Three-dimensional building model generation method based on CAD graph layer

Abstract

The invention discloses a three-dimensional building model generation method based on a CAD (Computer Aided Design) layer, which comprises the following steps of: importing CAD files of. Dwg and. Dxf suffixes by a user, performing layer conversion, filtering and validity judgment, acquiring a data set name by using an analytical algorithm, and generating a building component object and a user-defined field; the method comprises the following steps of: performing line-to-surface conversion, copying a surface data set, performing field assignment, merging the data set, performing special component re-assignment, removing a useless data set, generating a three-dimensional scene by utilizing a SuperMap component, performing cache switching, and finally issuing a service for a user to check a model online. The method aims at solving the problems that when a SuperMap component processes a CAD file, the layer name is lost, and general information cannot be obtained through the layer name.

Description

Technical Field

[0001] The invention belongs to the technical field of computer-aided design, and relates to a method for generating a three-dimensional building model based on a CAD layer. Background Art

[0002] In the fields of architectural design and engineering, with the rapid development of digital technology, 3D modeling and visualization technologies have become a key industry trend. CAD software, as a traditional design tool, has a deep foundation in 2D drawing creation. The .dwg and .dxf files it generates contain a wealth of architectural design information. In the architectural design and engineering fields, a vast amount of design data has long been accumulated in the form of 2D CAD drawings. These 2D drawings contain a wealth of architectural design information, such as building layouts and dimensioning. Directly undertaking 3D modeling means discarding these existing data resources and re-collecting and re-entering data, which not only wastes resources but also increases repetitive work. CAD software is renowned for its high-precision drawing capabilities, ensuring high accuracy of dimensions and positional information in 2D drawings. Converting this precise 2D data into 3D can largely ensure the accuracy of 3D models in terms of spatial position and component dimensions. Geographic Information System (GIS) platforms, such as SuperMap, with their powerful 3D visualization and analysis capabilities, play a vital role in processing geospatial data and displaying 3D models, providing new technical means for constructing 3D architectural models. Combining CAD files with SuperMap components to achieve efficient conversion from two-dimensional drawings to three-dimensional models is of great significance to improving the efficiency of architectural design and management.

[0003] However, in the actual process of using the SuperMap component to process CAD files, there are some problems that need to be solved urgently. Take the patent name: "A layer classification method for converting architectural drawings into three-dimensional BIM models, application number: 201811035500.3" as an example. Although this method is committed to solving the layer classification problem when converting architectural drawings to three-dimensional BIM models, it does not involve the problem of layer name loss when the SuperMap component processes CAD files. Layer names carry key information in architectural design, such as component type, location, etc. Losing layer names will make it difficult to accurately identify and utilize this information in subsequent modeling, increasing the difficulty and error rate of modeling. At the same time, the SuperMap component cannot obtain general information through layer names, which makes a large amount of valuable information contained in the CAD layer wasted and unable to fully play a role in the modeling process, thereby affecting the accuracy and efficiency of three-dimensional model generation. In addition, existing methods also have shortcomings in achieving automation, high-precision modeling and online display, and it is difficult to meet the growing digitalization needs of the construction industry.

[0004] The patent name is: A method for visualizing IFC2x3 data of tunnel main structure steel bars based on BIM, application number: 202110600044.8 By establishing a mapping between different data standards, data compatibility and visual display are achieved, which has certain advantages in data processing and visualization, but this method is mainly aimed at the steel bar data of the tunnel main structure, and its versatility is insufficient. It cannot directly solve the problem of layer name loss and information acquisition when the SuperMap component processes CAD files, or the patent name is: Three-dimensional model generation method, device, equipment and storage medium, application number: 202210730189.4 The drawing identification and information parsing methods are determined according to the component type, which improves the accuracy of modeling, but lacks specificity in dealing with layer-related issues when combining SuperMap components with CAD files, and cannot ensure the integrity of layer names and effective acquisition of information.

[0005] Although these existing methods have achieved certain results in their respective application scenarios, they have obvious defects in solving the problem of SuperMap component processing CAD files. Taking all factors into consideration, the 3D building model generation method based on CAD layers proposed in the present invention has unique advantages. This method can effectively solve the problems of layer name loss and inability to obtain general information when the SuperMap component processes CAD files through dataset name resolution and object generation, line conversion to surface, and surface dataset replication. At the same time, the 3D visualization function of the SuperMap component is used for scene generation and cache processing, realizing automatic and high-precision 3D building model generation, and having online display function, thereby improving the efficiency and quality of building information management and display, and providing a more effective solution for the digital development of the construction industry. Summary of the Invention

[0006] The present invention provides a method for generating a three-dimensional building model based on CAD layers to solve the problems of layer name loss and inability to obtain general information through layer names when the SuperMap component processes CAD files. It realizes automatic and high-precision generation of three-dimensional building models and enables online display, thereby improving the efficiency and quality of building information management and display.

[0007] In order to solve the above problems, the technical solution adopted by the invention is:

[0008] The method for generating a three-dimensional building model based on a CAD layer includes the following steps:

[0009] S01 file import: users import CAD files with suffixes of .dwg and .dxf;

[0010] S02 Layer Conversion and Filtering: Convert the imported CAD layers into data sets and exclude data sets that do not need to be parsed according to pre-set filtering rules;

[0011] S03 layer validity judgment: Check the converted data set. If there is no valid data set, feedback the modeling failure information to the user; if there is a valid data set, proceed to the next step;

[0012] S04 Dataset Name Parsing and Object Generation: Use parsing algorithms to extract the complete name in the valid dataset, and generate the corresponding building component object based on the name through the program's built-in method;

[0013] S05 Line to Surface: Converts the line objects in the dataset into a surface dataset, and writes user-defined fields into the surface dataset in combination with the generated building component objects to improve the attribute information of the surface dataset. S06 Surface Dataset Copy: Copy the surface dataset according to the number of building component objects. During the copying process, the values of some user-defined fields also change dynamically.

[0014] S07 Field Assignment: Get the value of the corresponding custom field from the generated building component object and assign it. For some dynamically changing values, calculate and assign them based on specific fields;

[0015] S08 Dataset Merging: Merge the datasets corresponding to the same building components;

[0016] S09 Special Component Re-assignment: For special data sets including but not limited to building floors, balconies, rooftop rooms, and parapets, modify the attribute values of their corresponding floor elevations in accordance with architectural design specifications and actual needs;

[0017] S10: Clean up useless data sets: clean up data sets generated during the modeling process that do not contribute to the final model;

[0018] S11 scene generation and cache processing: using the 3D visualization function of the SuperMap component to generate a 3D scene containing a building model;

[0019] The principles and advantages of this solution are:

[0020] After importing the CAD file, the layer is converted into a dataset and filtered to remove redundant information, reducing the burden of subsequent processing. Next, the dataset name is parsed, and the building component information contained in the name is used to generate the corresponding building component objects through the program's built-in methods to ensure the accuracy of model construction. When constructing a surface dataset, line objects are converted into surfaces and written into custom fields in combination with component objects. The surface dataset attributes are enriched, detailed information is added to the model, and special components are reassigned to ensure that the model meets industry standards based on building specifications and actual needs. Finally, the SuperMap component is used to generate a 3D scene, completing the conversion from 2D CAD layer information to a 3D building model.

[0021] Compared with the existing technology, in solving the problem of SuperMap component processing CAD files, the existing technology has difficulty in ensuring the integrity of layer names and the effectiveness of information acquisition. The present invention extracts information through data set name resolution, avoiding the trouble caused by the loss of layer names, making the generation of building component objects more accurate and reliable. In terms of the degree of automation, the present invention realizes the automation of the entire process from file import to model generation, greatly reducing manual participation, not only improving modeling efficiency, but also reducing the risk of errors that may be caused by manual operation. In terms of model accuracy, the present invention pays attention to the accuracy and integrity of data in each step, from line to surface conversion to improve attributes, to re-assignment of special components, improving model accuracy, and the generated three-dimensional model can more accurately reflect the actual situation of the building. In terms of display and management efficiency, the SuperMap component is used to generate three-dimensional scenes and perform cache processing, realizing the online display function, which is convenient for all parties involved in the project to view and communicate at any time, greatly improving the efficiency and quality of building information management and display, and promoting the coordinated advancement of construction projects.

[0022] In the above-mentioned S04, the parsing algorithm includes extracting and identifying key information of building components in the dataset name, where the key information includes but is not limited to the building name, building nature and number of floors.

[0023] Furthermore, in S05, the geometric transformation algorithm adopted is a transformation algorithm based on topological relationship and coordinate transformation.

[0024] Furthermore, in S09, the calculation basis of the building base related attribute value includes the attribute information of the building component object and the standard parameters in the building design code.

[0025] Furthermore, in S02, the specific method of filtering rules is:

[0026] Extract the names of all layers in the imported CAD file, create a list of common non-modeling related keywords, and compare each layer name with the keyword list. If the layer name contains any keyword in the list, the dataset corresponding to the layer is determined to be a dataset that does not need to be parsed and is excluded. The corresponding dataset will be filtered out.

[0027] For the remaining data sets that have not been excluded, their data types are detected and only specific data types are allowed to participate in subsequent modeling. If the data type of the data set is not within the allowed range, it will be excluded.

[0028] For the retained datasets of geometric data types, their geometric features are analyzed, the number and distribution density of geometric objects in the dataset are calculated, and a lower limit value for the number of geometric objects and a reasonable range for the distribution density are set. If the number of geometric objects in the dataset is too small, below the lower limit, or the distribution density of geometric objects is too high or too low, then the dataset is judged to have little contribution to modeling and is excluded.

[0029] Convert the obtained dataset names to lowercase, remove special characters from the names, and retain only letters, numbers, and necessary connectors;

[0030] Establish a keyword library of commonly used terms in the architectural field, including building types, building parts, floor information, and building functions. Use natural language processing to segment the preprocessed names. Extract keywords that match the keyword library from the segmentation results. Based on the extracted keywords, use semantic analysis technology to determine the logical relationships between them and construct a semantic network model, with nodes representing keywords and edges representing the relationships between keywords.

[0031] According to the results of semantic analysis, the properties of the building component object are inferred, the building component object is generated through the program's built-in methods, and these properties are assigned to the object as user-defined fields;

[0032] After the building component object is generated, its attributes are verified by comparing it with the existing building knowledge base or actual building design specifications to check whether the attribute value is abnormal. If anomalies are found, they are filtered out.

[0033] Furthermore, when the S01 user imports a CAD file with the suffix .dwg and .dxf, the system uses a custom program based on file flow monitoring and feature recognition to extract the original name information of each layer in the CAD file in advance before the SuperMap component reads the file, and temporarily stores it in an independent cache space. The cache space is stored in a hash table structure, with the unique identifier of the layer in the CAD file as the key value to ensure the accurate correspondence between the name information and the layer.

[0034] Furthermore, in said S04, the steps of the program built-in method are:

[0035] Establish a keyword library covering a wide range of terms in the construction field, including but not limited to building types, building parts, floor markings, and building materials;

[0036] Using the semantic segmentation algorithm in natural language processing technology, the obtained layer or dataset full name is processed. Then, these semantic units are matched with the keyword library to identify the building information category corresponding to each part.

[0037] The fuzzy matching algorithm is introduced to identify the error and correct it to ensure the accuracy of information extraction;

[0038] The program predefines class templates for various building components. Based on the identified building component type, the corresponding class template is called to create a building component object.

[0039] The extracted information is directly assigned to the corresponding basic attributes of the building component object. For some attributes that need to be calculated, they are calculated and assigned based on the existing attribute values and preset formulas;

[0040] Establish spatial relationships and hierarchical associations between building component objects;

[0041] Verify the properties of the building component objects with set properties. If it is found that the property value does not meet the requirements, make corrections or issue a warning.

[0042] Furthermore, in S04, the specific steps of the parsing algorithm are:

[0043] Before name resolution, pre-process the dataset names and unify their formats;

[0044] Build a keyword library containing common terms in the architectural field, use the word segmentation technology in natural language processing to segment the preprocessed names, compare the segmentation results with the keyword library, and extract the keywords.

[0045] Construct a semantic network model, with keywords as nodes and edges between nodes representing the relationships between keywords. For some implicit information, semantic reasoning is used to supplement it and clarify the role and relationship of each keyword in the architectural model.

[0046] Based on the relationships and supplementary information obtained from semantic analysis, the complete name of the dataset is reconstructed. The integrity of the reconstructed name is verified to check whether it contains the necessary information. If key information is missing, it is supplemented according to the semantic analysis results or default rules. If the building name is not clear in the name, it can be supplemented according to the project name or other relevant information. If it cannot be supplemented, it is marked as a possible problem with the dataset name and requires manual intervention to supplement it.

[0047] Furthermore, in the preprocessing stage, an abbreviation library is established for matching and expansion, and abbreviations are expanded into complete terms. For dataset names that do not conform to common naming patterns, they are identified and processed through machine learning algorithms. A classification model is trained using the labeled dataset names to identify irregularly named names, and they are parsed and reconstructed based on similar naming patterns or semantic features. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 Flowchart of the present invention. DETAILED DESCRIPTION

[0049] Example 1 is basically as shown in the attached Figure 1 As shown, the method for generating a three-dimensional building model based on a CAD layer includes the following steps:

[0050] S01 file import: users import CAD files with suffixes of .dwg and .dxf;

[0051] S02 Layer Conversion and Filtering: Convert the imported CAD layers into data sets and exclude data sets that do not need to be parsed according to pre-set filtering rules;

[0052] S03 layer validity judgment: Check the converted data set. If there is no valid data set, feedback the modeling failure information to the user; if there is a valid data set, proceed to the next step;

[0053] S04 Dataset Name Parsing and Object Generation: Use parsing algorithms to extract the complete name in the valid dataset, and generate the corresponding building component object based on the name through the program's built-in method;

[0054] S05 Line to Surface: Converts the line objects in the dataset into a surface dataset, and writes user-defined fields into the surface dataset in combination with the generated building component objects to improve the attribute information of the surface dataset. S06 Surface Dataset Copy: Copy the surface dataset according to the number of building component objects. During the copying process, the values of some user-defined fields also change dynamically.

[0055] S07 Field Assignment: Get the value of the corresponding custom field from the generated building component object and assign it. For some dynamically changing values, calculate and assign them based on specific fields;

[0056] S08 Dataset Merging: Merge the datasets corresponding to the same building components;

[0057] S09 Special Component Re-assignment: For special data sets including but not limited to building floors, balconies, rooftop rooms, and parapets, modify the attribute values of their corresponding floor elevations in accordance with architectural design specifications and actual needs;

[0058] S10: Clean up useless data sets: clean up data sets generated during the modeling process that do not contribute to the final model;

[0059] S11 scene generation and cache processing: Use the 3D visualization function of the SuperMap component to generate a 3D scene containing a building model.

[0060] After importing the CAD file, the layer is converted into a dataset and filtered to remove redundant information, reducing the burden of subsequent processing. Next, the dataset name is parsed, and the building component information contained in the name is used to generate the corresponding building component objects through the program's built-in methods to ensure the accuracy of model construction. When constructing a surface dataset, line objects are converted into surfaces and written into custom fields in combination with component objects. The surface dataset attributes are enriched, detailed information is added to the model, and special components are reassigned to ensure that the model meets industry standards based on building specifications and actual needs. Finally, the SuperMap component is used to generate a 3D scene, completing the conversion from 2D CAD layer information to a 3D building model.

[0061] In solving the problem of SuperMap component processing CAD files, the existing technology is difficult to ensure the integrity of layer names and the effectiveness of information acquisition. The present invention extracts information through data set name resolution, avoiding the trouble caused by the loss of layer names, making the generation of building component objects more accurate and reliable. In terms of automation, the present invention realizes the automation of the entire process from file import to model generation, greatly reducing manual participation, not only improving modeling efficiency, but also reducing the risk of errors that may be caused by manual operation. In terms of model accuracy, the present invention pays attention to the accuracy and integrity of data in each step, from line to surface conversion to improve attributes, to re-assignment of special components, improving model accuracy, and the generated three-dimensional model can more accurately reflect the actual situation of the building. In terms of display and management efficiency, the SuperMap component is used to generate three-dimensional scenes and perform cache processing, realizing the online display function, which is convenient for all parties involved in the project to view and communicate at any time, greatly improving the efficiency and quality of building information management and display, and promoting the coordinated advancement of construction projects.

[0062] When users import CAD files with .dwg and .dxf suffixes into the system, the system will convert the imported CAD layers into datasets and exclude those datasets that do not need to be parsed based on pre-set filtering rules. The system will also judge the validity of the converted datasets. If no valid dataset is found, the system will provide feedback to the user indicating modeling failure. If a valid dataset is found, the system can continue with subsequent processing steps.

[0063] When there is a valid dataset, the parsing algorithm is used to obtain the full name of the dataset. Then, the building component objects are generated through the built-in method of the program. The attributes of these objects are user-defined fields. The area objects in the line dataset are converted into a surface dataset, and the user-defined fields in the building component objects are written into the surface dataset.

[0064] The surface dataset is copied according to the number of floors recorded in the building component object. During the copying process, some user-defined field values will change dynamically according to the actual situation.

[0065] Get the value of the corresponding custom field from the generated building component object and assign it. For those values that change dynamically, calculate and assign them based on specific fields.

[0066] The datasets corresponding to the same building components are merged. Then, for the datasets of special building components, such as building floors, balconies, rooftops, parapets, etc., the building base-related attribute values are calculated, and the attribute values such as the floor elevation to which they belong are modified according to the building design specifications and actual needs.

[0067] After completing all the above processing steps, clear the data sets generated during the modeling process that do not contribute to the final model, use the 3D visualization function of the SuperMap component to generate a 3D scene containing the building model, and finally, publish the service and generate a link that can be accessed by users, so that users can easily view the generated 3D building model online.

[0068] Users only need to import CAD files, and the subsequent operations such as layer conversion, model construction, scene generation and service release are automatically completed by the system, which greatly reduces manual intervention, improves modeling efficiency and reduces the probability of human errors.

[0069] High modeling accuracy: Modeling data is derived from information contained in CAD layer names and remains unmodified during the modeling process, ensuring data authenticity and accuracy. Through precise dataset name parsing and object generation, as well as strict adherence to architectural design specifications for special component handling, the resulting 3D building model accurately reflects the building's physical presence.

[0070] Through parsing algorithms and built-in methods of the program, the layer name information was successfully obtained and utilized, and the generation of three-dimensional building models based on CAD layers was realized, which made up for the shortcomings of the SuperMap component.

[0071] The generated three-dimensional building model can be viewed online through a network link, facilitating information sharing and communication between different users. It is especially suitable for fields that require multi-party collaboration, such as urban planning and architectural design.

[0072] In the S04, the parsing algorithm includes extracting and identifying key information of building components in the dataset name, where the key information includes but is not limited to the building name, building nature and number of floors. By accurately extracting information such as the building name, building nature and number of floors, the properties of the building components can be precisely defined, and key information can be quickly identified. The dataset can be classified and screened based on this information. When the amount of data is huge, the datasets of the same building are grouped together by building name, the data of buildings with different functions are distinguished by building nature, and the data of different floors are sorted by the number of floors. In this way, when performing subsequent operations such as line-to-surface conversion, dataset copying, and field assignment, relevant data can be quickly located and processed, avoiding invalid processing of a large amount of irrelevant data, greatly improving modeling efficiency, and reducing modeling time costs.

[0073] In S05, the geometric transformation algorithm adopted is a transformation algorithm based on topological relationships and coordinate transformation. The transformation is based on topological relationships and can accurately identify the connection relationship and relative position relationship between each line in the line data set. At the same time, the algorithm based on topological relationships and coordinate transformation has strong adaptability and can handle various complex line data situations.

[0074] In S09, the calculation basis of the building base-related attribute values includes the attribute information of the building component objects and the standard parameters in the building design specifications. The attribute information of the building component objects includes the specific characteristics of the building, such as the shape, size, and weight distribution of each part of the building. Combining this information with the standard parameters in the specifications can comprehensively consider various factors affecting the building base, thereby more accurately calculating the base-related attribute values.

[0075] In said S02, the specific method of filtering rules is:

[0076] Extract the names of all layers in the imported CAD file, create a list of common non-modeling related keywords, and compare each layer name with the keyword list. If the layer name contains any keyword in the list, the dataset corresponding to the layer is determined to be a dataset that does not need to be parsed and is excluded. The corresponding dataset will be filtered out.

[0077] For the remaining data sets that have not been excluded, their data types are detected and only specific data types are allowed to participate in subsequent modeling. If the data type of the data set is not within the allowed range, it will be excluded.

[0078] For the retained datasets of geometric data types, their geometric features are analyzed, the number and distribution density of geometric objects in the dataset are calculated, and a lower limit value for the number of geometric objects and a reasonable range for the distribution density are set. If the number of geometric objects in the dataset is too small, below the lower limit, or the distribution density of geometric objects is too high or too low, then the dataset is judged to have little contribution to modeling and is excluded.

[0079] Convert the obtained dataset names to lowercase, remove special characters from the names, and retain only letters, numbers, and necessary connectors;

[0080] Establish a keyword library of commonly used terms in the architectural field, including building types, building parts, floor information, and building functions. Use natural language processing to segment the preprocessed names. Extract keywords that match the keyword library from the segmentation results. Based on the extracted keywords, use semantic analysis technology to determine the logical relationships between them and construct a semantic network model, with nodes representing keywords and edges representing the relationships between keywords.

[0081] According to the results of semantic analysis, the properties of the building component object are inferred, the building component object is generated through the program's built-in methods, and these properties are assigned to the object as user-defined fields;

[0082] After generating the building component object, its properties are verified. By comparing it with the existing building knowledge base or actual building design specifications, the property values are checked for anomalies. If anomalies occur, they are filtered out. By establishing a list of non-modeling related keywords and comparing layer names, layers not related to modeling, such as labeling and annotation layers, can be quickly filtered out to reduce the amount of data processing. The data type is detected, and only specific types of data are retained to avoid processing useless data types and save computing resources. Geometric features are analyzed to exclude data sets with too few geometric objects or unreasonable distribution density to prevent invalid data from interfering with the modeling process. Subsequent modeling operations can focus on processing valid data, greatly improving overall modeling efficiency. The properties of the building component objects are verified, and abnormal properties are checked according to the building knowledge base or design specifications to ensure that the model properties meet the actual building requirements and avoid model deviations caused by incorrect or unreasonable properties. The generated three-dimensional building model is accurate and reliable, meets industry standards and actual application needs, and improves the practicality and credibility of the model.

[0083] The S01 user imports a CAD file with the suffix .dwg and .dxf. During the import, the system uses a custom program based on file flow monitoring and feature recognition to extract the original name information of each layer in the CAD file in advance before the SuperMap component reads the file, and temporarily stores it in an independent cache space. The cache space adopts a hash table structure for storage, using the unique identifier of the layer in the CAD file as the key value to ensure the accurate correspondence between the name information and the layer. Through file flow monitoring and feature recognition technology, the name of each layer can be accurately located and extracted and temporarily stored in an independent cache space. The hash table structure is used to store the unique identifier of the layer in the CAD file as the key value, ensuring the accurate correspondence between the name information and the layer. In the subsequent modeling process, even if the SuperMap component loses the layer name, it can still be obtained from the cache, ensuring that subsequent operations based on the layer name (such as dataset name resolution, building component object generation, etc.) are carried out smoothly, effectively solving the inherent defects of the SuperMap component.

[0084] In said S04, the steps of the program built-in method are:

[0085] Establish a keyword library covering a wide range of terms in the construction field, including but not limited to building types, building parts, floor markings, and building materials;

[0086] Using the semantic segmentation algorithm in natural language processing technology, the obtained layer or dataset full name is processed. Then, these semantic units are matched with the keyword library to identify the building information category corresponding to each part.

[0087] The fuzzy matching algorithm is introduced to identify the error and correct it to ensure the accuracy of information extraction;

[0088] The program predefines class templates for various building components. Based on the identified building component type, the corresponding class template is called to create a building component object.

[0089] The extracted information is directly assigned to the corresponding basic attributes of the building component object. For some attributes that need to be calculated, they are calculated and assigned based on the existing attribute values and preset formulas;

[0090] Establish spatial relationships and hierarchical associations between building component objects;

[0091] Attribute verification is performed on building component objects with assigned properties. If attribute values do not meet requirements, corrections are made or warnings are issued. Establishing a keyword library encompassing a wide range of architectural terminology ensures comprehensive coverage and accurate identification of various building-related information, avoiding omissions or errors caused by missing keywords. Using semantic segmentation and fuzzy matching algorithms from natural language processing technologies, building information can be more accurately extracted from the complete names of layers or datasets. Semantic segmentation algorithms can meticulously analyze text structure, while fuzzy matching algorithms can correct potential recognition errors, collectively improving the accuracy and reliability of information extraction.

[0092] In S04, the specific steps of the parsing algorithm are:

[0093] Before name resolution, pre-process the dataset names and unify their formats;

[0094] Build a keyword library containing common terms in the architectural field, use the word segmentation technology in natural language processing to segment the preprocessed names, compare the segmentation results with the keyword library, and extract the keywords.

[0095] Construct a semantic network model, with keywords as nodes and edges between nodes representing the relationships between keywords. For some implicit information, semantic reasoning is used to supplement it and clarify the role and relationship of each keyword in the architectural model.

[0096] Based on the relationships and supplementary information obtained from semantic analysis, the complete name of the dataset is reconstructed, and the integrity of the reconstructed name is verified to check whether it contains necessary information. If key information is missing, it is supplemented according to the semantic analysis results or default rules. If the building name is not clear in the name, it can be supplemented according to the project name or other relevant information; if it cannot be supplemented, the dataset name is marked as having problems and requires manual intervention to supplement it; the dataset name is preprocessed and formatted in a unified manner, providing a standardized data basis for subsequent analysis, reducing processing difficulties and errors caused by inconsistent formats, and improving processing efficiency and accuracy. A semantic network model is constructed to clarify the relationship between keywords, which helps to deeply understand the architectural semantics contained in the dataset name, and to explore potential structural and logical information, thereby more accurately reconstructing the dataset name.

[0097] In the preprocessing stage, an abbreviation library is established for matching and expansion, and abbreviations are expanded into complete terms. For dataset names that do not conform to common naming patterns, they are identified and processed through machine learning algorithms. A classification model is trained using the labeled dataset names to identify irregularly named names, and they are parsed and reconstructed based on similar naming patterns or semantic features. An abbreviation library is established for matching and expansion, which can convert common abbreviations into complete terms, avoiding ambiguity or information loss caused by abbreviations, making dataset names clearer and complete, and facilitating subsequent analysis and processing.

[0098] Machine learning algorithms are used to identify and process dataset names that do not conform to common naming patterns, greatly improving the system's adaptability to diverse and irregular naming. It is no longer limited to common, standardized naming methods and can handle more complex and special cases.

[0099] Using labeled dataset names to train a classification model can fully leverage historical data and prior knowledge to improve the accuracy and efficiency of identifying irregular names. The model can learn the characteristics and patterns of different naming patterns, allowing for more accurate identification of unusual names.

[0100] Parsing and reconstruction based on similar naming patterns or semantic features can, to a certain extent, infer and recover information that may be missing or ambiguous in irregular naming, so that these names can be effectively understood and utilized rather than being directly excluded or mishandled.

[0101] The above are only embodiments of the present invention. Common knowledge such as the known specific structures and characteristics in the scheme are not described in detail here. Ordinary technicians in the relevant field are aware of all the common technical knowledge in the technical field of the invention before the application date or priority date, can obtain all the existing technologies in the field, and have the ability to apply conventional experimental means before that date. Ordinary technicians in the relevant field can improve and implement this scheme in combination with their own abilities under the inspiration given by this application. Some typical known structures or known methods should not become obstacles for ordinary technicians in the relevant field to implement this application. It should be pointed out that for those skilled in the art, several variations and improvements can be made without departing from the structure of the present invention. These should also be regarded as the scope of protection of the present invention. These will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A method for generating a three-dimensional building model based on a CAD layer, comprising the following steps: S01 file import: users import CAD files with suffixes of .dwg and .dxf; S02 Layer Conversion and Filtering: Convert the imported CAD layers into data sets and exclude data sets that do not need to be parsed according to pre-set filtering rules; S03 layer validity judgment: detect the converted data set. If there is no valid data set, feedback modeling failure information to the user; If there is a valid data set, proceed to the next step; S04 Dataset Name Parsing and Object Generation: Use parsing algorithms to extract the complete name in the valid dataset, and generate the corresponding building component object based on the name through the program's built-in method; S05 Line to Surface: Converts the line objects in the dataset into a surface dataset, and writes user-defined fields into the surface dataset in combination with the generated building component objects to improve the attribute information of the surface dataset. S06 Surface Dataset Copy: Copy the surface dataset according to the number of building component objects. During the copying process, the values of some user-defined fields also change dynamically. S07 Field Assignment: Get the value of the corresponding custom field from the generated building component object and assign it. For some dynamically changing values, calculate and assign them based on specific fields; S08 Dataset Merging: Merge the datasets corresponding to the same building components; S09 Special Component Re-assignment: For special data sets including but not limited to building floors, balconies, rooftop rooms, and parapets, modify the attribute values of their corresponding floor elevations in accordance with architectural design specifications and actual needs; S10: Clean up useless data sets: clean up data sets generated during the modeling process that do not contribute to the final model; S11 scene generation and cache processing: Use the 3D visualization function of the SuperMap component to generate a 3D scene containing a building model.

2. The method for generating a three-dimensional building model based on a CAD layer according to claim 1, wherein: In S04, the parsing algorithm includes extracting and identifying key information of building components in the dataset name, where the key information includes but is not limited to the building name, building nature, and number of floors.

3. The method for generating a three-dimensional building model based on a CAD layer according to claim 1, wherein: In the step S05 , the geometric transformation algorithm used is a transformation algorithm based on topological relationships and coordinate transformation.

4. The method for generating a three-dimensional building model based on a CAD layer according to claim 1, wherein: In the above-mentioned S06, the dynamically changing user-defined field value is adjusted accordingly according to the increasing or decreasing number of floors.

5. The method for generating a three-dimensional building model based on a CAD layer according to claim 1, wherein: In the above-mentioned S09, the calculation basis of the building base related attribute value includes the attribute information of the building component object and the standard parameters in the building design code.

6. The method for generating a three-dimensional building model based on a CAD layer according to claim 1, wherein: In said S02, the specific method of filtering rules is: Extract the names of all layers in the imported CAD file, create a list of common non-modeling related keywords, and compare each layer name with the keyword list. If the layer name contains any keyword in the list, the dataset corresponding to the layer is determined to be a dataset that does not need to be parsed and is excluded. The corresponding dataset will be filtered out. For the remaining data sets that have not been excluded, their data types are detected and only specific data types are allowed to participate in subsequent modeling. If the data type of the data set is not within the allowed range, it will be excluded. For the retained datasets of geometric data types, their geometric features are analyzed, the number and distribution density of geometric objects in the dataset are calculated, and a lower limit value for the number of geometric objects and a reasonable range for the distribution density are set. If the number of geometric objects in the dataset is too small, below the lower limit, or the distribution density of geometric objects is too high or too low, then the dataset is judged to have little contribution to modeling and is excluded. Convert the obtained dataset names to lowercase, remove special characters from the names, and retain only letters, numbers, and necessary connectors; Establish a keyword library of commonly used terms in the architectural field, including building types, building parts, floor information, and building functions. Use natural language processing to segment the preprocessed names. Extract keywords that match the keyword library from the segmentation results. Based on the extracted keywords, use semantic analysis technology to determine the logical relationships between them and construct a semantic network model, with nodes representing keywords and edges representing the relationships between keywords. According to the results of semantic analysis, the properties of the building component object are inferred, the building component object is generated through the program's built-in methods, and these properties are assigned to the object as user-defined fields; After the building component object is generated, its attributes are verified by comparing it with the existing building knowledge base or actual building design specifications to check whether the attribute value is abnormal. If anomalies are found, they are filtered out.

7. The method for generating a three-dimensional building model based on a CAD layer according to claim 1, wherein: The S01 user imports CAD files with suffixes of .dwg and .dxf. During the import, the system uses a custom program based on file flow monitoring and feature recognition to extract the original name information of each layer in the CAD file in advance before the SuperMap component reads the file, and temporarily stores it in an independent cache space. The cache space is stored in a hash table structure, with the unique identifier of the layer in the CAD file as the key value to ensure the accurate correspondence between the name information and the layer.

8. The method for generating a three-dimensional building model based on a CAD layer according to claim 1, wherein: In said S04, the steps of the program built-in method are: Establish a keyword library covering a wide range of terms in the construction field, including but not limited to building types, building parts, floor markings, and building materials; Using the semantic segmentation algorithm in natural language processing technology, the obtained layer or dataset full name is processed. Then, these semantic units are matched with the keyword library to identify the building information category corresponding to each part. The fuzzy matching algorithm is introduced to identify the error and correct it to ensure the accuracy of information extraction; The program predefines class templates for various building components. Based on the identified building component type, the corresponding class template is called to create a building component object. The extracted information is directly assigned to the corresponding basic attributes of the building component object. For some attributes that need to be calculated, they are calculated and assigned based on the existing attribute values and preset formulas; Establish spatial relationships and hierarchical associations between building component objects; Verify the properties of the building component objects with set properties. If it is found that the property value does not meet the requirements, make corrections or issue a warning.

9. The method for generating a three-dimensional building model based on a CAD layer according to claim 1, wherein: In S04, the specific steps of the parsing algorithm are: Before name resolution, pre-process the dataset names and unify their formats; Build a keyword library containing common terms in the architectural field, use the word segmentation technology in natural language processing to segment the preprocessed names, compare the segmentation results with the keyword library, and extract the keywords. Construct a semantic network model, with keywords as nodes and edges between nodes representing the relationships between keywords. For some implicit information, semantic reasoning is used to supplement it and clarify the role and relationship of each keyword in the architectural model. Based on the relationships and supplementary information obtained from the semantic analysis, the complete name of the dataset is reconstructed and the integrity of the reconstructed name is verified to check whether it contains the necessary information. If key information is missing, it is supplemented according to the semantic analysis results or default rules. If the building name is not clear in the name, it can be supplemented according to the project name or other relevant information. If it cannot be supplemented, it is marked that there may be problems with the dataset name and manual intervention is required to supplement it.

10. The method for generating a three-dimensional building model based on a CAD layer according to claim 9, wherein: In the preprocessing stage, an abbreviation library is established for matching and expansion, and abbreviations are expanded into complete terms. For dataset names that do not conform to common naming patterns, they are identified and processed through machine learning algorithms. A classification model is trained using the labeled dataset names to identify irregularly named names, and then parse and reconstruct them based on similar naming patterns or semantic features.

Images