. For the first type of missing portion, a second threshold is used. Replace the first threshold The selected vector features are re-identified as point entities. The identified point entities are paired with text features within the defect area according to the aforementioned nearest neighbor matching. Those that cannot be paired are discarded, thus obtaining candidate point entities. Subsequently, the selected line features are used as objects, with a first buffer distance. Or second buffer distance Re-execute the spatial connection to obtain candidate line entities connected to the candidate point entities. For the second type of missing portion, take the selected line features as the object and use the second buffer distance. Re-execute the spatial connection to obtain candidate line entities, and if necessary, simultaneously use a second threshold. Re-identify point entities. When re-performing spatial connections, only the selected line features are used as the starting point for the recursive search. To determine whether a line entity can be formed connecting two different point entities, the recursive search can extend beyond the defective region along the line features connected to its beginning and end. However, the line features extended beyond the defective region are only used for connectivity tracing and are not re-identified as point entities using the second threshold for vector features outside the region. When the same path constraint identifies multiple defective regions, re-identification is performed within these defective regions. Candidate point entities and candidate line entities are collectively referred to as candidate entities.

[0104] Second threshold Second buffer distance The following method is used to determine the similarity between each point entity template and other categories of graphics in the template library during the template library calibration phase. The maximum similarity is then taken. Other categories of graphics include point entity templates of other resource types and pre-registered non-resource symbols. When a point entity template includes ancillary graphics, the similarity is calculated as a combination similarity. Second threshold. In greater than and less than It takes values ​​within a range, for example: (8) When re-identifying within the defect area, both the basic threshold and the secondary similarity threshold are set according to... and The scaling ratio is reduced proportionally, while the consistency of the scaling ratio, relative position, and primitive attribute conditions remain unchanged. Taking a manhole template as an example... The maximum similarity score is 0.75, representing the combined similarity between the manhole template and other types of graphics such as handhole templates and solid round holes. If it is 0.60, then The base similarity threshold and the associated similarity threshold are reduced from 0.8 to 0.8 × 0.65 / 0.75, which is approximately 0.69.

[0105] When there are at least two identified point entities in the routing map, let the minimum distance between any two identified point entities be... .when > At that time, there exist conditions that simultaneously satisfy a condition greater than the first buffer distance. and less than Second buffer distance The second buffer distance is determined, for example, by the following formula: (9) when For 1.0, When the value is 9.6, d2 is obtained as 2.0 according to equation (9). When there are fewer than two identified point entities, or ≤ At that time, the method of not adopting the method of A second buffer distance is defined; at this point, only the second threshold is used to re-identify point entities, and if necessary, the corresponding path constraints and defect areas are output for manual confirmation.

[0106] Greater than Therefore, other categories of graphics already registered in the template library will not reach the combined similarity judgment threshold of that type of point entity under the template calibration conditions; in the case of using the second buffer distance, Less than Half of the buffers of any two identified point entities do not overlap, and the endpoint of a line feature falls into at most one buffer of an identified point entity; the scaling ratio consistency, relative position and primitive attribute conditions remain unchanged, and the relaxation only applies to similarity or spatial connectivity tolerance, while the structural verification of symbols is still performed according to the original conditions.

[0107] Step S107: After adding the candidate entity to the routing topology graph, re-determine the closure of the path constraint; if the path constraint is closed, retain the candidate entity and update the routing topology graph accordingly; if the path constraint is still not closed, discard the candidate entity.

[0108] When retaining candidate entities, you can either retain all candidate entities re-identified within the defective region, or only retain candidate entities located on the connected paths that close the path constraint, discarding the rest. Retained candidate entities are marked with a recovery flag, which records the defective region where the candidate entity is located, the relaxation parameters used during re-identification, and the path constraint that led to its retention. When discarding candidate entities, the path constraint and its defective region are output for manual verification.

[0109] As one specific embodiment, such as Figure 2 As shown, in this embodiment, steps S104 to S107 can be performed according to steps S201 to S210.

[0110] Step S201: Take out one path constraint to be determined in the order of first the pipe hole occupancy constraint and then the optical cable segment constraint.

[0111] Step S202: Determine whether the starting point name and the ending point name of the path constraint both have corresponding nodes in the routing topology diagram. If not, proceed to step S203; if yes, proceed to step S204.

[0112] Step S203: Using unmatched text elements with the same name as anchor points, determine the circular defect area, and proceed to step S206.

[0113] Step S204: Determine whether there is a connected path between two corresponding nodes. If yes, determine that the path constraint is closed and proceed to step S210. If no, proceed to step S205.

[0114] Step S205: Using the shortest line connecting the closest pair of candidate elements of two connected components as the center line, determine the strip-shaped defect area, and proceed to step S206.

[0115] Step S206: Re-identification is initiated from vector features intersecting the defect area: Re-identification of point entities is limited to vector features intersecting the defect area; for spatial connectivity, line features intersecting the defect area are used as the starting point for recursive search, and connectivity tracing is allowed to extend beyond the defect area along line features that are connected end to end; during re-identification, at least one of a second threshold and a second buffer distance is used to obtain candidate entities.

[0116] Step S207: Add the candidate entity to the routing topology graph and determine whether the path constraint is closed. If yes, proceed to step S208; otherwise, proceed to step S209.

[0117] Step S208: Retain candidate entities and attach recovery tags, update the routing topology graph, and re-determine the closure of the remaining unclosed path constraints, then proceed to step S210.

[0118] Step S209: Discard the candidate entity, output the path constraint and its defect area, and proceed to step S210.

[0119] Step S210: Determine whether all path constraints have been processed. If not, return to step S201. If yes, execute step S108.

[0120] Each path constraint is re-identified at most once; in step S208, the remaining path constraints are only re-determined for closure, and the already executed re-identification is not triggered again. Therefore, the above process ends within a finite number of steps. Pipe occupancy constraints generally correspond to a single pipe segment between two adjacent nodes. The defect area located by this is relatively small, so it is processed before optical cable segment constraints. The correspondence between pipes and pipe segments established after the pipe occupancy constraints are closed can also serve as the basis for subsequent selection of optical cable segment routes.

[0121] Step S108: Write the point entities and line entities in the routing topology map into the resource management system, and for each closed path constraint, establish the topological relationship between the corresponding pipe or optical cable segment and the line entity on the connected path according to its connected path.

[0122] Before writing, the identification results are checked for topological consistency, coding standardization, and attribute integrity: the start and end points of line entities must be existing point entities; resource codes must conform to naming rules and be unique; required attributes must not be empty, and numerical attributes must be within a reasonable range. After steps S104 to S107, all path constraints that can be closed have been closed. Path constraints that are not yet closed, along with their defective areas, are output and processed after manual confirmation.

[0123] Before writing, a topology review can be performed between the routing topology map and existing routing data in the resource management system. During the topology review, a side-view model is constructed from the routing topology map, and a system side-view model is constructed from existing routing data corresponding to the communication project in the resource management system, such as laying segment data entered during the design phase. Both models use routing points as nodes and pipe segments as edges. Node labels indicate resource type, and edge attributes include the number of pipe holes and length. If the resource type identifiers on both sides are inconsistent, they are first unified according to the mapping configuration. To determine whether two graph models are isomorphic, a backtracking method is used to search for one-to-one mappings between nodes: Nodes on the graph side are sorted by degree from largest to smallest. For each node, candidate nodes with the same node label and degree are selected from the unmapped nodes on the system side, prioritizing those with the same name. After each selection, it is checked whether each edge between the selected node and the mapped node has a corresponding edge on the system side, whether the number of holes on the corresponding edge is equal, and whether the difference in length is within a preset tolerance (e.g., the larger of 1 meter and 1% of the length). The same reverse check is performed on the system side. If the check passes, candidate nodes are selected for the next node; if the check fails, the current selection is canceled and the next candidate node is tried. If all candidate nodes fail, the process backtracks to the previous node. When a mapping covering all nodes is found, the two graph models are considered isomorphic.

[0124] When a mapping covering all nodes cannot be found, the two graph models are determined to be non-homogeneous. Based on the mapping with the most mapped nodes during the search process, unmapped nodes and edges whose endpoints are mapped but lack corresponding edges on the other side or have inconsistent edge attributes are considered inconsistent nodes and edges. For inconsistent nodes and edges, the portion involving entities with restoration markers is output as an identification restoration item, along with the defect area and relaxation parameters recorded in the restoration markers, for reviewers to verify against the drawings; the remaining portions are output as topological inconsistencies. Even when determined to be isomorphic, entities with restoration markers can be listed separately for reviewers. To control computation time, a step limit can be set for backtracking searches. Routing for individual projects. Figure 1 It typically contains dozens to hundreds of nodes, and after pruning by node labels and degrees, the search volume is relatively small.

[0125] When writing pipeline resources, each point entity generates a point resource such as a manhole, handhole, or utility pole, and each line entity generates a pipeline segment resource. The pipeline segment resource includes a unique identifier, resource type identifier, creation marker, initial attributes, and pipe hole information, and uses the point resources corresponding to the start and end points of the line entity as its starting and ending resources, respectively.

[0126] For each closed pipe occupancy constraint, the pipe occupancy record from which it originates is associated with the line entity on the connected path, that is, the pipe belongs to the pipe segment corresponding to the line entity. The pipe number, occupancy status and name of the fiber optic cable are written into the resource management system to establish the topological relationship between the pipe and the pipe segment.

[0127] For each closed optical cable segment constraint, the route of the optical cable segment is determined as follows: Using the two corresponding nodes of the optical cable segment constraint as the start and end points respectively, a depth-first search is performed on the routing topology graph to enumerate candidate paths between them that do not contain duplicate nodes. During the search, a set of visited nodes is maintained to avoid path loops; when the cumulative length of some paths exceeds the disk length... With respect to preset tolerance When the sum of the two is equal, that is ,in Stop searching along this part of the path, preset tolerance. This is used to accommodate length differences caused by reservations, inventory build-ups, etc.; duplicate candidate paths are deduplicated using strings concatenated from node sequences. For each candidate path, the occupancy rate E and length deviation are calculated. : (10) Where n is the total number of entities online on the candidate path. The number of line entities whose names contain the name of the optical cable segment in the duct occupancy record associated with the candidate path, where L is the length of the candidate path, i.e., the sum of the lengths of all line entities on the path. The candidate path with the highest occupancy evidence rate E is selected. If the occupancy evidence rates are the same, the one with the smallest length deviation δ is selected as the route for the optical cable segment. The occupancy relationship between the optical cable segment and the ducts in the corresponding duct segments of each line entity is established along the route: For each line entity on the route, the optical cable segment is associated with the duct in the corresponding duct segment where the name of the optical cable contains the name of the optical cable segment; if no such duct exists, the optical cable segment is recorded as passing through the duct segment, and supplementary duct information is prompted in the output results.

[0128] like Figure 6As shown, taking a ring network as an example, the lengths of the pipe segments between nodes A, B, C, and D are as follows: AB is 200m, BD is 220m, AC is 260m, and CD is 245m. The optical cable segment GL-01 starts at A and ends at D, with a cable tray length of 450m and a preset tolerance of 90m. Depth-first search yields the first candidate path 61 and the second candidate path 62. The first candidate path 61 is ABD with a length of 420m, and the second candidate path 62 is ACD with a length of 505m. The cumulative lengths of both do not exceed 540m. Pipe occupancy records show that GL-01 is routed through pipe segments AB and BD, while other optical cables are routed through pipe segments AC and CD. Therefore, the occupancy evidence rate for the first candidate path 61 is 1.0, and the occupancy evidence rate for the second candidate path 62 is 0. The first candidate path 61 is selected as the route for GL-01, with a length deviation of 30m. If no pipe holes occupy the map area, the occupancy evidence rate of both candidate paths is 0. Therefore, the path is selected based on length deviation. The length deviation of the first candidate path 61 (30 m) is less than the length deviation of the second candidate path 62 (55 m), resulting in the same outcome. Both candidate paths have a jump count of 2, and the jump count alone cannot distinguish between them.

[0129] When an optical cable segment is written into the resource management system, its ownership is determined as follows: Using the two corresponding nodes after the optical cable segment constraint closure as the endpoints of the segment, the optical cable splices located at the endpoints are identified: if the endpoint node itself is an optical cable splice, or if there is a point entity with the resource type of optical cable splice within the manhole where the endpoint node is located, that point entity is designated as the optical cable splice at that endpoint. The resource type of the optical cable splice can include multiple splice type identifiers. Subsequently, a disjoint-set data structure is constructed. Initially, each optical cable segment is a separate set. The optical cable splices are traversed, and optical cable segments with the same endpoint are merged into the same set. The search operation uses path compression, and the merging operation is performed by rank. After merging, each set becomes a connected component.

[0130] For each connected component, determine its affiliation in the following order: If there is an optical cable segment with an existing network identifier, the entire connected component is determined to be reused, and the existing network identifier is used. If there is no optical cable segment with an existing network identifier, but there is a reused optical cable connector, and an existing optical cable is found in the resource management system through the reused optical cable connector, then the entire connected component is determined to be reused, and the network identifier of the existing optical cable is used. If neither of the above two steps is satisfied, the entire connected component is considered newly created, and a new network identifier, such as a universally unique identifier, is generated, shared by all optical cable segments within the connected component. Here, an optical cable segment with an existing network identifier refers to an optical cable segment that already exists in the resource management system, and a reused optical cable connector refers to an optical cable connector marked as reused in the drawings or already existing in the resource management system. If multiple different existing network identifiers appear in the same connected component, the connected component is output for manual confirmation.

[0131] like Figure 7 As shown, optical cable segment , , Using the same optical cable connector 73 as the endpoint, they are merged into the first optical cable segment connectivity component 71. Since none of the optical cable segments have existing network identifiers, and optical cable connector 73 is not a reused optical cable connector, the first optical cable segment connectivity component 71 is determined to be newly created. The three optical cable segments share a newly generated network identifier. Figure 7 Zhong 75 is one of the optical cable segments; optical cable segment , Using the reused optical cable connector 74 as the endpoint, the components are merged into the second optical cable segment connected component 72. Existing optical cables are found through the reused optical cable connector 74, and the second optical cable segment connected component 72 is determined to be reused. The optical cable connector is determined based on the closed node; the manhole where the connector is located will not be missed due to misidentification. Therefore, optical cable segments sharing the same connector will not be split into different connected components and thus obtain different network identifiers.

[0132] Finally, establish the inclusion relationship between the pipe section and the manhole / handhole, as well as the occupancy relationship between the optical cable section and the pipe hole, and update the status of the project to "connected to the network".

[0133] As a specific example of pipeline engineering implementation, the method of this embodiment will be further explained.

[0134] The submitted CAD drawings for this project are in DWG format, containing cover sheets, routing sheets, and pipe / hole occupancy sheets within the same file. After conversion and parsing in step S101, vector elements corresponding to 326 straight lines, 185 polylines, 412 text elements, 58 circles, and 23 fill elements are obtained. After removing irrelevant layers and zero-length lines, 780 valid vector elements are obtained. The routing sheet and pipe / hole occupancy sheet are then determined based on the drawing frame and keywords.

[0135] In step S102, 11 manholes, 8 handholes, and 16 pipe segments were identified in the routing map. The average feature size of the identified point entities was 4.2, and the minimum distance between any two identified point entities was 9.6. In step S103, 36 pipe occupancy records were extracted from the pipe occupancy map, generating 36 pipe occupancy constraints.

[0136] After step S104, 30 terms are closed and 6 terms are not closed. Among the unclosed constraints, 4 terms come from routing subkeys "#006-#007" and "#007-#008", where the name "#007" has no corresponding node in the routing topology graph; 2 terms come from routing subkeys "#012-#013", where the two corresponding nodes are located in different connected components.

[0137] For the name "#007" without a corresponding node, the text element containing "#007" in the routing map is not paired with any point entity. Using this text element as the anchor point and a radius of 2 × 4.2 = 8.4, a defect area is determined. The manhole symbol within this area is drawn incorrectly; its base graphic has a total similarity of 0.70 with the template base graphic, and the total similarities of its two subordinate graphics are 0.74 and 0.72 respectively. The combined similarity is 0.5 × 0.70 + 0.5 × 0.73, approximately 0.72. Based on a base threshold of 0.8, this base graphic failed to become a candidate base graphic in the initial identification. Within the defect area, the base threshold and subordinate similarity threshold decrease to 0.69, and the combined similarity of 0.72 is not lower than the second threshold of 0.65, resulting in a candidate point entity, which is then paired with the text "#007". After re-performing spatial connection with the first buffer distance, two candidate line entities connecting "#006" and "#007", and "#007" and "#008" are obtained. After adding the candidate entity, all four path constraints mentioned above are closed, and the candidate entity is retained.

[0138] For the routing subkey "#012-#013", node "#013" constitutes a single connected component. Among the candidate elements of the first connected component 41, there is a line element that intersects with the manhole symbol of "#012" but is not identified as a line entity. The distance between its end and the manhole symbol of "#013" is 1.6, which is greater than the first buffer distance of 1.0. Using the shortest connection between the two as the center line and a width of 3.0, the defect area is determined. Spatial connection is re-executed with a second buffer distance of 2.0, resulting in a candidate line entity connecting "#012" and "#013". The above two path constraints are closed, and the candidate entity is retained.

[0139] After correction, the routing topology map includes 12 manholes, 8 handholes, and 19 pipe segments, with all 36 pipe occupancy constraints closed. During topology review, the drawing side map model and the system side map model are isomorphic. The recovered 1 point entity and 3 line entities are listed separately as identification recovery items for review. After being written into the resource management system, 20 point resources and 19 pipe segments are generated, and all 36 pipes are associated with their corresponding pipe segments. If the second threshold and second buffer distance are applied to the entire routing map, line elements at other locations will also be connected to point entities with a buffer distance of 2.0, such as pipes that pass near manholes but do not connect to them. In this embodiment, only vector elements within the two defect areas are used to initiate re-identification and spatial connection using relaxed parameters. The extension of spatial connection outside the defect area is only used for connectivity tracing, and candidate entities are only retained after path constraint review.

[0140] Taking another optical cable project as an example, this project includes three optical cable segments, all of which share the same optical cable connector as their endpoints. After the optical cable segments corresponding to the three segments are closed by constraint, they are merged into a single connected component through a lookup table. None of the optical cable segments have any existing network identifiers, and no existing optical cables are found through the optical cable connector. Therefore, the entire connected component is determined to be newly created, and the three optical cable segments share a newly generated network identifier.

[0141] As one specific embodiment, such as Figure 8 As shown, the present invention provides a system 100 for the identification and resource network access of CAD drawings for communication engineering, including a drawing sheet determination module 101, a route identification module 102, a constraint generation module 103, a closure determination module 104, a defect correction module 105, and a network access module 106.

[0142] Map sheet determination module 101 is used to obtain vector elements obtained by parsing communication engineering CAD drawings, and to determine the routing map sheet and associated map sheet from the vector elements.

[0143] The route identification module 102 is used to identify point entities based on a first threshold, identify the spatial connection of line entities based on a first buffer distance, and construct a route topology graph with point entities as nodes and line entities as edges. The acquisition of point entity names, template matching, and spatial connection processes adopt the corresponding processing methods in Embodiment 1.

[0144] The constraint generation module 103 is used to extract records including start-point and end-point names from associated map sheets, and generate path constraints from each record that require a connected path between corresponding nodes.

[0145] The closure determination module 104 is used to determine the closure state of the path constraint based on whether the corresponding node exists and whether there is a connected path between the corresponding nodes.

[0146] The defect correction module 105 is used to determine the defect region based on the missing part of the unclosed path constraint, take the vector elements that intersect with the defect region as the local re-identification object and start spatial connection, and re-identify by at least one of the following methods: lowering the similarity judgment threshold and increasing the spatial connection buffer distance to obtain candidate entities; after adding the candidate entities to the routing topology map, re-execute the closure judgment, and retain or discard the candidate entities according to the re-closure judgment result.

[0147] The network entry module 106 is used to write the point entities and line entities in the updated routing topology map into the resource management system 200, and associate the resources corresponding to their source records with the line entities on the corresponding connected paths according to the closed path constraints.

[0148] The specific implementation method of the module can be referred to the corresponding steps in the embodiment, which will not be repeated here.

[0149] The above modules can be implemented through software, hardware, or a combination of both. The map sheet determination module 101 may include a conversion unit that converts DWG format to DXF format and a parsing unit that parses DXF files into vector features. The vector features are stored in a spatial database. The route identification module 102, constraint generation module 103, closure determination module 104, and defect correction module 105 read and process the vector features through the spatial query function provided by the spatial database.

[0150] As one specific embodiment, such as Figure 9 As shown, an electronic device 300 is provided, including a processor 301, a memory 302, and a communication interface 303. The processor 301, memory 302, and communication interface 303 are interconnected via a bus 304. The memory 302 stores a computer program. When the processor 301 executes the computer program, it implements the method for recognizing and integrating communication engineering CAD drawings into the network.

[0151] The processor 301 can be a central processing unit, a graphics processing unit, or other general-purpose processor, or a digital signal processor, an application-specific integrated circuit, or a field-programmable gate array, etc.; the memory 302 can include random access memory, read-only memory, flash memory, or hard disk, etc.; the communication interface 303 is used to communicate with external devices such as resource management systems and file storage servers.

[0152] The above methods and steps can be executed by computer program instructions. Each module in the system can be implemented by software or by hardware circuits capable of performing corresponding data processing functions.

[0153] The above-described specific embodiments are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Various modifications, substitutions, and improvements made by those skilled in the art to the technical solutions of the present invention based on the provided textual description and drawings, without departing from the design concept and spirit of the present invention, should all fall within the scope of protection of the present invention. The scope of protection of the present invention is determined by the claims.

Claims

1. A method for identifying and integrating CAD drawings used in communication engineering into the network, characterized in that, include: The vector elements obtained from the parsed CAD drawings of communication engineering are obtained, and the vector elements include line elements and text elements; The routing map sheet and associated map sheet are determined from the vector elements, wherein the associated map sheet includes at least one of the pipe occupancy map sheet and the optical cable reel distribution map sheet; In the routing map, vector elements are matched for graphic similarity based on point entity templates to identify point entities, and text elements are associated with the point entities to obtain the names of the point entities; Using a first buffer distance as a tolerance, spatial connections are made between line elements and point entities, and line entities connecting different point entities are identified based on the spatial connection results. A routing topology graph is constructed using the point entities as nodes and the line entities as edges, wherein the first identification of the point entities uses a first threshold as a similarity determination threshold. Records are extracted from the associated map sheets, and each record includes a start point name and an end point name; A path constraint is generated from each of the records, the path constraint requiring that there is a connected path between the node corresponding to the starting point name and the node corresponding to the ending point name in the routing topology graph; For each of the path constraints, a closure determination is performed. When both the starting point name and the ending point name have corresponding nodes in the routing topology graph, and there is a connected path between the two corresponding nodes, the path constraint is determined to be closed. Otherwise, the path constraint is determined to be unclosed. For unclosed path constraints, the defect area is determined based on the position of its missing part in the routing map. The vector elements intersecting with the defect area are used as local re-identification objects and spatial connections are initiated. During re-identification, at least one of the following relaxation methods is adopted: the similarity judgment threshold is reduced from the first threshold to the second threshold; the buffer distance of spatial connection is increased from the first buffer distance to the second buffer distance to obtain candidate entities. After adding the candidate entity to the routing topology graph, the closure determination of the path constraint is performed again. If the path constraint is closed, the candidate entity is retained and the routing topology graph is updated. If the path constraint is not closed, the candidate entity is discarded. Write the point entities and line entities in the updated routing topology graph into the resource management system, and associate the resources corresponding to their source records with the line entities on the corresponding connected paths according to the closed path constraints.

2. The method according to claim 1, characterized in that, Associating text elements with point entities includes: performing neighbor matching in a loop, determining the nearest text element to each unpaired point entity in each round, and pairing the two when the nearest point entity to the text element is the point entity, until no new pairings are generated. When the missing part is an endpoint name that has no corresponding node in the routing topology map, select a text element that does not match any point entity from the text elements whose content is consistent with the endpoint name as an anchor point, take the anchor point as the center, and use a preset multiple of the average feature size of the identified point entities in the routing map as the radius to determine the defect area. When the missing part is a missing connected path between two corresponding nodes, the point entities and line entities in the connected components of the two corresponding nodes are taken as candidate elements, and the line elements that intersect with the point entities in the connected component and have not yet been identified as line entities are also taken as candidate elements. The pair of elements with the closest distance between the candidate elements of the two connected components are determined, and the shortest connection between the pair of elements is taken as the center line, and the defect area is determined according to a preset width.

3. The method according to claim 2, characterized in that, The text elements are obtained by performing high-fidelity parsing on the CAD drawings of the communication engineering project. The high-fidelity parsing includes: The block reference is disassembled and text elements generated by invisible attributes are filtered out; multi-byte encoded text represented by MIF escape sequences is decoded back to the original text. Before comparing the starting point name and the ending point name with the name of the point entity and the content of the text element, name normalization is performed on both sides of the comparison. The name normalization includes performing character replacement according to the mapping table and performing truncation according to the truncation rules. Each side adopts its own configured truncation rules.

4. The method according to claim 1, characterized in that, The second threshold is greater than the maximum similarity between the point entity template and other categories of graphics in the template library, and less than the first threshold. The other categories of graphics include point entity templates of other resource types and pre-registered non-resource symbols. The second buffer distance is greater than the first buffer distance, but less than half of the minimum distance between any two identified point entities in the routing map.

5. The method according to claim 4, characterized in that, The point entity template includes a base graphic and an auxiliary graphic. The graphic similarity matching includes: performing geometric normalization on the graphic to be identified and the template graphic, and calculating the total similarity based on the normalized geometric features; first determining candidate base graphics based on the total similarity, then matching auxiliary graphics in the neighborhood of the candidate base graphics, and forming a combined similarity based on the candidate base graphics and the successfully matched auxiliary graphics. Both the first threshold and the second threshold are used as thresholds for determining the combined similarity; When re-identifying within the defective region, the similarity threshold used for basic graphic matching is reduced, and the similarity threshold used for auxiliary graphic matching is also reduced, while keeping the scaling ratio consistency condition, relative position condition, and primitive attribute condition unchanged.

6. The method according to claim 1, characterized in that, The associated map sheet includes the duct hole occupancy map sheet and the optical cable reel distribution map sheet. The duct hole occupancy record also includes the name of the optical cable being laid, and the optical cable segment record also includes the optical cable segment name and the reel length. First, process the path constraints generated by the pipe occupancy record and establish the pipe occupancy relationship corresponding to the closed path; then process the path constraints generated by the optical cable segment record. For the closed path constraint generated by the optical cable segment record, enumerate the candidate paths that do not have duplicate nodes between their two corresponding nodes, and calculate the occupancy evidence rate and length deviation of each candidate path. The occupancy evidence rate is determined based on the number of line entities on the candidate path that have the name of the optical cable segment that matches the name of the optical cable segment, and the length deviation is determined based on the difference between the length of the candidate path and the length of the reel. Candidate paths with higher evidence occupancy rates are selected first. If the evidence occupancy rates are the same, candidate paths with smaller length deviations are selected as the connected paths on which the topological relationship is based.

7. The method according to claim 6, characterized in that, The write resource management system includes: using two corresponding nodes of the closed path constraint generated by the optical cable segment record as the endpoints of the optical cable segment, and identifying the optical cable joint located at the endpoint; Based on the relationship of sharing the same optical cable connector, the optical cable segment is divided into connected components; For each connected component, if at least one of the following reuse conditions is met, the connected component is determined to be reused and the corresponding existing network identifier is retained: The first condition for reuse is that there is an optical cable segment with an existing network identifier in the connected component. The second condition for reuse is that there is a reused optical cable connector in the connected component, and the reused optical cable connector is associated with an existing optical cable. The reused optical cable connectors include optical cable connectors marked as reused in the drawings and optical cable connectors that already exist in the resource management system. When neither the first nor the second reuse condition is met, the connected component is determined to be newly created, and the optical cable segment therein is assigned the same newly generated network identifier.

8. The method according to claim 1, characterized in that, Also includes: A recovery marker is attached to the retained candidate entity; the recovery marker records the defect region where the candidate entity is located, and records the relaxation parameters used during re-identification and the path constraints that allow the candidate entity to be retained; Before writing to the resource management system, graph models are constructed based on the routing topology map and the existing routing data corresponding to the communication project in the resource management system, and the topology consistency of the two graph models is compared based on node labels and edge attributes. When inconsistencies exist in the comparison results, the inconsistent portions involving entities with the recovery markers will be output separately from the remaining inconsistent portions.

9. A system for identifying and integrating CAD drawings for communication engineering into the network, used to implement the method for identifying and integrating CAD drawings for communication engineering into the network as described in any one of claims 1-8, characterized in that, include: The map sheet determination module is used to obtain vector elements from the parsed CAD drawings of communication engineering projects and to determine the routing map sheets and associated map sheets; The route identification module is used to identify point entities based on a first threshold, identify the spatial connection of line entities based on a first buffer distance, and construct a route topology graph with point entities as nodes and line entities as edges. The constraint generation module is used to extract records including start-point names and end-point names from the associated map sheets, and generate path constraints from each record that require a connected path between the corresponding nodes. The closure determination module is used to determine the closure status of the path constraint based on whether the corresponding node exists and whether there is a connected path between the corresponding nodes. The defect correction module is used to determine the defect area based on the missing part of the unclosed path constraint. Vector elements intersecting the defective region are used as local re-identification objects and spatial connections are initiated; at least one parameter relaxation method is used for re-identification, including lowering the similarity judgment threshold and increasing the spatial connection buffer distance; After obtaining the candidate entities, the candidate entities are processed according to the re-closure determination result; The network entry module is used to write the point entities and line entities in the updated routing topology map into the resource management system, and associate the resources corresponding to their source records with the line entities on the corresponding connected paths according to the closed path constraints.

10. The system according to claim 9, characterized in that, The defect correction module includes: The endpoint missing correction unit is used to determine the anchor point from text elements whose content is consistent with the endpoint name but not paired with the point entity when there is an endpoint name without a corresponding node in an unclosed path constraint, and to determine the defect area based on the anchor point. The path missing correction unit is used to determine the defect region based on the spatial position between the connected components of the two corresponding nodes when there is no connected path between the two nodes corresponding to the unclosed path constraint. The local re-identification unit is used to perform local re-identification for the defective area, add the obtained candidate entities to the routing topology map, and re-perform the closure determination; if the re-closure determination passes, the candidate entities are retained; if the re-closure determination fails, the candidate entities are discarded.