Construction calibration method based on 4D building model

Abstract

The application provides a construction calibration method based on a 4D building model, and relates to the technical field of building construction management. The method comprises the following steps: constructing a 4D building model of a building project; for the building project in any construction stage, scanning a construction site of the building project to obtain a 3D building model of the site, and generating a deviation diagram based on a comparison result of the 3D building model of the site and a planned state model; wherein the planned state model is a state of the 4D model in the current construction stage; and generating a construction calibration scheme of the building project in the current construction state based on the deviation diagram. The method realizes dynamic visualization and accurate calibration of the construction process by constructing the 4D building model and combining the 3D scanning technology, and can improve the efficiency of construction progress management.

Description

Technical Field

[0001] This application relates to the field of building construction management technology, specifically to a construction calibration method based on a 4D building model. Background Technology

[0002] In traditional construction management, project progress monitoring and quality control primarily rely on a combination of two-dimensional drawings and manual inspection. Construction teams typically conduct on-site construction based on floor plans, elevations, and sections provided by the design institute. Progress management involves recording the completion status of each process using paper or spreadsheets, which are then manually compared with pre-defined Gantt charts or bar charts. In this model, verifying the correlation between design intent and construction status often depends on the experience of technical personnel, using periodic manual measurements to confirm whether on-site construction meets design requirements. Because two-dimensional drawings cannot visually represent the spatial relationships of building components, errors, omissions, and incompleteness frequently occur during construction due to misunderstandings of the drawings. These problems are often only discovered after the process is completed, increasing rework and rectification costs.

[0003] In terms of schedule management, traditional methods lack dynamic visualization tools. Adjustments to construction plans are usually based on static schedules, making it difficult to reflect actual on-site progress in real time. Data recorded by manual inspections is subjective and lagging, failing to accurately capture three-dimensional spatial information during construction, leading to inaccurate location of quality defects. Furthermore, there is a lack of effective linkage between the construction schedule and the design model; design changes or on-site adjustments are difficult to promptly reflect in the schedule, easily causing misallocation of resources. For example, when a process is delayed due to design issues, the labor, materials, and machinery arrangements for subsequent processes may not be adjusted in time, resulting in idle work or rushed work, affecting overall construction efficiency.

[0004] While existing Building Information Modeling (BIM) applications have improved design visualization to some extent, most projects remain at the level of static 3D models. Model data and schedule management systems are independent, creating information silos. Construction teams cannot use model data for dynamic schedule simulation, and it is even more difficult to predict potential design conflicts or schedule risks before construction begins. For example, collisions involving MEP (Mechanical, Electrical, and Plumbing) lines usually only surface during construction, and temporary adjustments may affect the construction sequence of other disciplines, thus slowing down the overall schedule. Furthermore, traditional manual measurement and recording methods are insufficient to construct a digital twin mapping of the entire construction process, making real-time monitoring and historical tracking of construction status impossible. This results in a lack of data support for quality control and schedule optimization, ultimately leading to inefficient construction schedule management. Summary of the Invention

[0005] This application provides a construction calibration method based on a 4D building model. This technical solution realizes dynamic visualization and precise calibration of the construction process by constructing a 4D building model and combining it with on-site 3D scanning technology, which can improve the efficiency of construction progress management.

[0006] In a first aspect, the present invention provides a construction calibration method based on a 4D building model, comprising: constructing a 4D building model of a building project; wherein the 4D building model includes multiple component sub-models, each sub-model being bound to a corresponding timeline and keyframes, the timeline representing the time span of the construction plan of the component from start to finish, and the keyframes being task nodes defining the state changes of the component; for the building project at any construction stage, scanning the construction site of the building project to obtain a site 3D building model, generating a deviation diagram based on the comparison results between the site 3D building model and the planned state model; wherein the planned state model is the state of the 4D model at the current construction stage; and generating a construction calibration scheme for the building project in the current construction state based on the deviation diagram.

[0007] According to one embodiment of the present invention, for any construction stage of the building project, scanning the construction site of the building project to obtain a 3D building model of the site, and generating a deviation diagram based on the comparison results between the 3D building model of the site and the planned state model, includes: scanning the construction site with a laser scanning device to obtain point cloud data of the construction site; constructing the 3D model of the site based on the point cloud data; spatially aligning and matching the coordinate system of the 3D model of the site with the planned state model; comparing the 3D model of the site with the planned state model based on a three-dimensional difference detection algorithm to determine the deviation value of each detection point; and generating the deviation diagram based on the deviation value of each detection point.

[0008] According to one embodiment of the present invention, in the deviation diagram, the first color area represents the normal range that meets the design requirements, the second color area represents the range that exceeds the design requirements, and the third color area represents the range that does not meet the design requirements.

[0009] According to one embodiment of the present invention, constructing a 4D building model of a building project includes: acquiring construction plan data of the building project; generating baseline plan features and a plan layout based on the construction plan data; wherein the baseline plan features include a work breakdown structure encoding, a time reference coordinate system, and a critical path feature chain characterizing the building project, and the plan layout characterizes the execution logic of the building project; performing lightweight processing on a 3D building model based on the baseline plan features; performing spatial reconstruction on the lightweight 3D building model based on the baseline plan features to obtain a sub-model corresponding to each component in the 3D building model; for any sub-model of a component, binding the sub-model to a time axis and keyframes respectively; wherein the time axis characterizes the time span of the construction plan of the component from start to finish, and the keyframes are task nodes defining the state changes of the component; driving each sub-model based on the plan layout to obtain the 4D building model.

[0010] According to one embodiment of the present invention, the step of acquiring construction plan data of a building project and generating baseline plan features and plan layout based on the construction plan data includes: performing a structured decomposition of the engineering tasks of the building project based on the construction plan data to obtain multiple task units, each task unit corresponding to a work breakdown structure code; establishing the time baseline coordinate system based on the construction period data in the construction plan data; determining the dependency relationship of each task unit based on the critical path algorithm, and generating the critical path feature chain based on the dependency relationship.

[0011] According to one embodiment of the present invention, the lightweighting process of the 3D building model based on the baseline plan features includes: preprocessing the 3D building model; adjusting the level of detail of the 3D building model based on the critical path feature chain; filtering the attribute fields of the components in the 3D building model based on the work breakdown structure encoding; compressing the high-resolution textures in the 3D building model; and merging similar materials in the 3D building model into a single material group to obtain the lightweighted 3D building model.

[0012] According to one embodiment of the present invention, the step of spatially reconstructing the lightweight 3D building model based on the baseline plan features to obtain sub-models corresponding to each component in the 3D building model includes: acquiring the geometric topological relationships and spatial distribution features of the 3D building model, and establishing a component attribute index table of the 3D building model based on the geometric topological relationships and spatial distribution features; determining a target splitting strategy based on the construction method of the building project, and splitting the 3D building model into multiple levels of sub-models based on the target splitting strategy; for a first component that exists only in one construction stage, identifying the attribute information of each first component in the component attribute index table, and grouping the first component based on the attribute information to obtain a sub-model corresponding to each first component; for a second component that exists in multiple construction stages, cutting the second component according to the actual construction joint position based on a spatial classification algorithm to obtain a sub-model corresponding to each second component.

[0013] According to one embodiment of the present invention, binding the sub-model to a timeline and keyframes for any of the aforementioned components includes: creating a timeline for the sub-model; determining the position of the sub-model's time window on the timeline; inserting an initial keyframe at the start time of the time window and defining the initial state of the sub-model; inserting one or more progress keyframes in the middle of the time window and defining the progress state of the sub-model; inserting an end keyframe at the end time of the time window and defining the end state of the sub-model; and generating a transition effect according to the state transition requirements between adjacent keyframes.

[0014] According to one embodiment of the present invention, the step of driving each of the sub-models based on the planned layout to obtain the 4D building model includes: establishing a mapping relationship between the sub-model of each component and the work breakdown structure code; locating the target component group based on the mapping relationship, and filtering the component states that should be activated in the target component group according to the time axis; dynamically displaying the component according to the component state, and prioritizing the rendering of the component when the critical task time window is triggered.

[0015] In a second aspect, the present invention also provides a computer device, the computer device including 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 method of the above embodiments.

[0016] Compared with existing technologies, the beneficial effects of this application are: by deeply integrating 4D building models with 3D scanning, a closed-loop control system for construction progress management is constructed. In the model building phase, a 4D building model dynamically reflects the construction progress in the time dimension; in the progress monitoring phase, a 3D building model reflecting the current state of the construction site is constructed. By comparing the 3D building model with the corresponding model state in the 4D building model at the current construction stage, it is possible to intuitively show whether the current construction project is ahead of schedule or behind schedule, thereby improving the efficiency of construction progress management. Attached Figure Description

[0017] Figure 1 A schematic diagram illustrating the steps of the construction calibration method based on a 4D building model provided in this application embodiment.

[0018] Figure 2 A schematic diagram of the construction site provided for an embodiment of this application.

[0019] Figure 3 A schematic diagram of the deviation provided for an embodiment of this application.

[0020] Figure 4 A comparative diagram of the 4D building model and the actual construction situation provided in the embodiments of this application. Detailed Implementation

[0021] The present application will now be described in further detail with reference to experimental examples and specific embodiments. However, this should not be construed as limiting the scope of the subject matter of the present application to the following embodiments. All technologies implemented based on the content of the present application fall within the scope of protection of the present application.

[0022] Unless otherwise specified, the terms "upper," "lower," "left," "right," "center," "inner," "outer," and "side" used in the description of specific embodiments of this application to indicate orientation or positional relationships are based on the orientation or positional relationships shown in the accompanying drawings, or the orientation or positional relationship in which the product / equipment / device is usually placed during use. These terms are merely for the purpose of facilitating the description of the solution in this application or simplifying the description in specific embodiments, so as to enable those skilled in the art to quickly understand the solution, and do not indicate or imply that a particular device / component / element must have a specific orientation, or be constructed and operated in a specific positional relationship. Therefore, they should not be construed as limitations on this application.

[0023] In the description of the embodiments of this application, technical terms such as "first" and "second" only distinguish one entity or operation from another, and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary or secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0024] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0025] Please refer to Figure 1 , Figure 1 A schematic diagram illustrating the steps of the construction calibration method based on a 4D building model provided in this application embodiment. The construction calibration method based on a 4D building model may include:

[0026] S1. Construct a 4D architectural model of the building project.

[0027] The meaning of 4D is to add a time dimension to the 3D BIM building model. By driving the 3D building model in the time dimension, the 3D building model can display corresponding states (such as showing / hiding, color, position) according to the changes in time, so as to obtain a 4D building model that can dynamically reflect the construction progress in the time dimension.

[0028] The 4D building model includes multiple component sub-models, each of which is bound to a corresponding timeline and keyframes. The timeline represents the time span of the construction plan for the component from start to finish, and the keyframes are task nodes that define the changes in the state of the component.

[0029] S2. For any construction project at any construction stage, scan the construction site of the project to obtain a 3D building model. Generate a deviation diagram based on the comparison between the 3D building model and the planned state model.

[0030] The planned state model represents the current state of the 4D model during the current construction phase. For example, assuming a high-rise building project is currently in the main structural construction phase on the 10th floor, the planned state model would show that the walls, beams, and columns on the 10th floor should have been poured, the formwork support for the 11th floor should have been erected, and the construction preparation work for the 12th floor should be underway. This model not only includes the three-dimensional geometry of these components but also clearly distinguishes between completed, under-construction, and unfinished sections through visualization methods such as color or transparency.

[0031] When the actual 3D model obtained by on-site scanning is compared with the planned status model, the system automatically identifies the differences between the actual progress and the plan. For example, it may find that a section of the wall on the 10th floor has not been poured (it should have been completed but has not been constructed), or that the scaffolding on the 11th floor is ahead of schedule (it was not required by the plan but has been constructed). This generates an intuitive diagram of the deviation.

[0032] S3. Generate a construction calibration scheme for the building project under the current construction status based on the deviation diagram.

[0033] For example, if quality issues (dimensions, defects, installation, etc.) arise after calibration, on-site rectification will be carried out accordingly, including repairs, reinstallation, and, if necessary, re-construction or design optimization.

[0034] After discrepancies are found during calibration, a decision is made on whether to rescan. If the defect is minor, no rescan is needed; if there is a major defect requiring rework, a rescan is required. If design defects are found during construction and design optimization is performed, the 4D model needs to be optimized and the 3D model needs to be rescanned.

[0035] In the above implementation process, a closed-loop control system for construction progress management was constructed through the deep integration of 4D building models and 3D scanning. During the model building phase, a 4D building model dynamically reflects the construction progress over time. In the progress monitoring phase, a 3D building model reflecting the current state of the construction site is constructed. By comparing the 3D building model with the corresponding model state in the 4D building model at the current construction stage, it is possible to intuitively show whether the current construction project is ahead of schedule or behind schedule, thereby improving the efficiency of construction progress management.

[0036] In some embodiments, the specific implementation process of step S2 above may include:

[0037] The construction site was scanned to obtain point cloud data of the construction site. Please refer to [link / reference]. Figure 2 , Figure 2This is a schematic diagram of a construction site provided for an embodiment of this application. Specifically, a 3D laser scanner can be used to perform a full-range scan of the construction site, obtaining dense spatial coordinate points on the building surface through the principle of laser ranging, forming a high-precision point cloud dataset. This point cloud data contains the spatial location and geometric feature information of the actual construction state of the building components.

[0038] The on-site 3D building model is constructed based on the point cloud data. Specifically, the collected raw point cloud data is imported into modeling software, and through processing steps such as point cloud filtering to remove noise, point cloud registration to integrate multi-station scan data, and surface reconstruction to generate triangular meshes, a 3D model reflecting the actual on-site conditions is constructed.

[0039] The on-site 3D building model and the planned state model are spatially aligned and their coordinate systems matched. This can be achieved by selecting at least three corresponding feature points from both models as references and using the least squares method to calculate coordinate transformation parameters, ensuring that the two models have a consistent spatial reference system under a unified coordinate system.

[0040] A 3D difference detection algorithm is used to compare the on-site 3D building model with the planned state model to determine the deviation value at each detection point. After the two models are spatially aligned, a spatial grid index is established using the 3D difference detection algorithm, and the shortest distance from each detection point on the on-site model to the corresponding surface on the planned model is calculated as the deviation value. For structural components, key parameters such as their outline dimensions, positional offset, and installation angle can be detected; for building surfaces, quality indicators such as flatness and finished surface elevation can be evaluated. By determining the deviation value and direction of each detection point, a complete deviation dataset is formed.

[0041] The deviation diagram is generated based on the deviation value of each detection point. Finally, a visual deviation diagram is generated based on the calculated deviation values. In the diagram, a color gradient is used to visually represent the degree of deviation. In the deviation diagram, the first color area represents the normal range that meets the design requirements, the second color area represents the range that exceeds the design requirements, and the third color area represents the range that does not meet the design requirements. For example, please refer to... Figure 3 , Figure 3 A schematic diagram of the deviation provided for an embodiment of this application. Figure 3 In the design, the first color is green, the second color is blue, and the third color is red.

[0042] In some embodiments, the method of constructing a 4D architectural model of a building project may include:

[0043] S11. Obtain construction plan data for the building project, and generate baseline plan features and plan layout based on the construction plan data.

[0044] The baseline plan features include a Work Breakdown Structure (WBS) code, a time reference coordinate system, and a critical path feature chain representing the construction project. The WBS code serves as an index label for spatial reconstruction of the 3D building model, breaking down the project into manageable task units. Each component is bound to a WBS code, thus associating it with the project's schedule tasks. The WBS code is retained throughout the component development process of the 4D building model to index each component's corresponding task within the entire construction project. The time reference coordinate system serves as the time reference anchor point for the construction project. The critical path feature chain represents the task logic of the construction project, using the planned start date of the construction project as the zero point of the time axis, with the time of all tasks referenced to this point.

[0045] The plan layout represents the execution logic of the construction project. In this embodiment, the plan layout can be a complete process generated by Primavera P6, which arranges the time sequence, binds logical relationships, and optimizes resource allocation for each task in the construction project according to project objectives, resource constraints, and construction logic.

[0046] S12. Lightweighting of 3D building models based on baseline plan features.

[0047] The 3D building model can be a Revit model or a Civil 3D model. Lightweight processing refers to reducing the complexity of the original 3D BIM model through geometric simplification, attribute filtering, and texture compression to meet the requirements of real-time rendering and interaction. Specifically, based on the engineering semantics contained in the feature point baseline plan features, it selectively retains details of important structural components and key parts, simplifies the model complexity of non-critical areas such as standard repetitive components, while retaining all attribute data related to schedule management (such as component numbers, construction stages, etc.). This targeted processing reduces the amount of data in the model while fully preserving all the information required for 4D schedule simulation of the building project, thus laying the foundation for subsequent visualization of the construction process.

[0048] S13. Based on the characteristics of the baseline plan, the 3D building model after lightweighting is spatially reconstructed to obtain the sub-model corresponding to each component in the 3D building model.

[0049] The decomposition process during spatial reconstruction can be based on rules formulated according to construction logic, task allocation, and rendering requirements. For example, according to the construction sequence (such as foundation and main structure), the model can be decomposed into "earthwork excavation model", "basement structure model", "above-ground main structure model", etc.; according to professional subcontracting or system type, the model can be decomposed into "structural model", "mechanical and electrical model", "decorative model", or further refined into "water supply and drainage system", etc.; according to material or function, the model can be decomposed into "concrete component model", "steel structure model", "temporary support model", etc.; according to floors, sections, or construction phases, the model can be decomposed into "1-5 floor model", "6-10 floor model", etc.; or according to the task level in the schedule plan, the model can be decomposed into sub-models corresponding to the WBS code in the P6 plan.

[0050] The splitting can be done through rule-based splitting using Dynamo (a Revit plugin), which reads the attribute fields of model components (such as Phase, System Type, and WBS Code) and filters components according to preset rules (such as filtering components with Phase = Structural Construction and System Type = Concrete). The filtered results can be exported as independent NWC / NWD files, or sub-models can be created directly in Revit.

[0051] S14. For any component's sub-model, bind the sub-model to the timeline and keyframes respectively.

[0052] The timeline represents the time span of the construction plan for the component from start to finish. The timeline can be a linear time scale axis, usually divided by day / week / month, to represent the time span of the project from start to finish. When the user drags the timeline pointer, it can trigger the model status update, such as switching the display status of the model component or controlling the color change of the model component.

[0053] All the split sub-models are imported into the rendering software for binding operations. This application embodiment uses Fuzor software for illustration.

[0054] Import the schedule and model. First, import the Primavera P6 schedule (XML format) into Fuzor, parse the time range of the tasks, including start and end dates, etc.; then load the split 3D model and mechanical components.

[0055] Components are automatically associated with tasks. Select a component in Fuzor and bind it to a timeline task. For unmatched components, you can manually associate them with the corresponding task by dragging and dropping.

[0056] Keyframe definition and animation settings. By selecting a component, setting key nodes on the timeline, and inserting keyframes, you can set the component's display color to a different color at each stage. For example, locate the start time of its task on the timeline and add a start keyframe; locate the end time and add an end keyframe; you can set the first color at the start time of the task in a stage, and change it to a second color at the actual completion time of the task.

[0057] S15. Drive each sub-model based on the planned layout to obtain a 4D building model.

[0058] By incorporating a chronological order into the constructed 3D building model, a 4D building model is created, linking construction progress with modeling elements. This allows non-technical personnel and design professionals to not only grasp the final static design intent but also identify problems arising during planning changes, phased construction, and demolition. The 4D building model can be used to identify design problems and planning errors before construction begins, when resolving them is less costly. Furthermore, the 4D building model provides a visual overview of the entire process, showing the progress of construction activities throughout the project lifecycle.

[0059] For example, the detailed implementation steps of driving each of the sub-models based on the planned layout may include:

[0060] Timeline parameter configuration. Set the time granularity, associate with external data sources, and enable the timeline to dynamically read actual dates. Synchronize the actual progress dates to the Fuzor timeline via API.

[0061] Select the component, set keyframes on the timeline, and select the component's display status, display color, and display position at the task start time, important task node time, and task end time.

[0062] Dynamic state-driven architecture allows the use of expressions to associate component state with progress data, such as:

[0063] / / Component transparency = 1 - (Actual progress percentage / 100)

[0064] element.opacity=1-(task.actual_progress / 100)

[0065] Please refer to Figure 4 , Figure 4This diagram illustrates a comparison between the 4D building model and the actual construction status provided in this application embodiment. The left side shows the planned construction status of the CCC Building project as displayed in the 4D building model, while the right side shows the actual construction status. The upper part of the diagram shows the timeline of the 4D building model, the lower right corner shows a progress comparison between the planned and actual construction status, and the upper right corner shows the time progress. The comparison diagram clearly shows that the current construction progress is 39.12% of the total project progress, while the planned construction progress at the current time is 44.81% of the total project progress. Therefore, the current construction progress is slower than the planned construction progress.

[0066] In some optional embodiments, when implementing step S11 based on Primavera P6 or Microsoft Project, it may specifically include the following steps:

[0067] Based on the construction plan data, the engineering tasks of the building project are decomposed into multiple task units, each corresponding to a work breakdown structure code. For example, "WBS-STR-01" might represent "first-floor steel structure installation". These codes not only identify the subordinate relationships of tasks but also provide an index for the subsequent association of model components.

[0068] The time reference coordinate system is established based on the schedule data in the construction plan. In Primavera P6 or Microsoft Project, a unified time reference coordinate system can be established based on the project contract duration, construction process logic, and resource constraints. This coordinate system takes the planned start date as the origin, and the start and end times of all tasks are referenced to this. For example, "T+30 days" corresponds to a specific calendar date.

[0069] The critical path algorithm is used to determine the dependencies of each task unit, and the critical path feature chain is generated based on the dependencies.

[0070] The critical path algorithm can be a Clique Percolation Method (CPM) algorithm. By calculating the dependencies between tasks, the system automatically identifies the critical task chains affecting the overall project duration, forming a critical path feature chain, such as the sequence "pile foundation inspection → steel structure hoisting → curtain wall installation". Ultimately, these feature points (WBS coding, time base, critical path) are written into the metadata of the project plan, becoming the foundational data for subsequent lightweight model processing, 4D simulation, and schedule comparison. This entire process ensures a dynamic connection between the plan and the model, enabling traceability and real-time response capabilities in construction schedule management.

[0071] In some alternative embodiments, the way 3D BIM models are developed must ensure that each object (e.g., wall, slab, etc.) can be assigned to a different work package when necessary. Objects in the model must contain an Activity_ID attribute field. For efficient 4D sorting, an Activity_ID value must be entered for all elements. Based on its project schedule, a corresponding WBS value needs to be populated for each element and entered into the Activity_ID field; in the project schedule, this may be referred to as the WBS code or Activity ID. This value associates the element with a specific installation date or date range.

[0072] Therefore, taking the Revizto lightweighting process as an example, the steps for lightweighting a 3D building model based on the aforementioned baseline plan features may include:

[0073] The 3D architectural model undergoes preprocessing settings. During the lightweighting process based on baseline plan features, the entire process must balance data simplification with the preservation of key features. This ensures the lightweight model meets the performance requirements of real-time rendering while supporting subsequent 4D progress simulations. The first step in lightweighting is model data preprocessing, including coordinate system alignment and unit unification. All model files involved in integration must be forced to use a unified global coordinate system origin to avoid spatial misalignment issues in subsequent steps. Simultaneously, it's crucial to ensure consistent units across all models, such as converting them all to millimeters or meters to prevent scale distortion. This stage also requires checking the model's integrity and geometric errors, fixing common issues like broken or overlapping surfaces, laying the foundation for subsequent processing.

[0074] The 3D building model is adjusted for level of detail (LOD) based on the critical path feature chain. After preprocessing, the core step is geometric simplification. This step involves differentiated LOD adjustments based on the construction importance of components and the critical path identifiers in the baseline plan features. For repetitive elements such as standard doors and windows and pipe supports in non-critical path areas, their geometric representation is replaced from high-precision parametric families to simplified triangular meshes, retaining only the outer contour information. For components related to critical path tasks, such as steel structure nodes and equipment foundations, higher-level geometric details are retained, including critical construction features such as bolt holes and welds. Simultaneously, the system automatically identifies and optimizes duplicate components, merging identical family instances into lightweight shared references, significantly reducing the amount of geometric data.

[0075] The attribute fields of components in the 3D building model are filtered based on the Work Breakdown Structure (WBS) coding. In terms of attribute data processing, the lightweight engine intelligently filters the attribute fields of model components according to the WBS coding system in the baseline plan features. Core attributes directly related to construction progress, such as Activity_ID, WBS_Code, and Installation_Phase, are retained; these fields are crucial for binding components to schedule tasks in subsequent 4D simulations. Redundant data such as design notes, version history, and unused shared parameters from the design phase are removed. For critical path components, the system also additionally retains attributes required for construction control, such as material strength and installation techniques, to ensure the accuracy of the schedule simulation.

[0076] The high-resolution textures in the 3D architectural model are compressed, and similar materials in the model are merged into single material groups to obtain a lightweight 3D architectural model. Lossy compression of the high-resolution textures is performed, for example, downsampling 4K stone textures to 512x512 resolution, while generating multi-level mipmap textures and dynamically loading textures of different precisions based on the view distance. For material spheres, a merging strategy is used, merging materials with similar appearances but slightly different parameters (such as concrete with different labels) into single material groups, distinguishing specific types by vertex color or UV offset. This process significantly reduces video memory usage without noticeable loss of visual quality.

[0077] The final lightweight model output must strictly inherit the baseline plan's feature information. This is achieved by writing lightweight geometric data, simplified attributes, and compressed textures to each component, while ensuring that key feature fields such as WBS encoding and plan timestamps are fully preserved. The output format is typically an open standard NWD or GLTF file, which supports hierarchical data structures, facilitating subsequent model splitting by construction stage or professional system. The entire lightweighting process is automated through batch processing scripts, but a manual review process is triggered for components in critical path areas to ensure that the model accuracy of core construction nodes is not affected.

[0078] In some optional embodiments, when spatially reconstructing the lightweighted 3D building model based on baseline plan features, the decomposition process needs to be closely integrated with construction logic and schedule management requirements. To ensure that the generated component set can both meet the requirements of dynamic schedule simulation and support multi-disciplinary collaborative work, the specific implementation of spatial reconstruction in step S13 may include:

[0079] Obtain the geometric topological relationships and spatial distribution characteristics of the 3D building model, and establish a component attribute index table of the 3D building model based on the geometric topological relationships and spatial distribution characteristics.

[0080] The entire decomposition process begins with model parsing. The system reads the baseline plan feature data inherited from the lightweight model, including WBS coding, timestamps, and critical path identifiers, while simultaneously scanning the model's geometric topology and spatial distribution features. This stage requires establishing a complete component attribute index table, linking the geometric data of each component with its construction tasks, installation stages, and other metadata, providing a data foundation for subsequent rule-based decomposition.

[0081] Based on the construction method of the building project, a target splitting strategy is determined, and based on the target splitting strategy, the 3D building model is split into multiple levels of sub-models.

[0082] After model analysis is completed, the next step is to define the splitting rules. This step requires developing a multi-dimensional splitting strategy based on the specific construction organization of the project. Common splitting dimensions include: dividing by construction phase, such as splitting the model into "earthwork excavation phase model," "main structure phase model," and "decoration and finishing phase model"; dividing by professional system, such as dividing into "structural model," "mechanical and electrical model," and "curtain wall model," where the mechanical and electrical system can be further refined into "water supply and drainage system," "electrical system," and "heating, ventilation, and air conditioning system"; dividing by spatial area, such as splitting high-rise buildings into units of 5 floors, or dividing factory projects by axis; and dividing by task hierarchy in the schedule plan, so that each sub-model corresponds to a specific WBS coding range. In addition, these splitting rules can also be combined, for example, considering both construction phase and professional system simultaneously to generate a composite model unit of "main structure phase - steel structure subsystem."

[0083] For a first component that exists only in one construction phase, identify the attribute information of each first component in the component attribute index table, and group the first components into models based on the attribute information to obtain a sub-model corresponding to each first component.

[0084] When actually performing model splitting, components can be intelligently filtered and grouped based on preset rules. Taking splitting by WBS encoding as an example, by scanning the Activity_ID field of all first components, components with the same WBS prefix are automatically classified. For example, all structural components whose Activity_ID starts with "WBS-STR-" are classified into the structural model group.

[0085] For a second component that exists in multiple construction stages, the second component is cut according to the actual construction joint position based on a spatial classification algorithm to obtain a sub-model corresponding to each second component.

[0086] The second component is a composite component that spans multiple construction stages in complex situations. For the second component, spatial segmentation algorithms, such as mesh generation, quadtree, and octree, can be used to cut the component geometry according to the actual construction joint location.

[0087] Furthermore, after spatial reconstruction, the model packages constructed from all the sub-models can be optimized. The optimization steps may include:

[0088] Reorganize the geometric data of each of the sub-models; reorganization means instantiating and referencing duplicate components in the same sub-model to reduce storage space.

[0089] Establish a spatial index for each of the sub-models; spatial queries during subsequent rendering can be accelerated using an octree or BVH structure.

[0090] The Level of Detail (LOD) corresponding to each sub-model is adjusted based on its importance. For example, critical path components (such as main structural nodes and large equipment bases) can be automatically marked as high importance, and high importance sub-models can retain LOD400 level details, including construction process features such as bolt holes and welds. Standard floor beams, slabs, columns, and precast concrete wall panels of non-load-bearing partitions in the structural system can be of medium importance and can adopt LOD300 level, retaining the main geometric features but simplifying the connection nodes. Standard doors, windows, lighting fixtures, etc. can be simplified to LOD200 level details, retaining only the basic outline.

[0091] A metadata file is generated for each of the sub-models. The metadata file includes the work breakdown structure coding range, time window, and critical path tasks associated with the sub-model. The metadata file is used as an index for the sub-model corresponding to the metadata file when each sub-model is driven based on the plan arrangement.

[0092] The final output stage needs to ensure that the decomposed model retains complete engineering semantics. This is achieved by assigning a unique version identifier to each sub-model, establishing a traceability relationship with the original model, and checking the attribute inheritance of all components, particularly the integrity of baseline plan feature data. Furthermore, for very large projects, the decomposed model can be optimized for spatial partitioning and storage, managed using a spatial database or distributed file system, supporting concurrent access by multiple users.

[0093] In some optional embodiments, the implementation of binding the sub-model with the timeline and keyframes in step S14 may include:

[0094] A timeline is created for each sub-model; the timelines are organized according to the WBS hierarchy, forming a tree-like relationship similar to the planned tasks. Key parameters of the timelines may include a scaling factor representing the handling of planned compression or extension, and a cycle pattern representing the period defining repetitive construction tasks. For complex sub-models containing multiple construction phases, composite timelines can be configured; for example, a steel structure installation sub-model may contain three parallel sub-tracks: "factory prefabrication," "transportation," and "on-site hoisting." After configuration, a cross-reference table of timelines and sub-models is automatically generated, recording the time parameters associated with each sub-model.

[0095] The position of the sub-model's time window within the time track is determined, an initial keyframe is inserted at the start time of the time window, and the initial state of the sub-model is defined. The initial state of the sub-model can be a completely transparent state, indicating that the task corresponding to the sub-model has not yet started.

[0096] Insert one or more progress keyframes in the middle of the time window and define the progress state of the sub-model; wherein, the progress state of the sub-model can be semi-transparent, and when there are multiple progress states, different transparency can be set. Alternatively, other colors can be used to display the entity, and different brightness can be set when there are multiple progress states.

[0097] Insert an end keyframe at the end time of the time window and define the end state of the sub-model; wherein, the end state of the sub-model can be that the entity is displayed and the color is green, indicating that the task corresponding to the sub-model has been completed.

[0098] Furthermore, the state binding process transforms keyframe definitions into concrete visualization rules. A state machine is created for each sub-model component, including attributes such as display status (hidden / semi-transparent / solid), color coding (planned color / actual color / warning color), spatial location (for movable devices), and special effects (blinking / highlighting). Batch operations are supported, such as selecting all components with the same WBS coding to uniformly set state rules, and individual fine-tuning of specific components is also supported. For sub-models on the critical path, an additional warning mechanism is bound, automatically triggering state changes when actual progress deviates from the plan. A binding configuration file can be generated to record the state rules and parameter override relationships of each sub-model; these configurations can be exported as templates for reuse in similar projects.

[0099] Transition effects are generated based on the state transition requirements between adjacent keyframes.

[0100] In some embodiments, the implementation of step S15, which drives each of the sub-models based on the planned layout to obtain the 4D building model, may include the following steps:

[0101] Establish a mapping relationship between the sub-model of each component and the work breakdown structure code; wherein, in the planning data parsing stage, firstly extract structured data containing key fields such as WBS code, task name, plan start / completion time, and prerequisite relationships from the plan layout, and parse the baseline feature point data embedded in the 3D component library to establish a mapping relationship table between component ID and WBS code.

[0102] The target component group is located based on the mapping relationship, and the component states that should be activated in the target component group are filtered according to the time axis; wherein, the target component group is quickly located by WBS encoding index, and the component states that should be activated are filtered according to the current position of the time axis.

[0103] The component is dynamically displayed based on its status, and the component is rendered preferentially when a critical task time window is triggered.

[0104] Based on the same concept, embodiments of this application also provide a computer device, which may include 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 methods described above.

[0105] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A construction calibration method based on a 4D building model, characterized in that, include: Construct a 4D architectural model of the building project; wherein the 4D architectural model includes multiple component sub-models, each of which is bound to a corresponding timeline and keyframes. The timeline represents the time span of the construction plan of the component from start to finish, and the keyframes are task nodes that define the state changes of the component. For any construction phase of the building project, the construction site of the building project is scanned to obtain a 3D building model of the site. A deviation diagram is generated based on the comparison results between the 3D building model of the site and the planned state model. The planned state model is the state of the 4D model at the current construction phase. A construction calibration scheme for the building project under the current construction status is generated based on the deviation diagram.

2. The method according to claim 1, characterized in that, For any construction phase of the building project, the construction site is scanned to obtain a 3D building model. A deviation diagram is generated based on the comparison between the 3D building model and the planned state model, including: The construction site is scanned to obtain point cloud data of the construction site; The on-site 3D building model is constructed based on the point cloud data; The on-site 3D building model and the planned state model are spatially aligned and their coordinate systems are matched. The deviation value of each detection point is determined by comparing the on-site 3D building model with the planned state model based on the three-dimensional difference detection algorithm. The deviation diagram is generated based on the deviation value of each detection point.

3. The method according to claim 2, characterized in that, In the deviation diagram, the first color area represents the normal range that meets the design requirements, the second color area represents the range that exceeds the design requirements, and the third color area represents the range that does not meet the design requirements.

4. The method according to claim 1, characterized in that, The 4D architectural model of the construction project includes: The construction plan data of the building project is obtained, and a baseline plan feature and a plan layout are generated based on the construction plan data; wherein, the baseline plan feature includes a work breakdown structure code, a time base coordinate system and a critical path feature chain that characterize the building project, and the plan layout characterizes the execution logic of the building project; The 3D building model is lightweighted based on the aforementioned baseline plan features; Based on the characteristics of the baseline plan, the 3D building model after lightweighting is spatially reconstructed to obtain the sub-model corresponding to each component in the 3D building model. For any of the component's sub-models, the sub-models are bound to a timeline and keyframes respectively; wherein, the timeline represents the time span of the component's construction plan from start to finish, and the keyframes are task nodes that define the changes in the component's state; The 4D building model is obtained by driving each of the sub-models based on the planned layout.

5. The method according to claim 4, characterized in that, The process of acquiring construction plan data for building projects and generating baseline plan features and plan layouts based on the construction plan data includes: Based on the construction plan data, the engineering tasks of the construction project are decomposed into multiple task units, and each task unit corresponds to a work breakdown structure code. Establish the time reference coordinate system based on the construction period data in the construction plan data; The critical path algorithm is used to determine the dependencies of each task unit, and the critical path feature chain is generated based on the dependencies.

6. The method according to claim 4, characterized in that, The lightweighting process of the 3D building model based on the baseline plan features includes: The 3D building model is preprocessed and configured. The 3D building model is adjusted in level of detail based on the critical path feature chain. The attribute fields of the components in the 3D building model are filtered based on the work breakdown structure encoding. The high-resolution textures in the 3D building model are compressed, and similar materials in the 3D building model are merged into a single material group to obtain a lightweight 3D building model.

7. The method according to claim 4, characterized in that, The spatial reconstruction of the lightweighted 3D building model based on the baseline plan features yields sub-models corresponding to each component in the 3D building model, including: Obtain the geometric topological relationships and spatial distribution characteristics of the 3D building model, and establish a component attribute index table of the 3D building model based on the geometric topological relationships and spatial distribution characteristics; Based on the construction method of the building project, a target splitting strategy is determined, and based on the target splitting strategy, the 3D building model is split into multiple levels of sub-models; For a first component that exists only in one construction phase, identify the attribute information of each first component in the component attribute index table, and group the first components into models based on the attribute information to obtain a sub-model corresponding to each first component; For a second component that exists in multiple construction stages, the second component is cut according to the actual construction joint position based on a spatial classification algorithm to obtain a sub-model corresponding to each second component.

8. The method according to claim 4, characterized in that, For any of the aforementioned components, the sub-model is bound to a timeline and keyframes, respectively, including: Create a timeline for the sub-model; Determine the position of the sub-model's time window on the time track, insert an initial keyframe at the start time of the time window, and define the initial state of the sub-model; Insert one or more progress keyframes in the middle of the time window and define the progress state of the sub-model; Insert an end keyframe at the end time of the time window and define the end state of the sub-model; Transition effects are generated based on the state transition requirements between adjacent keyframes.

9. The method according to claim 4, characterized in that, The process of driving each sub-model based on the planned layout to obtain the 4D building model includes: Establish a mapping relationship between the sub-model of each component and the work breakdown structure code; The target component group is located based on the mapping relationship, and the component states that should be activated in the target component group are filtered according to the time axis. The component is dynamically displayed based on its status, and the component is rendered preferentially when a critical task time window is triggered.

10. A computer device, characterized in that, The computer device includes 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 method as claimed in any one of claims 1 to 9.

Images