Annotation generation method of multi-party collaborative design bridge model, medium and equipment

Abstract

The invention provides an annotation generation method for a multi-party collaborative design bridge model, a medium and equipment, and the method comprises the steps: obtaining annotation information inputted by an auditor, and the annotation information comprises the opinion content of a part of components in the bridge model; extracting a component number pointed by the opinion content from the annotation information, and determining a target component corresponding to the component number from the bridge model; calling an API (Application Program Interface), and generating an annotation containing corresponding opinion content at a target component in the bridge model; and determining a target designer responsible for the target component based on the component number of the target component, and generating task prompt information to prompt the target designer. According to the scheme, the annotation information of the auditor is automatically analyzed, the component number is identified, the target component is positioned in the model, the API is called to automatically generate the annotation, and the associated designer is determined by using the component number, so that the task prompting function is realized, and the working efficiency of multi-party collaborative design of the bridge model is effectively improved.

Description

Technical Field

[0001] The present application relates to the technical field of bridge model design, and in particular to a method, medium, and device for generating annotations for a multi-party collaborative design of a bridge model. Background Art

[0002] As bridge projects grow larger and more complex, collaborative design models based on BIM (Building Information Modeling) have become mainstream in the industry. During the bridge design process, the design team, auditors, and construction parties must frequently review the models and provide feedback.

[0003] Currently, in most cases, reviewers record issues through verbal communication, paper documents, or independent annotation files (such as PDF annotations), requiring designers to manually assign their comments to specific components in the model. While mainstream BIM software currently supports text annotations (even voice and image annotations), reviewers who use this method to comment rely on manual selection to bind the annotations to model components. This often requires reviewers to spend considerable time locating components, as automatic component location through natural language descriptions is not possible. Furthermore, after the annotations are generated, the responsible person (designer) must be manually notified. This lack of an automated task flow mechanism can easily lead to missed modifications, delays, and difficulties in tracing responsibilities. Summary of the Invention

[0004] The purpose of the embodiments of the present application is to provide a method, medium and equipment for generating annotations for multi-party collaborative design of bridge models, automatically parsing the reviewer's annotation information, identifying component numbers and locating target components in the model, calling APIs to automatically generate annotations, and using component numbers to determine the associated designers, realizing the task prompt function, and effectively improving the work efficiency when multi-party collaborative design of bridge models.

[0005] In order to achieve the above objectives, the embodiments of the present application are implemented in the following manner: In a first aspect, an embodiment of the present application provides a method for generating annotations for a multi-party collaborative design of a bridge model, comprising: obtaining annotation information entered by a reviewer, wherein the annotation information includes opinions on some components in the bridge model; extracting the component number to which the opinion content points from the annotation information, and determining the target component corresponding to the component number from the bridge model; calling an API to generate an annotation containing the corresponding opinion content at the target component in the bridge model; determining the target designer responsible for the target component based on the component number of the target component, and generating task prompt information to prompt the target designer.

[0006] In combination with the first aspect, in a first possible implementation method of the first aspect, the annotation information is text data, and the component number pointed to by the opinion content is extracted from the annotation information, including: extracting the component number using a first extraction scheme; detecting whether the component number extracted by the first extraction scheme is valid, wherein a valid component number indicates that the component number complies with the format specification and a corresponding component exists in the bridge model database; if the component number extracted by the first extraction scheme is valid, determining that the component number is the component number pointed to by the opinion content.

[0007] In combination with the first possible implementation method of the first aspect, in the second possible implementation method of the first aspect, if the component number extracted by the first extraction scheme is invalid, the method also includes: using the second extraction scheme to extract the component number; detecting whether the component number extracted by the second extraction scheme is valid; if the component number extracted by the second extraction scheme is valid, determining that the component number is the component number pointed to by the opinion content; if the component number extracted by the second extraction scheme is invalid, prompting the reviewer to re-enter a valid component number, and re-using the first extraction scheme to extract the component number.

[0008] In combination with the second possible implementation of the first aspect, in a third possible implementation of the first aspect, the first extraction scheme is regular matching, and the second extraction scheme is a bridge domain fine-tuning model based on the Transformer architecture.

[0009] In combination with the first possible implementation method of the first aspect, in the fourth possible implementation method of the first aspect, the format specification of the component number is: "number" + "#" + "component type", "number" + "number" + "component type", "number and letter combination" + "#" + "component type", "number and letter combination" + "number" + "component type", where "number" includes the form of pure numbers and the form of multiple groups of numbers connected by separators, and "number and letter combination" includes the form of direct combination of numbers and letters and the form of combination of numbers and letters through separators.

[0010] In combination with the second possible implementation method of the first aspect, in a fifth possible implementation method of the first aspect, determining the target component corresponding to the component number from the bridge model includes: based on the component number, querying the corresponding target component data in the bridge model database, wherein each component data includes the component number, the belonging model, the BIM software internal ID, and the designer ID; based on the belonging model in the target component data, determining the bridge model where the target component is located; based on the BIM software internal ID, determining the target component object in the bridge model where the target component is located.

[0011] In combination with the fifth possible implementation method of the first aspect, in the sixth possible implementation method of the first aspect, an API is called to generate an annotation containing corresponding opinion content at the target component in the bridge model, including: obtaining the geometric center or bounding box center of the target component object as the annotation anchor point of the target component; calling the BIM software API to create the annotation, and filling the opinion content into the annotation; generating a leader based on the annotation anchor point of the target component and the point where the annotation is located.

[0012] In combination with the fifth possible implementation method of the first aspect, in the seventh possible implementation method of the first aspect, based on the component number of the target component, the target designer responsible for the target component is determined, and task prompt information is generated to prompt the target designer, including: based on the designer ID in the target component data corresponding to the component number, determining the target designer responsible for the target component; based on the component number and reviewer information, generating task prompt information with a timestamp, and sending the task prompt information to the target designer.

[0013] In a second aspect, an embodiment of the present application provides a storage medium, which is set in a device and includes a stored program, wherein when the program is running, the device where the storage medium is located is controlled to execute the annotation generation method for the multi-party collaborative design of a bridge model as described in the first aspect or any one of the possible implementation methods of the first aspect.

[0014] In a third aspect, an embodiment of the present application provides an electronic device comprising a memory and a processor, wherein the memory is used to store information including program instructions, and the processor is used to control the execution of the program instructions, and when the program instructions are loaded and executed by the processor, the steps of the method for generating annotations for a multi-party collaborative design of a bridge model are implemented as described in the first aspect or any one of the possible implementation methods of the first aspect.

[0015] Beneficial effects: This solution obtains the annotation information entered by the reviewer (including opinions on some components in the bridge model), extracts the component number pointed to by the opinion content from the annotation information, and determines the target component corresponding to the component number from the bridge model. It then calls the API to generate an annotation containing the corresponding opinion content at the target component in the bridge model. Based on the component number of the target component, it determines the target designer responsible for the target component and generates task prompt information to prompt the target designer. Based on this, a complete automated process of annotation information parsing - component detection and positioning - automatic annotation generation - task prompt generation can be achieved, which improves the reviewer's annotation efficiency and the communication efficiency of multi-party collaboration, and can realize task flow in a timely manner, reduce modification omissions and delays, facilitate clear responsibilities, and effectively improve the work efficiency of multi-party collaborative design of bridge models.

[0016] Component numbers are extracted through a two-tiered extraction scheme combining regular expression matching and a domain-tuned model. This is combined with fast querying of the component database (traversal query is used when the bridge model database is a single-model file database, while an index database can be constructed for fast querying when the bridge model database is a multi-model file database). Regular expression matching (the first extraction scheme) can quickly process numbers that conform to standard formats (such as "3# bridge pier"), while the Transformer-based bridge domain-tuned model (the second extraction scheme) has stronger parsing capabilities for colloquial descriptions (such as "the third bridge pier"), thereby covering a wider range of input scenarios. Format specifications are defined (with limited extensibility) to support number / letter combinations (such as "P3-1# temporary support") and multi-level delimiters (such as "3-2# beam body"), allowing simple and reliable regular expression matching to cover a wider range of standard input formats, thereby improving component number recognition capabilities.

[0017] By calling the BIM software's native API, annotation objects (such as Revit annotation families and AutoCAD leads) that conform to BIM platform specifications are generated. A unified component database is used to store cross-platform metadata (such as component numbers and BIM software internal IDs), ensuring data consistency across different software and facilitating cross-platform compatibility. Based on component-designer mappings, responsible individuals (designers) are automatically associated. Furthermore, time-stamped task reminders can be pushed via message queues (such as RabbitMQ), triggering notifications through common channels such as email.

[0018] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments of the present application. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without creative work.

[0020] Figure 1 This is a flowchart of a method for generating annotations for a multi-party collaborative bridge model design provided in an embodiment of the present application.

[0021] Figure 2 A schematic diagram with automatically generated annotations. DETAILED DESCRIPTION

[0022] The technical solutions in the embodiments of the present application will be described below in conjunction with the drawings in the embodiments of the present application.

[0023] See also Figure 1 , Figure 1 This is a flow chart of a method for generating annotations for a multi-party collaborative bridge model design, as provided in an embodiment of this application. In this embodiment, the method is applied to electronic devices (computers equipped with a BIM platform, currently supporting mainstream BIM software such as Revit and AutoCAD), and can be run as a script.

[0024] Before introducing this scheme, the format specification of component numbering is defined here as follows: (1) "number" + "#" + "component type", such as 3# bridge pier, 5-1# support; (2) "number" + "number" + "component type", such as 6# beam column, 8-3 truss; (3) "number and letter combination" + "#" + "component type", such as P3# bridge pier, GL-1# support, etc.; (4) "number and letter combination" + "number" + "component type", such as L3# support, GU-4 beam column, etc. "Number" includes the form of pure numbers and the form of multiple groups of numbers connected by separators, "number and letter combination" includes the form of direct combination of numbers and letters and the form of combination of numbers and letters through separators, and "#" is equivalent to "number". The various standard formats defined here are to adapt to different component numbering formats, with the forms of "number" + "#" + "component type" and "number and letter combination" + "#" + "component type" being the most common. Of course, the component numbers displayed in the bridge model drawings are usually "numbers" (including pure numbers and multiple groups of numbers connected by separators), "numbers and letters" (including numbers and letters directly combined and numbers and letters combined with separators), etc., without the need to display "#" or "component type".

[0025] When the reviewer reviews the bridge model designed by multi-party collaboration, the script can be run, and based on this, the electronic device can run the annotation generation method of the bridge model designed by multi-party collaboration.

[0026] First, the electronic device may execute step S10.

[0027] Step S10: Obtain the comment information entered by the reviewer, wherein the comment information includes opinions on some components in the bridge model.

[0028] In this embodiment, when reviewing a bridge model designed by multiple parties, if the reviewer has opinions or questions about a component, they can add comments using the input box provided by the script. The comments added by the reviewer must include the component number and the content of the comments regarding the component. For example, "GL-1# truss: does not meet regulations, the diameter should be increased." After entering the comments, the electronic device can obtain the reviewer's comments and execute step S20 accordingly.

[0029] Step S20: extracting the component number pointed to by the opinion content from the annotation information, and determining the target component corresponding to the component number from the bridge model.

[0030] In this embodiment, the annotation information is text data, and the electronic device can use the first extraction scheme to extract the component number (the component number of each component in the bridge model is unique). Here, the first extraction scheme is regular matching: define the component naming rules in advance (such as "{number}#{component type}", that is, "number" + "#" + "component type"), build a regular expression library, and then perform regular matching on the annotation information. For example, the annotation information entered by the reviewer in the text box is: "The reinforcement ratio of pier 3# is insufficient, please review", and the matching logic code is as follows: import re # Match "number + # + component type" pattern = r"(\d+[\-#]?[\d]*)#(pier|support|beam|pile foundation|……)" text = "3# Pier reinforcement ratio is insufficient, please review" match = re.search(pattern, text) if match: component_id = match.group(1) # Output "3" component_ symbol = match.group(2) # Output "#" component_type = match.group(3) # Output "bridge pier" The first extraction scheme (regular matching) is simple and efficient in extracting component numbers, and is suitable for annotations that clearly comply with naming conventions.

[0031] After the component number is extracted using the first extraction scheme, the electronic device can detect whether the component number extracted by the first extraction scheme is valid, wherein a valid component number indicates that the component number complies with the format specification and a corresponding component exists in the bridge model database.

[0032] In this embodiment, the electronic device can query whether the corresponding component exists in the bridge model database based on the component number. If so, it means that the component number is valid; if not, it means that the component number is invalid.

[0033] If the component number extracted by the first extraction scheme is valid, the electronic device may determine that the component number is the component number pointed to by the opinion content.

[0034] In the case where only the first extraction scheme is used to extract the component number, if the component number extracted by the first extraction scheme is invalid, the reviewer needs to be reminded to re-enter the valid component number and re-use the first extraction scheme to extract the component number.

[0035] To improve the ability to extract component numbers from annotation information, if the component number extracted by the first extraction scheme is invalid, the electronic device can also use a second extraction scheme (a fine-tuned model based on the Transformer architecture for bridges, which requires optimized training for component number recognition in bridge design-related text) to extract the component number. The component number extracted by the second extraction scheme is then tested for validity (see the detection method above). If the component number extracted by the second extraction scheme is valid, it is determined to be the component number specified in the opinion content. If the component number extracted by the second extraction scheme is invalid, the reviewer is prompted to re-enter a valid component number and the first extraction scheme is used again to extract the component number.

[0036] The second extraction scheme is primarily designed to address colloquial descriptions. For example, it uses bridge design documents, component attribute annotations, and historical audit and comment records to perform entity annotation for training the BERT model and fine-tuning the BERT model for the bridge domain. This establishes an entity mapping relationship between colloquial component numbers and standard-compliant component numbers, allowing the extracted colloquial component numbers to be mapped to the correct standard-compliant component numbers. The currently trained bridge domain fine-tuning model in this embodiment has an overall accuracy rate of less than 60% in mapping colloquial component number descriptions to standard-compliant component numbers, which is less than ideal (possibly due to insufficient training data or an inappropriate training method for the bridge domain fine-tuning model, and no better improvement solution has yet been identified). However, it does have some effect. Therefore, the script in this embodiment still primarily relies on the first extraction scheme to extract component numbers. The second extraction scheme plays a supporting role and can be removed. Only the first extraction scheme is used to extract component numbers, requiring auditors to provide standard-compliant component number descriptions when entering annotation information.

[0037] Component numbers are extracted through a two-tiered extraction scheme combining regular expression matching and a domain-tuned model. This is combined with fast querying of the component database (traversal query is used when the bridge model database is a single-model file database, while an index database can be constructed for fast querying when the bridge model database is a multi-model file database). Regular expression matching (the first extraction scheme) can quickly process numbers that conform to standard formats (such as "3# bridge pier"), while the Transformer-based bridge domain-tuned model (the second extraction scheme) has stronger parsing capabilities for colloquial descriptions (such as "the third bridge pier"), thereby covering a wider range of input scenarios. Format specifications are defined (with limited extensibility) to support number / letter combinations (such as "P3-1# temporary support") and multi-level delimiters (such as "3-2# beam body"), allowing simple and reliable regular expression matching to cover a wider range of standard input formats, thereby improving component number recognition capabilities.

[0038] After extracting the valid component number, the electronic device can query the corresponding target component data in the bridge model database based on the component number, where each component data includes the component number, the model to which it belongs, the internal ID of the BIM software, the designer ID, etc. When multiple parties collaborate to design a bridge model, there are two situations for the bridge model database: a database of a single model file and a database of multiple model files. When the bridge model database belongs to a database of a single model file, the target component data can be queried by traversal; when the bridge model database belongs to a database of multiple model files, an index database can be constructed to achieve fast query. The internal ID of the BIM software is to adapt to different BIM software (such as Revit, AutoCAD, etc.) to achieve cross-software operation of the annotation generation plan. The designer ID is to correspond to the designer and realize the automatic flow of tasks.

[0039] Afterwards, the electronic device can determine the bridge model where the target component is located based on the model in the target component data, and then determine the target component object in the bridge model where the target component is located based on the internal ID of the BIM software (such as ElementId of Revit and component_id of AutoCAD).

[0040] After determining the target component object in the bridge model, the electronic device may execute step S30.

[0041] Step S30: calling the API to generate annotations containing corresponding opinion content at the target component in the bridge model.

[0042] In this embodiment, the electronic device may obtain the geometric center or the center of the bounding box of the target component object as the annotation anchor point of the target component.

[0043] For example, when using Revit to achieve the positioning of the target component object, the logic code is as follows: public void LocateComponent(string componentId) { / / 1. Query the database to get element_id var componentData = Database.Query($"SELECT * FROM componentsWHERE component_id = '{componentId}'"); / / 2. Find the Revit element based on element_id Document doc = Application.ActiveUIDocument.Document; ElementId elemId = new ElementId(componentData.ElementId); Element elem = doc.GetElement(elemId); / / 3. Get the geometric center of the component Options geomOptions = new Options(); GeometryElement geomElem = elem.get_Geometry(geomOptions); XYZ centroid = geomElem.GetBoundingBox().Min.Add(geomElem.GetBoundingBox().Max).Divide(2); / / 4. Return coordinates return new double[] { centroid.X, centroid.Y, centroid.Z}; } When using AutoCAD to locate the target component object, the logic code is as follows: def locate_component(component_id): / / 1. Query the database to get the block name / handle component_data = db.query(f"SELECT * FROM components WHEREcomponent_id='{component_id}'") / / 2. Traverse the model space to find the object doc = Application.DocumentManager.MdiActiveDocument db = doc.Database with doc.LockDocument(): with db as transaction: bt = transaction.GetObject(db.BlockTableId, OpenMode.ForRead) model_space = transaction.GetObject(bt[BlockTableRecord.ModelSpace],OpenMode.ForRead) for obj_id in model_space: obj = transaction.GetObject(obj_id, OpenMode.ForRead) / / 3. Check attributes or extend dictionary if hasattr(obj, "ComponentId") and obj.ComponentId ==component_id: / / 4. Get the center of the bounding box ext = obj.GeometricExtents centroid = (ext.MinPoint + ext.MaxPoint) / 2 return (centroid.X, centroid.Y, centroid.Z) After determining the annotation anchor point of the target component, the electronic device can call the BIM software API to create the annotation and fill the comments into the annotation. Here, the logic and scheme of generating annotations are different for different BIM software, because different BIM software supports different schemes for calling APIs. Revit mainly generates annotations by using annotation families, while CAD mainly uses multi-line text (i.e. MText) to generate annotations. After the annotation is generated, the comments can be filled into the annotation, and then based on the annotation anchor point of the target component and the point where the annotation is located, the leader line is generated to automatically generate the annotation. The effect is as follows: Figure 2 shown.

[0044] Of course, the target component can also be visualized, and the area where the target component is located can be highlighted or given a color to facilitate reviewers and designers to view, which is not limited here.

[0045] After the annotation is generated, if the reviewer completes the review, the electronic device may execute step S40.

[0046] Step S40: Based on the component number of the target component, determine the target designer responsible for the target component, and generate task prompt information to prompt the target designer.

[0047] In this embodiment, the electronic device can determine the target designer responsible for the target component based on the designer ID in the target component data corresponding to the component number, and then generate a task reminder message with a timestamp based on the component number and the reviewer information, and send the task reminder message to the target designer (through the contact channel associated with the designer to realize task notification, such as email).

[0048] By calling the BIM software's native API, annotation objects (such as Revit annotation families and AutoCAD leads) that conform to BIM platform specifications are generated. A unified component database is used to store cross-platform metadata (such as component numbers and BIM software internal IDs), ensuring data consistency across different software and facilitating cross-platform compatibility. Based on component-designer mappings, responsible individuals (designers) are automatically associated. Furthermore, time-stamped task reminders can be pushed via message queues (such as RabbitMQ), triggering notifications through common channels such as email.

[0049] It should be noted that, due to the current unsatisfactory recognition accuracy of component numbers expressed in spoken language, the solution of recording annotations by voice has not been further implemented. After overcoming the problem of the recognition accuracy of component numbers expressed in spoken language, the function of recording annotations by voice can be further expanded, and text recognition can be achieved by calling speech recognition-related APIs, such as the highly open speech-to-text API.

[0050] An embodiment of the present application also provides a storage medium, which is set in an electronic device and includes a stored program. When the program is running, the electronic device where the storage medium is located is controlled to execute the annotation generation method for the multi-party collaborative design of a bridge model of this embodiment.

[0051] In summary, the embodiments of the present application provide a method, medium, and device for generating annotations for a multi-party collaborative design of a bridge model. The method obtains the annotation information entered by the reviewer (including the opinions on some components in the bridge model), extracts the component number to which the opinions refer from the annotation information, and determines the target component corresponding to the component number from the bridge model. The API is then called to generate an annotation containing the corresponding opinion content at the target component in the bridge model. Based on the component number of the target component, the target designer responsible for the target component is determined, and task prompt information is generated to prompt the target designer. In this way, a complete automated process of parsing annotation information - component detection and positioning - automatic generation of annotations - generation of task prompts can be achieved, thereby improving the reviewer's annotation efficiency and the communication efficiency of multi-party collaboration, and realizing timely task transfer, reducing modification omissions and delays, and facilitating clear responsibilities, effectively improving work efficiency during multi-party collaborative design of bridge models.

[0052] Component numbers are extracted through a two-tiered extraction scheme combining regular expression matching and a domain-tuned model. This is combined with fast querying of the component database (traversal query is used when the bridge model database is a single-model file database, while an index database can be constructed for fast querying when the bridge model database is a multi-model file database). Regular expression matching (the first extraction scheme) can quickly process numbers that conform to standard formats (such as "3# bridge pier"), while the Transformer-based bridge domain-tuned model (the second extraction scheme) has stronger parsing capabilities for colloquial descriptions (such as "the third bridge pier"), thereby covering a wider range of input scenarios. Format specifications are defined (with limited extensibility) to support number / letter combinations (such as "P3-1# temporary support") and multi-level delimiters (such as "3-2# beam body"), allowing simple and reliable regular expression matching to cover a wider range of standard input formats, thereby improving component number recognition capabilities.

[0053] By calling the BIM software's native API, annotation objects (such as Revit annotation families and AutoCAD leads) that conform to BIM platform specifications are generated. A unified component database is used to store cross-platform metadata (such as component numbers and BIM software internal IDs), ensuring data consistency across different software and facilitating cross-platform compatibility. Based on component-designer mappings, responsible individuals (designers) are automatically associated. Furthermore, time-stamped task reminders can be pushed via message queues (such as RabbitMQ), triggering notifications through common channels such as email.

[0054] In this document, relational terms such as first and second, etc. are used merely to distinguish one entity or operation from another entity or operation, but do not necessarily require or imply any actual relationship or order between these entities or operations.

[0055] The above description is merely an embodiment of the present application and is not intended to limit the scope of protection of the present application. For those skilled in the art, various modifications and variations of the present application are possible. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A method for generating annotations for a multi-party collaborative bridge model design, characterized in that: include: Obtain the comment information entered by the reviewer, wherein the comment information includes opinions on some components in the bridge model; Extract the component number pointed to by the opinion content from the annotation information, and determine the target component corresponding to the component number from the bridge model; Call the API to generate annotations containing corresponding opinion content at the target component in the bridge model; Based on the component number of the target component, a target designer responsible for the target component is determined, and task prompt information is generated to prompt the target designer.

2. The annotation generation method for multi-party collaborative bridge model design according to claim 1 is characterized in that: The annotation information is text data. The component number pointed to by the opinion content is extracted from the annotation information, including: The first extraction scheme is used to extract the component number; Detecting whether the component number extracted by the first extraction scheme is valid, wherein a valid component number indicates that the component number complies with a format specification and a corresponding component exists in the bridge model database; If the component number extracted by the first extraction scheme is valid, the component number is determined to be the component number pointed to by the opinion content.

3. The annotation generation method for multi-party collaborative bridge model design according to claim 2 is characterized in that: If the component number extracted by the first extraction scheme is invalid, the method further includes: The second extraction scheme is used to extract the component number; Check whether the component number extracted by the second extraction scheme is valid; If the component number extracted by the second extraction scheme is valid, determining that the component number is the component number pointed to by the opinion content; If the component number extracted by the second extraction scheme is invalid, the reviewer is prompted to re-enter the valid component number and re-use the first extraction scheme to extract the component number.

4. The annotation generation method for multi-party collaborative bridge model design according to claim 3 is characterized in that: The first extraction scheme is regular matching, and the second extraction scheme is a bridge domain fine-tuning model based on the Transformer architecture.

5. The annotation generation method for multi-party collaborative bridge model design according to claim 2 is characterized in that: The format specifications of component numbers are: "number" + "#" + "component type", "number" + "number" + "component type", "number and letter combination" + "#" + "component type", "number and letter combination" + "number" + "component type", among which "number" includes the form of pure numbers and the form of multiple groups of numbers connected by separators, and "number and letter combination" includes the form of direct combination of numbers and letters and the form of combination of numbers and letters through separators.

6. The annotation generation method for multi-party collaborative bridge model design according to claim 3 is characterized in that: Determine the target component corresponding to the component number from the bridge model, including: Based on the component number, the corresponding target component data is searched in the bridge model database, where each component data includes the component number, the model to which it belongs, the internal ID of the BIM software, and the designer ID; Based on the model in the target component data, the bridge model where the target component is located is determined; Based on the internal ID of the BIM software, the target component object is determined in the bridge model where the target component is located.

7. The method for generating annotations for a multi-party collaborative bridge model design according to claim 6, characterized in that: Call the API to generate annotations containing corresponding opinion content at the target component in the bridge model, including: Obtain the geometric center or bounding box center of the target component object as the annotation anchor point of the target component; Call the BIM software API to create annotations and fill the comments into the annotations; Generates a leader based on the annotation anchor point and annotation location of the target component.

8. The annotation generation method for multi-party collaborative bridge model design according to claim 6 is characterized in that: Based on the component number of the target component, the target designer responsible for the target component is determined, and task prompt information is generated to prompt the target designer, including: Determine a target designer responsible for the target component based on the designer ID in the target component data corresponding to the component number; Based on the component number and the reviewer information, a task reminder message with a time stamp is generated and sent to the target designer.

9. A storage medium, characterized in that: The storage medium is set in the device and includes a stored program, wherein when the program is running, the device where the storage medium is located is controlled to execute the annotation generation method for multi-party collaborative design of a bridge model according to any one of claims 1 to 8.

10. An electronic device comprising a memory and a processor, wherein the memory is used to store information including program instructions, and the processor is used to control the execution of the program instructions, characterized in that: When the program instructions are loaded and executed by the processor, the steps of the method for generating annotations for a multi-party collaborative design bridge model according to any one of claims 1 to 8 are implemented.

Images