Component label automatic generation method and device based on three dimensions, and medium
By selecting the frame view and identifying the component family category, and combining the annotation scoring rules to optimize the annotation position, the problem of low efficiency of manual annotation in 3D views is solved, and the automatic generation of component annotations is realized, which improves the accuracy of annotation and the aesthetics of drawings.
Patent Information
- Application Number
- CN202511664839.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-13
- Publication Date
- 2026-02-06
AI Technical Summary
In architectural design and construction drawings, manual annotation in the 3D views of HVAC systems is inefficient, and annotations are prone to overlap and confusion, affecting the readability of the drawings and the accuracy of construction.
The view is selected by clipping the box, the component family category is identified, initial annotation information is generated, the annotation position is optimized according to the annotation scoring rules, and the component annotation is automatically generated by using global annotation optimal position control.
It significantly improves drawing output efficiency, reduces manual annotation time, enhances annotation accuracy and drawing aesthetics, supports differentiated annotation rules for various component types, and adapts to complex scenario requirements.
Smart Images
Figure CN121479903A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of three-dimensional views, and in particular to a method, device and medium for automatically generating component annotations in three dimensions. Background Technology
[0002] In the process of architectural design and construction drawing production, especially in the HVAC field, the layout of complex pipes, machine rooms, and other components is difficult to clearly express using only floor plans and sections. Traditional methods rely on designers manually adding annotations to 3D views, which is inefficient and prone to overlapping and confusion, affecting the readability of the drawings and the accuracy of construction. In other words, it has certain limitations. (1) The label position is fixed and cannot be automatically adjusted according to the view content; (2) The labels are prone to overlap and require manual intervention for adjustment; (3) Lack of intelligent annotation rules for different components; (4) It is impossible to achieve global avoidance and aesthetic layout of annotations. Summary of the Invention
[0003] This application provides a method, device, and medium for automatically generating component annotations in three dimensions, which addresses the following technical problem: In existing HVAC three-dimensional views, manual annotation is inefficient and prone to overlap and confusion, affecting the readability of drawings and the accuracy of construction.
[0004] The embodiments of this application adopt the following technical solutions: On one hand, this application provides an automatic component annotation generation method based on three-dimensional dimensions, including: using a clipping box to perform view selection processing on a three-dimensional building component to obtain a locked three-dimensional component view; performing keyword recognition and text annotation processing on the three-dimensional component view related to component family categories to obtain initial component annotation information; restricting the annotation area of a single three-dimensional building component in the three-dimensional component view according to a preset annotation format to obtain a view area to be annotated; performing planar division processing on the view area to be annotated to obtain a target annotation division area; performing local optimal scoring processing on the annotation positions in the target annotation division area according to annotation scoring rules to obtain a local optimal annotation position; performing relevant global optimal annotation position control on each of the local optimal annotation positions in the three-dimensional building component, and obtaining the final annotation content position information of the three-dimensional building component based on the initial component annotation information.
[0005] This application's embodiments significantly reduce the time and labor required for manual annotation by automating component annotation generation, thereby improving work efficiency. Furthermore, automatic identification and annotation reduce errors caused by manual annotation, improving accuracy. Simultaneously, the automatic annotation method can respond quickly and update annotation information in real time when the 3D model is updated. It can also be easily extended to different component family categories, suitable for annotation of various architectural models. Through keyword recognition and scoring rules, annotation optimization can be performed according to preset rules. Moreover, by utilizing annotation area restrictions and planar division processing, the layout of annotation positions can be optimized, improving space utilization efficiency.
[0006] In one feasible implementation, a view selection process is performed on the three-dimensional building component using a clipping frame to obtain a locked three-dimensional component view. Specifically, this includes: selecting the three-dimensional building component using a preset ratio based on the section selection range of the clipping frame to obtain a selection range view; and locking the elevation of the selection range view based on the corner point view of the bounding rectangle of the three-dimensional building component under the current view to obtain the three-dimensional component view.
[0007] In one feasible implementation, the three-dimensional component view is subjected to keyword recognition and text annotation processing related to component family categories to obtain initial component annotation information. Specifically, this includes: manually fine-tuning the three-dimensional component view; identifying HVAC family categories in the three-dimensional component view; wherein, the HVAC family categories include: mechanical equipment, duct accessories, conventional models, duct ends, duct systems, pipes, and pipe accessories; extracting keyword information for each object in the HVAC family category; and generating upper and lower label text for each object based on the keyword information; and performing annotation information generation processing on the HVAC family category, the upper and lower label text to obtain the initial component annotation information.
[0008] In one feasible implementation, according to a preset annotation format, the annotation area of a single three-dimensional building component in the three-dimensional component view is restricted to obtain the view area to be annotated. Specifically, this includes: determining the base point position based on the annotation object leader point of the three-dimensional building component; determining the leader object position based on the center point position of the cuboid and the center point position of the circumscribed cuboid in the non-cuboid state; annotating the base point position and the leader object position to obtain the annotation format; converting the single three-dimensional building component into a planar view to obtain a component planar view; marking the component planar view with a planar selection box according to the clipping box planar view in the three-dimensional component view to obtain a clipping plane area; and expanding the clipping plane area outwards with respect to the relevant annotation area to obtain the view area to be annotated.
[0009] In one feasible implementation, the view area to be labeled is divided into planar regions to obtain the target labeled region. Specifically, this includes: dividing the view area to be labeled into three-view regions based on the three-view line division rules, and removing the clipping plane to obtain three planes to be labeled; wherein the three-view regions include: front view, right view, and top view; and dividing the intersection of the three planes to be labeled into regions according to the order of the labeled regions to obtain the target labeled region.
[0010] In one feasible implementation, according to the annotation scoring rules, the annotation positions in the target annotation division area are subjected to local optimal scoring processing to obtain the local optimal annotation position. Specifically, this includes: constructing the annotation scoring rules based on the leader annotation area score, leader length score, and horizontal movement score; using the annotation scoring rules, performing local optimal scoring processing on the individual annotation position of each piece of information to be annotated in the target annotation division area to obtain the local optimal score; and determining the corresponding annotation position as the local optimal annotation position based on the local optimal score.
[0011] In one feasible implementation, each locally optimal annotation position in a 3D building component is subject to relevant global optimal annotation position control, and the final annotation content position information of the 3D building component is obtained based on the initial component annotation information. Specifically, this includes: processing each locally optimal annotation position into a bounding polygon to obtain an annotation rectangle; performing collision scoring processing on each annotation rectangle using a bounding shape algorithm to obtain a collision score; controlling the placement of extended regions under the relevant global optimal position based on the collision score for the spatial position of each annotation rectangle to obtain a global optimal annotation position based on the 3D building component; and integrating the initial component annotation information with the global optimal annotation position to obtain the final annotation content position information.
[0012] In one feasible implementation, after performing collision scoring on each of the labeled rectangles using a bounding box algorithm to obtain a collision score, the method further includes: adjusting the local optimal score corresponding to the local optimal labeling position of each piece of information to be labeled by reducing the collision score to meet the spatial position requirements of the global optimal position, thereby obtaining the global optimal labeling position under the local optimal score reduction processing.
[0013] Secondly, embodiments of this application also provide an automatic component annotation generation device based on three dimensions, the device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, so that the at least one processor can execute the automatic component annotation generation method based on three dimensions described in any of the above embodiments.
[0014] Thirdly, embodiments of this application also provide a non-volatile computer storage medium, which is a non-volatile computer-readable storage medium storing at least one program, each program including instructions, which, when executed by a terminal, cause the terminal to execute the automatic generation method for component annotation based on three dimensions as described in any of the above embodiments.
[0015] This application provides a method, device, and medium for automatically generating component annotations in three dimensions. Compared with the prior art, the embodiments of this application have the following beneficial technical effects: 1. Enables automatic generation and intelligent arrangement of annotations, significantly improving drawing efficiency; 2. Through local and global optimization mechanisms, overlapping annotations are effectively avoided, improving the aesthetics and readability of drawings; 3. Supports differentiated annotation rules for various component types to meet the needs of complex scenarios in the HVAC industry. Attached Figure Description
[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 A flowchart of a method for automatically generating component annotations in three dimensions is provided for embodiments of this application; Figure 2 A cross-sectional view of the selected area is provided for an embodiment of this application. Figure 3 This application provides a schematic diagram of the lead-out position of a labeled object. Figure 4 A schematic diagram of a marked and restricted area provided in an embodiment of this application; Figure 5 A schematic diagram of a bounded polygon based on global optimization is provided in this application embodiment; Figure 6This is a structural schematic diagram of a component annotation automatic generation device based on three dimensions, provided as an embodiment of this application. Detailed Implementation
[0017] To enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this application.
[0018] This application provides a method for automatically generating component annotations in three dimensions, such as... Figure 1 As shown, the automatic generation method for component annotations in 3D specifically includes steps S101-S106: It should be noted that this application starts with automatic recognition, proceeds to the formation of view clipping frames, adjusts annotation positions based on spatial relationships, and finally generates a partial 3D rendering. It primarily improves design efficiency by summarizing design experience, determining annotation content, and then using the optimal collision solution to meet relevant design requirements. The main components include a view generation module, an annotation generation module, a region division module, a local optimization module, a global optimization module, and a user interaction module. View generation module: Generates a 3D view with a scale of 1:50 based on the user's selected area, and supports showing and hiding the cropping box.
[0019] Label generation module: Automatically generates up and down label content based on component type (such as fan, valve, duct, etc.).
[0020] Region division module: Divides the view range into multiple regions and assigns label leader areas based on the center point position of the components.
[0021] Local optimization module: Determines the optimal position of a single annotation through a scoring mechanism.
[0022] Global optimization module: Employs a bounding box collision detection algorithm to ensure no overlap between annotations.
[0023] User interaction module: Provides view adjustment and annotation confirmation functions, and supports manual fine-tuning.
[0024] S101. Using the clipping box, the three-dimensional building components are selected in the view to obtain the locked three-dimensional component view.
[0025] Specifically, it is necessary to first select the 3D building components by using the section selection area of the clipping frame to perform preset scale selection processing, and then obtain the selection area view.
[0026] Furthermore, based on the corner view of the bounding rectangle of the three-dimensional building component, the selected range view is subjected to elevation locking processing under the current view to obtain the three-dimensional component view.
[0027] In one embodiment, Figure 2 A cross-sectional view of an embodiment of this application is provided, as shown in the figure. Figure 2 As shown, the format of the generated result is determined by whether the clipping frame is displayed based on the user's choice of the function. The clipping frame, also known as the section frame, generates a 3D view with a scale of 1:50 by controlling the size of the user-selected bounding box. For example, when the user-selected bounding box is 10000x10000, the clipping frame size seen in the top view of the generated 3D view is also 10000x10000. The angle is controlled at the southeast corner, the corner point of the front view + top view + right view, and the height of the clipping frame is the range between the two elevations below the current view.
[0028] S102. Perform keyword recognition and text annotation processing on the 3D component view for relevant component family categories to obtain initial component annotation information.
[0029] Specifically, the three-dimensional component views are first manually fine-tuned.
[0030] Furthermore, identify the HVAC family categories in the 3D component view. These HVAC family categories include: mechanical equipment, duct fittings, standard models, duct terminals, duct systems, pipes, and pipe fittings.
[0031] Furthermore, keyword information is extracted for each object within the HVAC family category. Based on this keyword information, the upper and lower heading text for each object's label is generated.
[0032] Furthermore, the HVAC family category, the upper and lower label texts are processed to generate label information, resulting in the initial component label information.
[0033] In one embodiment, after the 3D component view is generated, it also provides space for designers or users to manually modify it. When the designer or user has modified the view as needed, clicking the "Finish" button will lock the current view and annotate the HVAC objects.
[0034] Table 1
[0035] Table 1 shows the initial component annotation information under the HVAC family category in the 3D component view, i.e., the obtained object keywords, for example: Mechanical Equipment: Obtain mechanical equipment whose family name keyword contains 'fan'. Duct Accessories: Obtain duct accessories whose family name keyword contains 'valve'. Silencers are special; only those with the keyword "silencer" are obtained. Pipe Accessories: Obtain pipe accessories whose family name keyword contains 'valve'.
[0036] As a feasible implementation method, to address the issue of duplicate duct / pipe dimension markings, the system can automatically check whether the ducts / pipes on both sides of the connector (pipe fitting, pipe fitting, duct accessory) have the same dimensions. If they are the same, only one duct / pipe needs to be marked; if they are different, both need to be marked. The marked duct / pipe is the longest duct / pipe among the multiple ducts / pipes on the drawing.
[0037] S103. According to the preset annotation format, the annotation area of a single three-dimensional building component in the three-dimensional component view is restricted to obtain the view area to be annotated.
[0038] Specifically, the location of the base point must be determined based on the reference points of the annotation objects of the three-dimensional building components.
[0039] Furthermore, based on the position of the center point of the cuboid and the position of the center point of the circumscribed cuboid in the non-cuboid configuration, the position of the leading-out object is determined.
[0040] Furthermore, the base point position and the position of the leading object are labeled and configured to obtain the labeling format.
[0041] As a feasible implementation method, Figure 3 This application provides a schematic diagram of the marker object's lead-out position, as shown in the embodiment. Figure 3 As shown, the base point position needs to be determined first, where: Base point position: the point from which the annotation object originates. Point of origin: the position of the object's center point. If the object is not a cuboid, the center point of its circumscribed cuboid is used as the annotation origin point position. The logic for other object origin points is the same as that for the mechanical equipment family.
[0042] Furthermore, the single three-dimensional building component is transformed into a planar view to obtain a planar view of the component.
[0043] Furthermore, based on the clipping frame planar view in the 3D component view, it is also necessary to mark the planar selection area of the component planar view to obtain the clipping plane region.
[0044] Furthermore, the clipping plane area is expanded outwards by the relevant annotation area to obtain the view area to be annotated.
[0045] In one embodiment, Figure 4A schematic diagram of a marked and restricted area provided for an embodiment of this application, such as... Figure 4 As shown, a single annotation implementation (local optimization) treats the 3D component view as a planar view. The size of the clipping frame plane is then manually determined by the user and can be adjusted after the 3D view is generated. The adjusted clipping frame forms an inner hand-drawn schematic area (the clipping plane area), which is considered a planar view, allowing the determination of the length of each line within this inner frame. The outer hand-drawn frame is an extension of the inner hand-drawn frame (the area to be annotated), used to limit the length of the annotation leader lines, requiring annotations to be performed only within the area to be annotated.
[0046] S104. Divide the view area to be annotated into planes to obtain the target annotation area.
[0047] Specifically, based on the three-view line division rules, the area to be labeled needs to be divided into three views, and the clipping plane needs to be removed to obtain three planes to be labeled. The three views include: front view, right view, and top view.
[0048] Furthermore, the intersection of the three planes to be labeled is divided into regions according to the order of the labeling areas, thus obtaining the target labeling region.
[0049] In one embodiment, such as Figure 4 As shown, the area to be annotated (the schematic area between the two hand-drawn lines) is divided into three planes: 1 (front view), 2 (right view), and 3 (top view). Then, it determines which area within the green frame the center point of the object to be annotated is located in. The annotation content for the object to be annotated is located within the area to be annotated (this is not mandatory; if the system determines that the area to be annotated is full, it can be extended to other surfaces according to subsequent global scoring steps). If the center point of the object to be annotated is located at the boundary of an area, then... Figure 4 The arrows in the diagram indicate the rules for dividing the area.
[0050] S105. According to the annotation scoring rules, the annotation positions in the target annotation division area are processed for local optimal scoring to obtain the local optimal annotation positions.
[0051] Specifically, a labeling scoring rule is constructed based on the score of the lead-in label area, the score of the lead-in length, and the score of the horizontal movement.
[0052] Furthermore, by using the annotation scoring rules, the local optimal scoring process is performed on the individual annotation positions of each piece of information to be annotated in the target annotation division area to obtain the local optimal score.
[0053] Furthermore, based on the local optimal score, the corresponding annotation position is determined as the local optimal annotation position.
[0054] In one embodiment, (1) a rating for the labeled area is given, for example: the area to which the label belongs: 10 points; the area not to which the label belongs: 5 points.
[0055] (2) Leader length score: The distance from the starting point of the annotation area to the center point of the annotation object component is defined as the leader length; the initial leader length is 5000, and the smaller the absolute value of the difference from 5000, the higher the score, for example: |x-5000|=0 (5 points) 0 < |x - 5000| ≤ 500 (4 points) 500 < |x - 5000| ≤ 1000 2 points; 1000 < |x - 5000| ≤ 1500 1 point; 1500 < |x - 5000| ≤ 2000 0 points; The length of the leader line can be defined and varied by the red and green bounding boxes, and can be drawn vertically or horizontally from the center point of the object to be marked to the intersection of the red prohibited area and the green area.
[0056] (3) Horizontal movement score: such as Figure 4 As shown, the distance from the bottom right corner of the annotation area (cutting plane) (the corner closest to the center of the pipe) should be as small as possible (the minimum distance is specified as 900 to ensure the continuity of the leader line and the text annotation line). The horizontal movement distance L is as follows: |L|=900 5 points; 900<|L|≤1100 4 points; 1100 < |L| ≤ 1300 (2 points) 1300 < |L| ≤ 1500 1 point; 1500 < |L| gets 0 points.
[0057] Finally, based on the above scoring rules, the local optimal score of a single labeled part is calculated, and the corresponding labeled position is then determined as the local optimal labeled position.
[0058] S106. Perform relevant global annotation optimal position control on each local optimal annotation position in the three-dimensional building component, and obtain the final annotation content position information of the three-dimensional building component based on the initial component annotation information.
[0059] Specifically, each local optimal annotation position is processed by circumscribed polygons to obtain the annotation rectangle.
[0060] Furthermore, a collision score is obtained by performing a collision scoring process on each labeled rectangle using a bounding box algorithm.
[0061] Furthermore, based on the collision score, the spatial position of each labeled rectangle is controlled by the extension area placement under the relevant global optimal position, thus obtaining the global optimal label position based on the three-dimensional building components.
[0062] Furthermore, the initial component annotation information is integrated with the global optimal annotation position to obtain the final annotation content location information.
[0063] As a feasible implementation method, after performing collision scoring on each labeled rectangle using the bounding box algorithm to obtain the collision score, it is also possible to: reduce and adjust the local optimal score corresponding to the local optimal labeling position of each piece of information to be labeled according to the collision score, so as to meet the spatial position requirements of the global optimal position, and obtain the global optimal labeling position under the local optimal score reduction processing.
[0064] In one embodiment, such as Figure 4 As shown, the global optimal requirement for the annotation of 3D building components is that the annotations do not collide with each other in the 1st, 2nd, and 3rd faces of the region division. The bounding box of each text annotation is found by using the bounding shape method - precise collision detection polygon method.
[0065] In one embodiment, it is necessary to first approximate the bounding polygon of each annotation. Figure 5 A schematic diagram of a bounded polygon based on global optimization is provided for an embodiment of this application; as shown below. Figure 5 As shown in the figure, the longer side of the rectangle containing the upper and lower line text annotations is the length of the leader line below the text; the width of the rectangle is the height of the upper and lower line text, both of which are set to 3mm. The global optimum score is obtained by judging the bounding boxes between these annotations. 1) Method 1: Piecewise function scoring items: Collision count: x=0, score 100 points; Collision count: x=1, score 90 points; Collision count, x=2, score 60 points; Collision count, x=3, score 40 points; Collision count, x=4, score 0; Arrangement method: Each label can be generated separately. After each label is generated, it occupies a certain space. This space is the space to be avoided. If there is no room for the label in the end, the local label placement score can be reduced to place it (such as lengthening the label leader line, controlling the label's belonging area, etc.).
[0066] 2) Method Two: Continuous Function Scoring Item The optimal global score is calculated using a linear function with a deduction system: Score = 100 - 10x, i.e.: Collision count: x=0, score 100 points; Collision count: x=1, score 90 points; Collision count: x=2, score 80 points.
[0067] Finally, through global calculation based on piecewise function scoring terms or continuous function scoring terms, the optimal global annotation position for the three-dimensional building components can be obtained. Then, combined with the initial component annotation information, the final annotation content position information can be obtained.
[0068] One special case is duct labeling. When the duct length on the drawing is less than 500mm, the duct is not labeled in the 3D detail design. Only ducts with a length greater than 500mm are labeled.
[0069] In addition, embodiments of this application also provide an automatic component annotation generation device based on three-dimensional dimensions, such as... Figure 6 As shown, the 600 automatic component annotation generation device based on 3D modeling specifically includes: At least one processor 601; and a memory 602 communicatively connected to at least one processor 601; wherein the memory 602 stores instructions executable by at least one processor 601 to enable at least one processor 601 to execute: By using the clipping box, the 3D building components are selected in the view to obtain a locked 3D component view; The 3D component view is processed by keyword recognition and text annotation of component family categories to obtain initial component annotation information; Based on the preset annotation format, the annotation area of a single three-dimensional building component in the three-dimensional component view is restricted to obtain the view area to be annotated; The view area to be annotated is divided into planar regions to obtain the target annotation area; According to the annotation scoring rules, the annotation positions in the target annotation region are subjected to local optimal scoring processing to obtain the local optimal annotation positions; Each local optimal annotation position in the 3D building component is controlled by the relevant global optimal annotation position, and the final annotation content position information of the 3D building component is obtained based on the initial component annotation information.
[0070] This application's embodiments significantly reduce the time and labor required for manual annotation by automating component annotation generation, thereby improving work efficiency. Furthermore, automatic identification and annotation reduce errors caused by manual annotation, improving accuracy. Simultaneously, the automatic annotation method can respond quickly and update annotation information in real time when the 3D model is updated. It can also be easily extended to different component family categories, suitable for annotation of various architectural models. Through keyword recognition and scoring rules, annotation optimization can be performed according to preset rules. Moreover, by utilizing annotation area restrictions and planar division processing, the layout of annotation positions can be optimized, improving space utilization efficiency.
[0071] The various embodiments in this application are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the device and medium embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the description of the method embodiments.
[0072] The devices and media provided in this application are one-to-one with the methods. Therefore, the devices and media also have similar beneficial technical effects as their corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the devices and media will not be repeated here.
[0073] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0074] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0075] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0076] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0077] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.
[0078] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.
[0079] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.
[0080] It should also be noted that 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 limitation, 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.
[0081] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of this specification.
Claims
1. A method for automatically generating component annotations in three dimensions, characterized in that, The method includes: By using the clipping box, the 3D building components are selected in the view to obtain a locked 3D component view; The three-dimensional component view is processed by keyword recognition and text annotation of component family categories to obtain initial component annotation information; According to the preset annotation format, the annotation area of a single three-dimensional building component in the three-dimensional component view is restricted to obtain the view area to be annotated; The view area to be labeled is divided into planar regions to obtain the target labeled region. According to the annotation scoring rules, the annotation positions in the target annotation division area are subjected to local optimal scoring processing to obtain the local optimal annotation positions; Each of the local optimal annotation positions in the three-dimensional building components is subject to relevant global optimal annotation position control, and the final annotation content position information of the three-dimensional building components is obtained based on the initial component annotation information.
2. The method for automatically generating component annotations in three dimensions according to claim 1, characterized in that, By using a clipping box, the 3D building components are selected in the view to obtain a locked 3D component view, specifically including: By using the cross-sectional selection range of the clipping frame, the three-dimensional building component is selected at a preset ratio to obtain a selection range view; Based on the corner view of the bounding rectangle of the three-dimensional building component, the selected range view is subjected to elevation locking processing under the current view to obtain the three-dimensional component view.
3. The method for automatically generating component annotations in three dimensions according to claim 1, characterized in that, The three-dimensional component view is processed by keyword recognition and text annotation of component family categories to obtain initial component annotation information, specifically including: The three-dimensional component view is manually fine-tuned; Identify the HVAC family categories in the three-dimensional component view; wherein, the HVAC family categories include: mechanical equipment, duct accessories, conventional models, duct terminals, duct systems, pipes, and pipe accessories; Extract keyword information for each object in the HVAC family category; and generate the upper and lower label text for each object based on the keyword information; The initial component annotation information is obtained by processing the HVAC family category, the upper line text of the annotation, and the lower line text of the annotation.
4. The method for automatically generating component annotations in three dimensions according to claim 1, characterized in that, According to a preset annotation format, the annotation area of a single 3D building component in the 3D component view is restricted to obtain the view area to be annotated, specifically including: Based on the leader points of the annotation objects of 3D building components, the location of the base point is determined; The location of the leading object is determined based on the center point of the cuboid and the center point of the circumscribed cuboid in the non-cube form. The base point position and the lead-out object position are labeled and configured to obtain the labeling form; Transform a single 3D building component into a planar view to obtain a planar view of the component; Based on the clipping frame planar view in the 3D component view, the component planar view is marked with a planar selection box to obtain the clipping plane region; The clipping plane area is expanded outwards by the relevant annotation area to obtain the view area to be annotated.
5. The method for automatically generating component annotations in three dimensions according to claim 1, characterized in that, The view area to be labeled is divided into planar regions to obtain the target labeled region, specifically including: Based on the three-view line division rules, the area to be labeled is divided into three views, and the clipping plane is removed to obtain three planes to be labeled; wherein the three views include: front view, right view and top view; The target annotation area is obtained by dividing the intersection of the three planes to be annotated into an area with a sequential assignment of the relevant annotation regions.
6. The method for automatically generating component annotations in three dimensions according to claim 1, characterized in that, According to the annotation scoring rules, the annotation positions in the target annotation region are subjected to local optimal scoring processing to obtain the local optimal annotation positions, specifically including: The annotation scoring rules are constructed based on the leader label area score, leader length score, and horizontal movement score. Using the aforementioned annotation scoring rules, a local optimal scoring process is performed on the individual annotation position of each piece of information to be annotated in the target annotation division area to obtain a local optimal score; Based on the local optimal score, the corresponding annotation position is determined as the local optimal annotation position.
7. The method for automatically generating component annotations in three dimensions according to claim 1, characterized in that, Each locally optimal annotation position in the 3D building component is subject to relevant global annotation optimal position control, and based on the initial component annotation information, the final annotation content position information of the 3D building component is obtained, specifically including: Each of the local optimal annotation positions is processed by circumscribing a polygon to obtain an annotation rectangle; The collision score is obtained by performing collision scoring on each of the marked rectangles using the bounding box algorithm. Based on the collision score, the spatial position of each labeled rectangle is controlled by the extension area placement under the relevant global optimal position to obtain the global optimal label position based on the three-dimensional building component; The initial component annotation information is integrated with the global optimal annotation position to obtain the final annotation content position information.
8. The method for automatically generating component annotations in three dimensions according to claim 7, characterized in that, After performing collision scoring on each labeled rectangle using a bounding box algorithm to obtain a collision score, the method further includes: Based on the collision score, the local optimal score corresponding to the local optimal annotation position of each piece of information to be annotated is reduced and adjusted to meet the spatial position requirements of the global optimal position, thus obtaining the global optimal annotation position under the local optimal score reduction processing.
9. An automatic component annotation generation device based on three-dimensional model, characterized in that, The device includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, enabling the at least one processor to perform an automatic component annotation generation method based on any one of claims 1-8.
10. A non-volatile computer storage medium, characterized in that, The storage medium is a non-volatile computer-readable storage medium that stores at least one program, each program including instructions that, when executed by a terminal, cause the terminal to perform an automatic component annotation generation method based on any one of claims 1-8.