BIM model structure tree generation method and device based on AI, electronic equipment and storage medium

Abstract

The invention provides a BIM model structure tree generation method and device based on AI, electronic equipment and a storage medium, and relates to the technical field of road-related computer-aided architectural design, and the method comprises the steps: creating attribute class structure hierarchies; creating a model attribute instance of the BIM model; the method comprises the steps of generating a root node of a BIM model structure tree, storing preset attribute class information in the root node, determining an attribute class of the root node and sub-attribute classes of all levels of the root node, sequentially generating sub-nodes of all levels and correspondingly storing attribute class information of the sub-attribute classes of all levels, the father node of the leaf node is generated, and the attribute class information of the last-stage attribute class is stored; and generating leaf nodes and storing specific model attribute information. By adopting the technical scheme of the invention, the technical problems of model structure tree error generation and model error correspondence in the prior art can be solved.

Description

Technical Field

[0001] This invention relates to the field of computer-aided architectural design technology related to roads, specifically to an AI-based BIM model structure tree generation method, device, electronic device, and storage medium. Background Technology

[0002] With the rapid development of the national economy and the informatization transformation of road engineering design, the design of road traffic signs is a key component of the digital and intelligent design of traffic engineering. The generation of a structure tree based on section management and the 3D model of signs is a key and challenging aspect of BIM applications in traffic engineering. Dynamic generation of structure trees based on section and sign 3D models provides engineers with a convenient and intuitive window for project and model management, improving modeling and management efficiency. Existing structure tree generation methods suffer from the following technical problems: the structure tree cannot be correctly generated when interfering models are present; when multiple models have the same lower-level attributes but different upper-level attributes, the attribute information of these models will be mapped to different structure tree paths, which is inconsistent with the actual situation; due to differences in fields or formats, multiple leaf nodes that are actually identical to the corresponding 3D models in the structure tree are not located under the same structure tree path. Summary of the Invention

[0003] The purpose of this invention is to provide an AI-based BIM model structure tree generation method, device, electronic device, and storage medium to solve the technical problems of incorrect generation of model structure trees and corresponding model errors in the prior art.

[0004] To achieve the above objectives, this invention provides an AI-based BIM model structure tree generation method, comprising: creating an attribute class structure hierarchy, including multiple attribute classes and their attribute class information and hierarchical relationships; creating model attribute instances of the BIM model, including model attribute information corresponding to the attribute classes and their hierarchical relationships; generating a root node of the BIM model structure tree, storing preset attribute class information in the root node, determining the attribute class of the root node and its sub-attribute classes at each level based on the attribute class structure hierarchy, sequentially generating each level of child nodes and storing the attribute class information of each level of sub-attribute classes, until the parent node of the leaf node is generated and the attribute class information of the final level attribute class is stored; obtaining a model attribute instance whose model attribute information matches the attribute class information of the final level attribute class, obtaining specific model attribute information of the matched model attribute instance, generating a leaf node and storing the specific model attribute information.

[0005] This invention also provides an AI-based BIM model structure tree generation device, comprising: an attribute class structure hierarchy creation module for creating an attribute class structure hierarchy, including multiple attribute classes and their attribute class information and hierarchical relationships; a model attribute instance creation module for creating model attribute instances of a BIM model, including model attribute information corresponding to the attribute classes and their hierarchical relationships; a node generation module for generating the root node of the BIM model structure tree, storing preset attribute class information in the root node, determining the attribute class of the root node and its sub-attribute classes at each level based on the attribute class structure hierarchy, sequentially generating sub-nodes at each level and storing the attribute class information of the sub-attribute classes at each level, until the parent node of the leaf node is generated and the attribute class information of the final level attribute class is stored; and a model matching module for obtaining model attribute instances whose model attribute information matches the attribute class information of the final level attribute class, obtaining specific model attribute information of the matched model attribute instances, generating leaf nodes and storing the specific model attribute information.

[0006] The present invention also provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps in the AI-based BIM model structure tree generation method.

[0007] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps in the AI-based BIM model structure tree generation method.

[0008] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0009] The AI-based BIM model structure tree generation method of this invention determines the hierarchical relationship between attribute classes by creating an attribute class structure hierarchy. It achieves accurate matching between the model and the structure tree path by creating model attribute instances in the BIM model that correspond to the attribute structures of the attribute class structure hierarchy. The direct matching of model attribute instances with attribute information in nodes improves the matching efficiency between the BIM model and the model structure tree and avoids the omission of model attribute information. By generating nodes of the model structure tree level by level and correspondingly storing attribute class information in the attribute class structure hierarchy, the method can modify the attribute information stored in any node of the model structure tree by modifying the attribute class information in the attribute class structure hierarchy. This ensures that the attribute information stored in the nodes of the model structure tree is no longer limited to the attribute information contained in the BIM model and also ensures that even with interfering models, the model structure tree and its attribute class information can be correctly generated and stored.

[0010] Preferably, the step of sequentially generating child nodes at each level and correspondingly storing the attribute class information of the sub-attribute classes at each level until the parent node of the leaf node is generated and the attribute class information of the last-level attribute class is stored includes: sequentially generating child nodes at each level and correspondingly storing the attribute class information of the sub-attribute classes at each level until the target-level child node is generated and the corresponding attribute class information is stored; obtaining the attribute class information stored in the target-level child node, filtering model attribute instances containing the attribute class information as target model attribute instances, determining the sub-attribute classes of the attribute classes of the target-level child node, obtaining the model attribute information in the target model attribute instance corresponding to the sub-attribute class, generating the next-level child node of the target-level child node, and storing the attribute class information of the sub-attribute class in the next-level child node; sequentially generating child nodes at each level under the next-level child node and correspondingly storing the attribute class information of the sub-attribute classes at each level until the parent node of the leaf node is generated and the attribute class information of the last-level attribute class is stored.

[0011] By adopting the above technical solution, by activating attribute classes at a specific level, model attribute instances whose model attribute information matches the attribute class at that specific level can be filtered out. BIM models that do not belong to the activated attribute classes do not participate in the generation of the structure tree, thereby realizing flexible matching between the model structure tree and the BIM model and saving computing resources.

[0012] Preferably, the AI-based BIM model structure tree generation method further includes: identifying parent nodes of leaf nodes that do not store the specific model attribute information as invalid nodes; sequentially identifying nodes whose child nodes are all invalid nodes as invalid nodes; and removing all invalid nodes.

[0013] By adopting the above technical solution, all nodes on the local tree structure path that do not have a corresponding BIM model can be removed, making the model tree structure information concise and reducing the interference of invalid information in the tree structure to users.

[0014] Preferably, after creating model attribute instances of a BIM model, the method further includes: constructing an attribute field knowledge graph; training a semantic similarity model based on the semantic mapping relationship defined between attribute fields; calculating the similarity value between the original attribute name field in the model attribute information and the corresponding standard attribute name field in the mapping table using the semantic similarity model; if the similarity exceeds a preset threshold, replacing the original attribute name field with the standard attribute name field; calculating the similarity value between the original format attribute value in the model attribute information and the corresponding standard format attribute value in the mapping table using the semantic similarity model; if the similarity exceeds a preset threshold, extracting the corresponding standard format attribute value field in the mapping table using regular expressions, and replacing the original format attribute value field with the standard format attribute value field; identifying multiple model attribute instances with the same model attribute information, and retaining one of the multiple model attribute instances, wherein the model attribute information includes attribute name and attribute value.

[0015] The above technical solution can solve the technical problem that multiple 3D models have the same semantics for attributes at different levels but different fields or formats due to different versions of modeling software, etc. It can ensure accurate deduplication of BIM model attribute instances and ensure that each model attribute instance matched in the final generated model structure tree is unique.

[0016] Preferably, the node generation module includes: a first child node generation unit, used to sequentially generate child nodes at all levels and store the attribute class information of the sub-attribute classes at each level, until the target level child node is generated and the corresponding attribute class information is stored; a model attribute instance filtering unit, used to obtain the attribute class information stored in the target level child node, filter model attribute instances containing the attribute class information as target model attribute instances, determine the sub-attribute classes of the attribute classes of the target level child node, obtain the model attribute information corresponding to the sub-attribute classes in the target model attribute instances, generate the next level child node of the target level child node, and store the attribute class information of the sub-attribute classes in the next level child node; and a second child node generation unit, used to sequentially generate child nodes at all levels under the next level child node and store the attribute class information of the sub-attribute classes at each level, until the parent node of the leaf node is generated and the attribute class information of the last level attribute class is stored.

[0017] Preferably, the AI-based BIM model structure tree generation device further includes: an invalid parent node identification module, used to identify parent nodes of leaf nodes that do not store the specific model attribute information as invalid nodes; an invalid node identification module, used to identify nodes whose child nodes are all invalid nodes as invalid nodes; and an invalid node removal module, used to remove all invalid nodes.

[0018] Preferably, the model attribute instance creation module includes: a model training unit, used to construct an attribute field knowledge graph and train a semantic similarity model based on the semantic mapping relationship defined between attribute fields; an attribute name standardization unit, used to calculate the similarity value between the original attribute name field in the model attribute information and the corresponding standard attribute name field in the mapping table using the semantic similarity model, and if the similarity exceeds a preset threshold, the original attribute name field is replaced with the standard attribute name field; an attribute value format standardization unit, used to calculate the similarity value between the original format attribute value in the model attribute information and the corresponding standard format attribute value in the mapping table using the semantic similarity model, and if the similarity exceeds a preset threshold, the corresponding standard format attribute value field in the mapping table is extracted using regular expressions, and the original format attribute value field is replaced with the standard format attribute value field; and a model attribute instance deduplication unit, used to identify multiple model attribute instances with the same model attribute information and retain one of the multiple model attribute instances, wherein the model attribute information includes attribute name and attribute value. Attached Figure Description

[0019] To more clearly illustrate the embodiments of the present invention, the accompanying drawings used in the embodiments 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 based on these drawings without creative effort.

[0020] Figure 1 A flowchart of the AI-based BIM model structure tree generation method provided by the present invention;

[0021] Figure 2 This is a schematic diagram of the model structure tree provided by the present invention;

[0022] Figure 3 A schematic diagram of the structure of the AI-based BIM model structure tree generation device provided by the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Figure 1 This is a flowchart illustrating the AI-based BIM model structure tree generation method provided in this embodiment.

[0025] like Figure 1 As shown, this embodiment discloses an AI-based BIM model structure tree generation method, including:

[0026] Step S1: Create an attribute class structure hierarchy, containing multiple attribute classes, their attribute class information, and hierarchical relationships. The attribute class structure hierarchy can be created based on the attribute information bound to all dynamically read 3D models. In some preferred embodiments, to avoid incorrect generation of the model structure tree when there are errors in the model attribute information or when other software generates interfering models, and to achieve intelligent generation of the attribute class structure hierarchy, the creation of the attribute class structure hierarchy specifically includes:

[0027] Step S11: Construct a knowledge graph for the field of traffic engineering and form a traffic engineering classification code corresponding to multi-level node names;

[0028] Step S12: Extract traffic engineering classification hierarchy information from industry standards and historical project information through natural language processing (NLP) to correct the constructed traffic engineering classification code, such as supplementing missing node levels and correcting node names.

[0029] Step S13, constructing an attribute class structure hierarchy based on traffic engineering classification coding, including:

[0030] Step S131: Based on the top-level node name and the names of each level of child nodes in the traffic engineering classification code, generate the root node attribute class and the attribute classes of each level of nodes in the attribute class structure hierarchy until all leaf node attribute classes are generated.

[0031] Step S132 generates child node attribute classes for the leaf node attribute class; specifically, this includes: obtaining the leaf node attribute class name, then marking the leaf node attribute class name, for example, by adding "Node" to the end of the leaf node attribute class name; then inserting child node attribute classes under the leaf node attribute class, using the marked leaf node attribute class name as the name of the child node attribute class. Ordinary attribute classes and relational attribute classes are defined. Each relational attribute class includes a leaf node attribute class and its child node attribute classes, with the leaf node attribute class being the parent node attribute class in the relational attribute class. In the attribute class structure hierarchy, attribute classes other than the leaf node attribute class and its child node attribute classes in the relational attribute class are ordinary attribute classes.

[0032] Step S2: Create model attribute instances for the BIM model, containing model attribute information corresponding to the attribute classes and their hierarchical relationships. Each model attribute instance stores general attribute class information and leaf node attribute class information. General attribute class information includes location information such as sections and routes, as well as other categories of information, such as identification information and dimension information. Leaf node attribute class information is used to associate with the generated model structure tree, and the leaf node attribute class information of each model attribute instance is different from each other.

[0033] In some preferred embodiments, based on the information set of all model attribute instances, it is verified whether the traffic engineering classification code obtained in step S12 can satisfy the description of all models; if not, the missing node levels are supplemented.

[0034] Specifically, to avoid situations where multiple 3D models have the same semantics for attributes at different levels but different fields or formats due to different versions of modeling software, and to ensure accurate deduplication of BIM model attribute instances, ensuring that each model attribute instance matched in the final generated model structure tree is unique, after creating the model attribute instances of the BIM model, the following is also included:

[0035] Step S21: Construct an attribute field knowledge graph and train a semantic similarity model based on the semantic mapping relationship defined between attribute fields.

[0036] Step S22: Calculate the similarity value between the original attribute name field in the model attribute information and the corresponding standard attribute name field in the mapping table using the semantic similarity model. If the similarity exceeds a preset threshold, replace the original attribute name field with the standard attribute name field. When multiple candidate standard attribute name fields exist for the same attribute name, such as "ID" and "number", the field with the highest priority can be selected as the standard attribute name field based on historical usage frequency or user preference.

[0037] Step S23: Using the semantic similarity model, a clustering algorithm is employed to calculate the similarity between the original format attribute values ​​in the model attribute information and the corresponding standard format attribute values ​​in the mapping table. If the similarity exceeds a preset threshold, the corresponding standard format attribute value field in the mapping table is extracted using regular expressions, and the original format attribute value field is replaced with the standard format attribute value field. When multiple candidate standard formats exist for the same attribute value, such as "K1+900" and "K1+900.0", the format with the highest priority can be selected as the standard format based on historical usage frequency or user preference.

[0038] Step S24: Identify multiple model attribute instances with the same model attribute information, and retain one of the multiple model attribute instances, wherein the model attribute information includes attribute name and attribute value.

[0039] Step S3: Generate the root node of the BIM model structure tree, store the preset attribute class information in the root node, determine the attribute class of the root node and its sub-attribute classes at all levels based on the attribute class structure hierarchy, generate each level of child nodes in sequence and store the attribute class information of each level of sub-attribute classes accordingly, until the parent node of the leaf node is generated and the attribute class information of the last level attribute class is stored.

[0040] Specifically, the step of sequentially generating child nodes at each level and correspondingly storing the attribute class information of each level of child attribute class, until the parent node of the leaf node is generated and the attribute class information of the last-level attribute class is stored, includes:

[0041] Step S31: Generate child nodes at each level sequentially and store the attribute class information of each level of sub-attribute class accordingly, until the target level child nodes are generated and the corresponding attribute class information is stored;

[0042] Step S32: Obtain the attribute class information stored in the target level child node, filter the model attribute instances containing the attribute class information as target model attribute instances, determine the sub-attribute classes of the attribute class of the target level child node, obtain the model attribute information corresponding to the sub-attribute class in the target model attribute instance, generate the next level child node of the target level child node, and store the attribute class information of the sub-attribute class in the next level child node;

[0043] Step S33: Generate each level of child nodes under the next level child node in sequence and store the attribute class information of each level of child attribute class accordingly, until the parent node of the leaf node is generated and the attribute class information of the last level attribute class is stored.

[0044] Step S4: Obtain model attribute instances whose model attribute information matches the attribute class information of the last-level attribute class; obtain specific model attribute information of the matched model attribute instances; generate the leaf nodes and store the specific model attribute information.

[0045] like Figure 2As shown, in one specific implementation, firstly, a root node of the BIM model structure tree is generated, which stores the attribute class information "Traffic Engineering and Roadside Facility Components". Then, two child nodes are generated under the root node, storing the section attribute class information "1" and "2" respectively. In this embodiment, the section attribute class is determined to be a sub-attribute class of "Traffic Engineering and Roadside Facility Components" based on the attribute class structure hierarchy. In other implementations, when the attribute class structure hierarchy does not have a section attribute class, section attribute classes or other attribute classes can be obtained from model attribute instances, etc., and added to the sub-attribute class set of "Traffic Engineering and Roadside Facility Components", and the hierarchical relationship between the obtained attribute classes and other attribute classes is defined. Then, model attribute instances are filtered according to actual needs. In this embodiment, only the model structure tree with section information "1" is generated, so the model attribute instance with section attribute class information "1" in the ordinary attribute class information is filtered as the target model attribute instance, and the model attribute instance with section information "2" is no longer involved in the subsequent association with the model structure tree. Then, based on the attribute class structure hierarchy or defined hierarchical relationship, the sub-attribute class of the segment attribute class is determined to be the route attribute class. The model attribute information "K6" and "K7" corresponding to the route attribute class in the target model attribute instance are obtained. Two child nodes are generated under the node with segment attribute class information "1", storing the route attribute class information "K6" and "K7" respectively in these two child nodes. Next, model attribute instances with route attribute class information "K6" from the ordinary attribute class information are selected as target model attribute instances. Model attribute instances with route attribute class information "K7" are no longer involved in subsequent association with the model structure tree. Then, based on the attribute class structure hierarchy or defined hierarchical relationship, the sub-attribute class of the route attribute class is determined to be "traffic signs" in the sign information attribute class, and this sign information attribute class is a normal attribute class. Child nodes are generated under the "K6" node, storing the attribute class information "traffic signs" in them. Then, based on the attribute class structure hierarchy, the sub-attribute class of "traffic sign" is determined to be "single-column traffic sign" in the design information attribute class, and the design information attribute class is a normal attribute class. Child nodes under the "traffic sign" node are generated, and the attribute class information "single-column traffic sign" is stored in them.Then, based on the attribute class structure hierarchy, the sub-attribute classes of "single-column traffic sign" are determined to be "single-column traffic sign at roadbed" and "single-column traffic sign at bridge" in the location information attribute class, and both are relation attribute classes. Two child nodes are generated under "single-column traffic sign", which store the attribute class information "single-column traffic sign at roadbed" and "single-column traffic sign at bridge" respectively. Child nodes of the "single-column traffic sign at roadbed" node and the "single-column traffic sign at bridge" node are generated respectively, and leaf node attribute class information is stored in them respectively. The information is specific attribute information for model attribute instances of "single-column traffic sign at roadbed" and "single-column traffic sign at bridge". Each child node stores specific attribute information for one model attribute instance. For example, the "single-column traffic sign at roadbed" node has two child nodes. Each child node stores the start mileage, end mileage and type information of a model attribute instance: "461+046.668~461+046.668 (general directional sign)" and "461+339.662-461+339.662 (general directional sign)". Then, based on the attribute class structure hierarchy, the sub-attribute classes of "single-column traffic sign at roadbed" and "single-column traffic sign at bridge" are determined to be "single-column traffic sign at roadbed Node" and "single-column traffic sign at bridge Node" respectively. Since the sub-attribute class name contains the marker "Node", the next child node attribute class is traversed, and the above process is repeated to continue to determine whether other untraversed child node attribute classes have sub-attribute classes: if they exist, it is determined whether the sub-attribute class is a normal attribute class or a relational attribute class, and a node storing the corresponding information is generated; if they do not exist, the process is exited.

[0046] Specifically, to ensure that all nodes on local tree structure paths without corresponding BIM models are cleared, making the model tree structure information concise and reducing interference from invalid information in the tree structure, the BIM model tree structure generation method further includes:

[0047] Step S5: Identify leaf nodes that do not store child nodes containing the specific model attribute information as invalid nodes;

[0048] Step S6: Identify nodes whose child nodes are all invalid nodes as invalid nodes;

[0049] Step S7: Remove all invalid nodes.

[0050] For example, in the above implementation, the node "K6" also includes the child node "Traffic Markings," which in turn includes the child node "Lane Edge Line," which in turn includes the child nodes "Lane Edge Line - Hot-Melt Standard Type" and "Lane Edge Line - Hot-Melt Vibration Type." Both "Lane Edge Line - Hot-Melt Standard Type" and "Lane Edge Line - Hot-Melt Vibration Type" are leaf nodes and do not store specific model attribute information. Therefore, the leaf nodes "Lane Edge Line - Hot-Melt Standard Type" and "Lane Edge Line - Hot-Melt Vibration Type" are identified as invalid nodes. At this time, all child nodes of the child node "Lane Edge Line" are also invalid nodes, and the child node "Lane Edge Line" is identified as an invalid node. Similarly, the child node "Traffic Markings" is also identified as an invalid node. Since the child node "Traffic Signs" of the node "K6" is not an invalid node, the child node "K6" is identified as a valid node, and the child nodes "Traffic Markings," "Lane Edge Line," "Lane Edge Line - Hot-Melt Standard Type," and "Lane Edge Line - Hot-Melt Vibration Type" are removed from the tree structure.

[0051] Figure 3 This is a schematic diagram of the AI-based BIM model structure tree generation device provided in this embodiment.

[0052] like Figure 3 As shown, this embodiment provides an AI-based BIM model structure tree generation device, including:

[0053] The attribute class structure hierarchy creation module 101 is used to create an attribute class structure hierarchy, which includes multiple attribute classes, their attribute class information, and hierarchical relationships.

[0054] The model attribute instance creation module 102 is used to create model attribute instances of the BIM model, which contain model attribute information corresponding to the attribute class and its hierarchical relationship.

[0055] The node generation module 103 is used to generate the root node of the BIM model structure tree, store the preset attribute class information in the root node, determine the attribute class of the root node and its sub-attribute classes at all levels based on the attribute class structure hierarchy, generate each level of sub-nodes in sequence and store the attribute class information of each level of sub-attribute classes accordingly, until the parent node of the leaf node is generated and the attribute class information of the last level attribute class is stored.

[0056] The model matching module 104 is used to obtain model attribute instances that match the attribute class information of the final-level attribute class, obtain specific model attribute information of the matched model attribute instances, generate the leaf nodes, and store the specific model attribute information.

[0057] Specifically, the node generation module 103 includes:

[0058] The first child node generation unit is used to sequentially generate child nodes at each level and store the attribute class information of the child attribute class at each level, until the target level child node is generated and the corresponding attribute class information is stored.

[0059] The model attribute instance filtering unit is used to obtain the attribute class information stored in the target level child node, filter the model attribute instances containing the attribute class information as target model attribute instances, determine the sub-attribute classes of the attribute class of the target level child node, obtain the model attribute information corresponding to the sub-attribute class in the target model attribute instance, generate the next level child node of the target level child node, and store the attribute class information of the sub-attribute class in the next level child node.

[0060] The second child node generation unit is used to sequentially generate each level of child nodes under the next level child node and store the attribute class information of each level of child attribute class accordingly, until the parent node of the leaf node is generated and the attribute class information of the last level attribute class is stored.

[0061] Specifically, the BIM model structure tree generation device further includes:

[0062] The invalid leaf node identification module is used to identify leaf nodes that do not have child nodes storing the specific model attribute information as invalid nodes;

[0063] The invalid node identification module is used to identify nodes whose child nodes are all invalid nodes as invalid nodes.

[0064] The invalid node removal module is used to remove all invalid nodes.

[0065] Specifically, the model attribute instance creation module 102 includes:

[0066] The model training unit is used to construct an attribute field knowledge graph and train a semantic similarity model based on the semantic mapping relationship defined between attribute fields.

[0067] The attribute name standardization unit is used to calculate the similarity value between the original attribute name field in the model attribute information and the corresponding standard attribute name field in the mapping table using the semantic similarity model. If the similarity exceeds a preset threshold, the original attribute name field is replaced with the standard attribute name field.

[0068] The attribute value format standardization unit is used to calculate the similarity between the original format attribute value in the model attribute information and the corresponding standard format attribute value in the mapping table using the semantic similarity model. If the similarity exceeds a preset threshold, the corresponding standard format attribute value field in the mapping table is extracted by regular expression, and the original format attribute value field is replaced with the standard format attribute value field.

[0069] The model attribute instance deduplication unit is used to identify multiple model attribute instances with the same model attribute information and retain one of the multiple model attribute instances, wherein the model attribute information includes attribute name and attribute value.

[0070] This embodiment also provides an electronic device that stores a computer program, including: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the BIM model structure tree generation method of the claim.

[0071] This embodiment also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the BIM model structure tree generation method.

[0072] It should be noted that for descriptions of relevant parts of the AI-based BIM model structure tree generation device, electronic device, and computer-readable storage medium disclosed in this embodiment of the invention, please refer to the detailed description of the corresponding parts in the AI-based BIM model structure tree generation method provided in this embodiment of the invention, which will not be repeated here. Furthermore, parts of the technical solutions provided in this embodiment of the invention that are consistent with the implementation principles of corresponding technical solutions in the prior art have not been described in detail to avoid excessive elaboration.

[0073] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0074] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0075] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0076] It should also be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0077] The foregoing description of various embodiments of the invention is provided for the purpose of description to those skilled in the art. It is not intended to be exhaustive or to limit the invention to a single disclosed embodiment. As mentioned above, various alternatives and variations of the invention will be apparent to those skilled in the art to which the foregoing pertains. Therefore, while some alternative embodiments have been specifically discussed, other embodiments will be obvious or readily apparent to those skilled in the art. The invention is intended to include all alternatives, modifications, and variations of the invention already discussed herein, as well as other embodiments falling within the spirit and scope of the foregoing claims.

[0078] Although the invention has been described by way of embodiments, those skilled in the art will recognize that the invention has many variations and modifications without departing from its spirit, and the appended claims cover such variations and modifications without departing from its spirit.

Claims

1. An AI-based method for generating BIM model structure trees, including: Create an attribute class structure hierarchy, containing multiple attribute classes, their attribute class information, and hierarchical relationships; Create BIM model attribute instances, which contain model attribute information corresponding to the attribute class and its hierarchical relationship; Generate the root node of the BIM model structure tree, store the preset attribute class information in the root node, determine the attribute class of the root node and its sub-attribute classes at all levels based on the attribute class structure hierarchy, generate each level of sub-nodes in sequence and store the attribute class information of each level of sub-attribute classes accordingly, until the parent node of the leaf node is generated and the attribute class information of the last level attribute class is stored. Obtain model attribute instances whose model attribute information matches the attribute class information of the last-level attribute class; obtain specific model attribute information of the matched model attribute instances; generate leaf nodes and store the specific model attribute information.

2. The method according to claim 1, characterized in that, The process of sequentially generating child nodes at each level and correspondingly storing the attribute class information of each level of child attribute class, until the parent node of the leaf node is generated and the attribute class information of the last-level attribute class is stored, includes: The generation of child nodes at each level is carried out sequentially, and the attribute class information of each level of child attribute class is stored accordingly, until the generation of the target level child nodes and the storage of the corresponding attribute class information are completed; Obtain the attribute class information stored in the target level child node, filter the model attribute instances that contain the attribute class information as target model attribute instances, determine the sub-attribute classes of the attribute class of the target level child node, obtain the model attribute information corresponding to the sub-attribute class in the target model attribute instance, generate the next level child node of the target level child node, and store the attribute class information of the sub-attribute class in the next level child node; The process continues until the parent node of the leaf node is generated and the attribute class information of the last-level attribute class is stored.

3. The method according to claim 1, characterized in that, Also includes: Identify parent nodes of leaf nodes that do not store the specific model attribute information as invalid nodes; Nodes whose child nodes are all invalid nodes are identified as invalid nodes. Remove all invalid nodes.

4. The method according to claim 1, characterized in that, After creating an instance of the model properties of the BIM model, the following is also included: Construct an attribute field knowledge graph and train a semantic similarity model based on the semantic mapping relationships defined between attribute fields; The semantic similarity model is used to calculate the similarity value between the original attribute name field in the model attribute information and the corresponding standard attribute name field in the mapping table. If the similarity exceeds a preset threshold, the original attribute name field is replaced with the standard attribute name field. The semantic similarity model is used to calculate the similarity between the original format attribute value in the model attribute information and the corresponding standard format attribute value in the mapping table. If the similarity exceeds a preset threshold, the corresponding standard format attribute value field in the mapping table is extracted by regular expression, and the original format attribute value field is replaced with the standard format attribute value field. Identify multiple model attribute instances with the same model attribute information, and retain one of the multiple model attribute instances, wherein the model attribute information includes attribute name and attribute value.

5. An AI-based BIM model structure tree generation device, comprising: The attribute class structure hierarchy creation module is used to create attribute class structure hierarchies, which contain multiple attribute classes, their attribute class information, and hierarchical relationships. The model attribute instance creation module is used to create model attribute instances of the BIM model, which contain model attribute information corresponding to the attribute class and its hierarchical relationship. The node generation module is used to generate the root node of the BIM model structure tree, store the preset attribute class information in the root node, determine the attribute class of the root node and its sub-attribute classes at all levels based on the attribute class structure hierarchy, generate each level of sub-nodes in sequence and store the attribute class information of each level of sub-attribute classes accordingly, until the parent node of the leaf node is generated and the attribute class information of the last level attribute class is stored. The model matching module is used to obtain model attribute instances that match the attribute class information of the final-level attribute class, obtain specific model attribute information of the matched model attribute instances, generate leaf nodes, and store the specific model attribute information.

6. The apparatus according to claim 5, characterized in that, The node generation module includes: The first child node generation unit is used to sequentially generate child nodes at each level and store the attribute class information of the child attribute class at each level, until the target level child node is generated and the corresponding attribute class information is stored. The model attribute instance filtering unit is used to obtain the attribute class information stored in the target level child node, filter the model attribute instances containing the attribute class information as target model attribute instances, determine the sub-attribute classes of the attribute class of the target level child node, obtain the model attribute information corresponding to the sub-attribute class in the target model attribute instance, generate the next level child node of the target level child node, and store the attribute class information of the sub-attribute class in the next level child node. The second child node generation unit is used to sequentially generate each level of child nodes under the next level child node and store the attribute class information of each level of child attribute class accordingly, until the parent node of the leaf node is generated and the attribute class information of the last level attribute class is stored.

7. The apparatus according to claim 5, characterized in that, Also includes: The invalid parent node identification module is used to identify invalid nodes as leaf nodes that do not store the specific model attribute information; The invalid node identification module is used to identify nodes whose child nodes are all invalid nodes as invalid nodes. The invalid node removal module is used to remove all invalid nodes.

8. The apparatus according to claim 5, characterized in that, The model attribute instance creation module includes: a model training unit, used to construct an attribute field knowledge graph and train a semantic similarity model based on the semantic mapping relationship defined between attribute fields; The attribute name standardization unit is used to calculate the similarity value between the original attribute name field in the model attribute information and the corresponding standard attribute name field in the mapping table using the semantic similarity model. If the similarity exceeds a preset threshold, the original attribute name field is replaced with the standard attribute name field. The attribute value format standardization unit is used to calculate the similarity between the original format attribute value in the model attribute information and the corresponding standard format attribute value in the mapping table using the semantic similarity model. If the similarity exceeds a preset threshold, the corresponding standard format attribute value field in the mapping table is extracted by regular expression, and the original format attribute value field is replaced with the standard format attribute value field. The model attribute instance deduplication unit is used to identify multiple model attribute instances with the same model attribute information and retain one of the multiple model attribute instances, wherein the model attribute information includes attribute name and attribute value.

9. An electronic device having a computer program stored thereon, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the method according to any one of claims 1 to 4.

10. A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method according to any one of claims 1 to 4.

Images