Curved surface curtain wall keel construction method and system based on three-dimensional laser scanning

Abstract

The present application relates to the technical field of curved surface curtain wall construction, and particularly relates to a curved surface curtain wall keel construction method based on three-dimensional laser scanning, comprising the following steps: performing three-dimensional laser scanning on a completed structure entity, obtaining point cloud data and splicing to form complete point cloud; constructing a structure entity contour three-dimensional model based on the complete point cloud; constructing an appearance effect three-dimensional model based on the structure entity contour three-dimensional model, and performing spatial alignment; reversely deducing a curtain wall keel construction drawing that fits the completed structure entity based on the structure entity contour three-dimensional model and the appearance effect three-dimensional model; based on the curtain wall keel construction drawing, performing factory pre-fabrication, numbering and partition packaging of the keel and the veneer; and based on the curtain wall keel construction drawing, positioning and installing the keel and the veneer. The present application realizes digital pre-compensation and systematic elimination of construction error of the structure entity and welding error of the keel on site. The present application also discloses a corresponding system.

Description

Technical Field

[0001] This invention relates to the field of curved curtain wall construction technology, and in particular to a construction method and system for curved curtain wall keel based on three-dimensional laser scanning. Background Technology

[0002] In recent years, irregularly shaped curved curtain walls (especially hyperboloid aluminum panel curtain walls) have been widely used in large public buildings such as high-speed railway stations and airport terminals, significantly enhancing the artistic expression and contemporary feel of the buildings with their flowing shapes. Currently, the mainstream construction method for this type of curtain wall is a digital construction process based on Building Information Modeling (BIM): that is, a three-dimensional BIM model is created based on the design blueprints, the keel and decorative panels are prefabricated in the factory according to the model data, and then transported to the site for installation.

[0003] However, construction methods based on ideal BIM models have the following drawbacks: The construction methods based on the ideal BIM model ignore the objective dimensional and positional errors that exist after the main structure is constructed. During construction, the main building structure (such as concrete beams, columns, and eaves) inevitably exhibits deviations from the theoretical design model (typically reaching several centimeters). If curtain wall framing is directly produced based on the ideal BIM model, it often requires extensive on-site cutting, adjustments, or even remanufacturing due to incompatibility with the actual structure, leading to material waste, construction delays, and increased costs.

[0004] Furthermore, this method fails to effectively control secondary errors introduced during the on-site welding and installation of the keel. In traditional processes, the keel is mostly measured, cut, and welded on-site. The high-altitude working environment is harsh, measurement and positioning are difficult, and welding deformation is hard to control precisely, further accumulating installation errors and affecting the flatness and curvature of the final curtain wall surface. Summary of the Invention

[0005] To address at least one of the aforementioned technical problems, this invention proposes a construction method and system for curved curtain wall keel based on three-dimensional laser scanning. This method solves the technical challenges of traditional BIM model-based construction methods, which neglect on-site errors, resulting in inconsistent keel installation, extensive on-site cutting and welding, significant material waste, repeated rework, and distortion of complex curved surface shapes.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: The first aspect of this invention proposes a construction method for curved curtain wall keel based on three-dimensional laser scanning, comprising: S1, Perform 3D laser scanning on the completed structural entity to acquire point cloud data and stitch it together to form a complete point cloud; S2, Based on the complete point cloud, construct a three-dimensional model of the structural entity outline; S3, Based on the three-dimensional model of the structural entity outline, construct a three-dimensional model of the appearance effect, and spatially align the three-dimensional model of the appearance effect with the three-dimensional model of the structural entity outline; S4. Based on the three-dimensional model of the structural entity outline and the three-dimensional model of the appearance effect after spatial alignment, the construction drawings of the curtain wall keel that fit the completed structural entity are derived in reverse. S5. Based on the curtain wall keel construction drawings, the keel and decorative panels are prefabricated in the factory, numbered, and packaged in sections. S6. Based on the curtain wall keel construction drawings, position and install the keel and decorative panels.

[0007] Preferably, step S1 includes the following steps: S11, Confirm that there are no obstructions on the surface of the structure to be scanned that would affect the scanning accuracy; S12, Use a total station to set up and mark uniform construction survey benchmark points and baselines on site; S13, Based on the reference points and reference lines, a spatial reference control network consisting of multiple control targets with known coordinates is deployed in the scanning area; S14, Based on the spatial reference control network, plan multiple scanning stations with overlapping areas, so that adjacent scanning stations cover at least one of the control targets; S15, set up a three-dimensional laser scanner at each scanning station to scan the on-site structural entity and simultaneously collect point cloud data of the control target; S16, Based on the known coordinates of the control target, the point cloud data obtained by scanning at each station are stitched together to generate a complete point cloud.

[0008] Preferably, step S2 includes the following steps: S21, under the global construction coordinate system defined by the spatial reference control network, the complete point cloud is denoised and filtered to remove interference points and optimize data density, so as to obtain a clean point cloud dataset that characterizes the surface of the structural entity. S22, Analyze the clean point cloud dataset to identify and distinguish between regular geometric features and free-form surface features; S23, for the identified regular geometric feature elements, point cloud slices are generated by creating a cutting plane, two-dimensional contour lines are drawn based on the vectorization of the slice point cloud, and a parametric solid model is generated through extrusion and lofting three-dimensional operations; For the identified freeform surface features, the corresponding point cloud data is converted into a triangular mesh model, and surface fitting, editing and stitching are performed to generate a high-precision continuous surface model; S24 performs Boolean operations and splices the parametric solid model and the continuous surface model to form a three-dimensional model of the structural solid contour.

[0009] Preferably, the point cloud data includes the three-dimensional spatial coordinate information of the sampling points, laser reflection intensity information, and / or true color information.

[0010] Preferably, step S3 includes the following steps: S31, Obtain the original curved curtain wall design data and build or import a 3D model of the design appearance; S32, import the three-dimensional model of the design appearance effect into the global construction coordinate system defined by the spatial reference control network where the three-dimensional model of the structural entity outline is located; S33, select no fewer than three corresponding feature alignment points in the three-dimensional model of the structural entity outline, and perform spatial transformation on the three-dimensional model of the design appearance effect so that it is aligned with the three-dimensional model of the structural entity outline in terms of spatial position and orientation.

[0011] Preferably, the feature alignment points include the center point of the embedded part and / or the intersection of the structural axis and / or the control point or endpoint of the two-dimensional and / or three-dimensional feature curve used to define the curved shape of the curved curtain wall; the spatial transformation includes rotation and / or movement and / or scaling.

[0012] Preferably, step S4 includes the following steps: S41, Pre-set the layout rules of the curtain wall keel system, the layout rules including keel spacing, cross-sectional specifications and connection standards; S42, Based on the structural entity outline model, generate the spatial starting point of the keel according to the arrangement rules; S43, Based on the three-dimensional model of the design appearance effect, generate spatial target points or target surfaces for the keel; S44, by connecting the spatial starting point and the spatial target point, the centerline spatial data of each keel is calculated and generated; S45, Based on the centerline spatial data and cross-sectional specifications, generate a three-dimensional keel model; S46, The three-dimensional keel model is verified. Based on the verified three-dimensional keel model, the curtain wall keel construction drawings and component processing and cutting list are automatically generated. The three-dimensional keel model can be divided into several unit truss models to guide factory prefabrication.

[0013] Preferably, the verification of the three-dimensional keel model in step S46 includes at least one of the following verification methods: (1) Structural interference verification: In the three-dimensional design environment, collision detection analysis is performed on the three-dimensional keel model and the three-dimensional model of the structural entity outline to ensure that there is no physical interference and that there is no mutual interference between the components of the keel model; (2) Construction space verification: In the three-dimensional design environment, check the minimum distance between the outer contour surface of the three-dimensional keel model and the inner surface of the three-dimensional model of the design appearance effect to ensure that it is not less than the minimum construction space thickness required by the curtain wall system; (3) Design rule compliance verification: Check whether the arrangement spacing, cross-section selection and connection method of each component in the three-dimensional keel model are consistent with the arrangement rules of the preset curtain wall keel system; (4) Preliminary verification of mechanical properties: The three-dimensional keel model is imported into the structural analysis software, and a preset load is applied to conduct a preliminary stress analysis to verify whether its strength, stiffness and stability meet the design requirements.

[0014] Preferably, step S6 includes the following steps: S61. Based on the curtain wall keel construction drawings, using the benchmark points and benchmark lines as a joint benchmark, use a total station to measure and set out the keel installation control lines and control points. S62, according to the keel number, the prefabricated keel is hoisted to the designated area, and positioned, leveled and fixed according to the installation control line and control point; S63. According to the panel number and layout design drawing, the decorative panels are aligned and installed on the corrected keel. S64, sealing and surface cleaning of the decorative panel.

[0015] The second aspect of this invention proposes a curved curtain wall keel construction system based on three-dimensional laser scanning, used to implement the construction method described in the first aspect, comprising: The 3D laser scanning and point cloud fusion module is used to perform 3D laser scanning on completed structural entities, acquire point cloud data, and stitch them together to form a complete point cloud. The entity contour construction module is used to construct a three-dimensional model of the structural entity contour based on the complete point cloud. The spatial alignment module is used to construct a three-dimensional model of the appearance effect based on the three-dimensional model of the structural entity outline, and to spatially align the three-dimensional model of the appearance effect with the three-dimensional model of the structural entity outline. The reverse derivation module is used to reverse derive the curtain wall keel construction drawings that fit the completed structural entity based on the spatially aligned three-dimensional model of the structural entity outline and the three-dimensional model of the appearance effect. The prefabrication module is used to prefabricate, number, and package the keel and decorative panels in the factory based on the curtain wall keel construction drawings. The positioning and installation module is used to position and install the keel and decorative panels based on the curtain wall keel construction drawings.

[0016] Compared with the prior art, the present invention has the following beneficial effects: This invention, through scanning and measurement followed by reverse design, generates construction drawings that inherently include compensation values ​​for all structural errors. This enables precise "zero-cut" installation of prefabricated keels on-site, achieving digital pre-compensation and systematic elimination of structural construction errors and on-site welding errors of the keels. It solves the technical problems of traditional BIM-based construction methods, which neglect on-site errors, leading to inconsistencies in keel installation, excessive on-site cutting and welding, significant material waste, repeated rework, and distortion of complex curved surfaces. This method is particularly suitable for scenarios where traditional measurement and positioning are extremely difficult, such as cantilevered, large-span, and multi-curvature eaves, providing a standardized and replicable intelligent construction solution for solving the curtain wall construction challenges of such landmark buildings. Attached Figure Description

[0017] Figure 1 A flowchart illustrating a construction method for curved curtain wall keel based on 3D laser scanning; Figure 2 This is a framework diagram of a curved curtain wall keel construction system based on three-dimensional laser scanning; Detailed Implementation

[0018] To enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments of the present invention.

[0019] Example 1 Please refer to Figure 1 As shown, a construction method for curved curtain wall keel based on three-dimensional laser scanning includes: S1, Perform 3D laser scanning on the completed structural entity to acquire point cloud data and stitch it together to form a complete point cloud; S2, based on the complete point cloud, constructs a 3D model of the structural entity outline; S3: Based on the 3D model of the structural entity outline, construct the 3D model of the appearance effect, and spatially align the 3D model of the appearance effect with the 3D model of the structural entity outline. S4. Based on the three-dimensional model of the structural entity outline and the three-dimensional model of the appearance after spatial alignment, the curtain wall keel construction drawings that fit the completed structural entity are derived in reverse. S5, based on the curtain wall keel construction drawings, the keel and decorative panels are prefabricated in the factory, numbered, and packaged in sections; S6, based on the curtain wall keel construction drawings, positions and installs the keel and decorative panels.

[0020] This embodiment, through scanning and measurement followed by reverse design, generates construction drawings that inherently include compensation values ​​for all structural errors. This enables precise "zero-cut" installation of prefabricated keels on-site, achieving digital pre-compensation and systematic elimination of structural construction errors and on-site welding errors of the keels. It solves the technical problems of traditional BIM-based construction methods, which neglect on-site errors, leading to inconsistencies in keel installation, excessive on-site cutting and welding, significant material waste, repeated rework, and distortion of complex curved surfaces. This method is particularly suitable for scenarios where traditional measurement and positioning are extremely difficult, such as cantilevered, large-span, and multi-curvature eaves, providing a standardized and replicable intelligent construction solution for solving the curtain wall construction challenges of such landmark buildings.

[0021] In a preferred embodiment, S1 includes the following steps: S11, confirm that there are no obstructions on the surface of the structure to be scanned that would affect the scanning accuracy.

[0022] Specifically, 3D laser scanning is "what you see is what you get." Any obstructions (such as scaffolding, pipelines, and debris) will be scanned and become part of the point cloud, contaminating the data and interfering with subsequent modeling. Eliminating the influence of obstructions is a prerequisite for ensuring data quality. A "clean" scanning environment ensures that the acquired point cloud purely reflects the surface information of the target structural entity, forming the basis for building a high-fidelity entity contour model. The 3D laser scanner used in this embodiment is the Faruo S350.

[0023] S12, Use a total station to set up and mark uniform construction survey benchmark points and baselines on site.

[0024] Specifically, the benchmark point can be the intersection of the station building's designed center mileage and the designed vertical distance between the main elevation axis and the station's main line. Then, using a high-precision total station (such as the Zhongwei ZT20R), based on the permanent on-site benchmark stakes transferred and verified by the surveying unit, the benchmark point coordinates are accurately measured and marked on stable structures or dedicated surveying piers on site using polar coordinate methods or resection methods. The measurement process must comply with the "Engineering Surveying Standard" (GB50026-2020) and involve forward and backward measurements or multiple rounds of measurement for adjustment to ensure point accuracy is better than ±3mm.

[0025] The baseline can be established using the surveyed benchmark points as the origin. Utilizing the direction measurement function of a total station, the main construction baselines are laid out and marked according to the design axis direction. For example, clear ink lines or laser alignment lines are drawn along the axis of the station building's front facade and perpendicular to that axis, with permanent or semi-permanent markings made at key locations. These baselines, together with the benchmark points, constitute a two-dimensional plane control benchmark, providing direction and starting point for the subsequent establishment of a three-dimensional spatial control network.

[0026] Understandably, the reference point serves as the coordinate origin for scanner setup and subsequent point cloud stitching, ensuring that the scanned data is incorporated into the global construction coordinate system from the outset. Simultaneously, it also serves as the drawing origin for the 3D model of the structural entity outline and the curtain wall frame construction drawings. This "one-point-one-line" reference transfer mechanism ensures that all spatial information, from data acquisition and digital design to component installation, is based on the same coordinate reference, eliminating error accumulation caused by inconsistent references or multiple conversions. This provides a precise spatial foundation for subsequent reverse adaptation and accurate installation based on 3D laser scanning.

[0027] S13, based on reference points and reference lines, deploys a spatial reference control network consisting of multiple control targets with known coordinates in the scanning area.

[0028] Specifically, the control network is optimized based on the three-dimensional spatial shape, size, and complexity of the structure to be scanned (such as the curved eaves of a station building). The design principle is that control points (targets) should be evenly distributed around the scanning area and key feature parts to form a stable geometric shape (such as a closed traverse network, triangulation network, or edge-angle hybrid network). This ensures that any scanning station can observe a sufficient number (usually no less than 3) of control targets, and that the joint observation of targets by different stations can form a strong geometric intensity to constrain and control the overall deformation of the scanning data.

[0029] Select specialized cooperative targets suitable for 3D laser scanners. These targets typically have high reflectivity or specific geometric patterns (such as spherical targets, planar targets, checkerboard targets), enabling them to be quickly and accurately identified and centered by the scanner in point clouds.

[0030] Securely mount the selected control targets on a stable support within the scanning area (such as a dedicated tripod, permanent structure, or temporary support) to ensure absolute target stability without any movement or vibration throughout the scanning operation. The target placement height and orientation should avoid mutual obstruction and take into account the scanner's effective scanning distance and viewing angle.

[0031] Using the aforementioned unified construction survey benchmark as the coordinate origin and the baseline as the orientation basis, the three-dimensional spatial coordinates (X, Y, Z) of each control target center (or feature point) in the global construction coordinate system are measured one by one using surveying techniques such as polar coordinate method or forward intersection method. The measurement process must follow the requirements of high-level control surveying, conduct multiple rounds of observation, and eliminate observation errors through measurement adjustment calculations (such as the least squares method) to obtain the optimal estimated coordinates and accuracy assessment information of each control target.

[0032] It is understandable that all the control targets with measured coordinates together constitute a space reference control network. Physically, the control network is a spatial array with targets as nodes, and digitally, it is a set of spatial points with high-precision three-dimensional coordinates.

[0033] During subsequent 3D laser scanning, while acquiring the point cloud of the structural entity's surface, the scanner simultaneously captures the point cloud image of the control target in the scanner's instantaneous coordinate system. Since the global coordinates of the control target are known, by solving the spatial transformation parameters (rotation matrix and translation vector) between the scanner's coordinate system and the global coordinate system, the point cloud data obtained from each station's scan, which is in the instrument's local coordinate system, can be accurately transformed and unified to the global construction coordinate system. This enables seamless and high-precision stitching of multi-station point clouds, generating a complete structural entity point cloud model with unified coordinates.

[0034] S14, based on the spatial reference control network, plans multiple scanning stations with overlapping areas, so that adjacent scanning stations cover at least one control target.

[0035] Specifically, a detailed scanning requirements analysis is conducted by combining the three-dimensional shape, spatial scale, detailed features of the structural entity to be scanned (such as a complex curved eave), as well as the on-site environmental conditions (such as visibility and workspace limitations). The required point cloud resolution (point spacing), scanning accuracy, and areas requiring special attention (such as key connection nodes and areas with drastic changes in surface curvature) are determined.

[0036] Within the existing physical framework of the "space reference control network," the topology of the scanning stations is designed. Core design principles include: Complete coverage principle: The combination of scanning station positions in the design must ensure that point cloud data of all surfaces of the structural entity can be collected from all perspectives to avoid scanning blind spots.

[0037] Overlapping area guarantee principle: For any two adjacent or spatially close scanning stations, their scanning fields of view must have a sufficient overlap area. Furthermore, the overlapping area should not merely be a small overlap of edges, but should include a substantial surface with rich three-dimensional features (such as corners, edges, and curved surfaces) of a certain area. This ensures that point cloud data is stitchable at a "relative" level (based on surface features), enabling multi-source data fusion.

[0038] Control target co-view principle: During planning, it must be ensured that at least one control target is simultaneously located within the scanning field of view of adjacent scanning stations. The point cloud data obtained from scanning two adjacent stations contain images of the same (or several) control targets. The known global coordinates of this target provide the basis for direct and absolute coordinate alignment of the data from these two stations.

[0039] Simultaneously, for each planned scanning station, specific scanning parameters need to be further determined, including but not limited to: scanner mounting height, horizontal and tilt angles, scanning resolution settings, and scanning range (achieved by setting the scanning window or clipping box). The optimal working distance of the scanner must be considered during planning to ensure point cloud quality and avoid sparse or missing point clouds due to excessive incident angles.

[0040] The control target introduces an "absolute" constraint to point cloud stitching. In subsequent stitching processing, the software not only uses the point clouds of overlapping areas for matching, but also utilizes the known precise global coordinates of the common-view control target to correct and optimize the matching results, thereby directly and accurately placing the point cloud data of each station into the global construction coordinate system. Combining the dual advantages of feature matching (relative accuracy) and control point constraints (absolute accuracy), it can effectively resist the influence of different scanning distances, angles, and instrument errors, ensuring that the stitched "complete point cloud" has extremely high overall geometric fidelity and coordinate accuracy, providing a data authenticity foundation for subsequent reverse modeling.

[0041] S15. Three-dimensional laser scanners are set up at each scanning station to scan the on-site structural entities and simultaneously collect point cloud data of the control targets.

[0042] Specifically, according to the pre-established scanning plan, the operators sequentially set up and leveled the 3D laser scanner at the preset scanning sites, optimizing parameters such as scanning resolution, field of view, and laser intensity based on site conditions. After starting the scan, the instrument simultaneously acquires point cloud data of the structural entity surface and specific control targets in the spatial reference control network, ensuring that both are acquired "at the same station, from the same source, and synchronously" in the same local coordinate system.

[0043] This embodiment, through synchronous acquisition, embeds a physical reference (target point cloud) in each station's data that can be precisely linked to the global coordinate system. This provides real, reliable, and directly solvable raw observation data for subsequent high-precision point cloud stitching based on absolute control. This absolute orientation method based on physical targets has higher global accuracy, stronger anti-interference ability, and more reliable error detection capability than simply relying on the features of overlapping areas in the point clouds for relative stitching.

[0044] S16: Based on the known coordinates of the control target, the point cloud data obtained from scanning at each station are stitched together to generate a complete point cloud.

[0045] Specifically, the raw point cloud data files acquired from each scanning station are imported into dedicated point cloud processing software (such as Faraday, Leica Cyclone, or Autodesk ReCap). The software automatically or semi-automatically identifies and extracts the point cloud clusters of control targets from the data of each station, and accurately calculates the center or feature point coordinates of each target in the local coordinate system of that station using fitting algorithms (such as sphere fitting, plane fitting, or bullseye recognition).

[0046] The processing software calls the global coordinate database of the control targets, which was pre-measured and entered in step S13. For the data of each scanning station, the software establishes the following mathematical model: finding an optimal spatial rigid body transformation (usually including three translation parameters and three rotation parameters) such that after the transformation, the overall deviation between the fitted coordinates of all identified control targets in the station's data and their corresponding known global coordinates is minimized (usually solved using the least squares method). This process essentially uses the resection principle to accurately calculate the transformation parameters from the local coordinate system of the scanning station to the global construction coordinate system.

[0047] Based on the solved transformation parameters, a coordinate transformation is performed on all point cloud data (including structural entity point clouds) collected by the scanning station to convert them to the global construction coordinate system. After performing the above orientation and transformation operations on all scanning stations in sequence, all station point cloud data have a unified spatial reference benchmark.

[0048] Based on absolute orientation, a fine-grained registration algorithm (such as the Iterative Closest Point (ICP) algorithm) is executed to fine-tune and optimize the absolute orientation results by utilizing the rich surface features of the overlapping area of ​​the point cloud between adjacent scanning stations. This eliminates residual system errors and achieves a high degree of consistency of the point cloud in the overlapping area.

[0049] Finally, the converted multi-site point cloud data is fused to remove redundant points that may have been generated from scanning from different perspectives, resulting in a complete point cloud 3D dataset with a consistent structure, continuous details, and unified coordinates. This complete point cloud accurately reflects the actual 3D shape and spatial location of the completed structural entity.

[0050] In a preferred embodiment, S2 includes the following steps: S21. Under the global construction coordinate system defined by the spatial reference control network, the complete point cloud is denoised and filtered to remove interference points and optimize data density, resulting in a clean point cloud dataset representing the surface of the structural entity.

[0051] It is understandable that during the acquisition of a complete point cloud, environmental factors (such as moving people, vehicles, and floating objects), internal instrument errors, or multiple reflections of the light beam can cause discrete noise points that do not represent the surface of the entity and outliers far from the main point cloud cluster. These can be addressed using statistical filtering algorithms or radius filtering algorithms. For example, by calculating the average distance and standard deviation between each point and its K neighboring points, points whose distance exceeds a preset standard deviation threshold can be identified as noise and deleted; or by setting a search radius, isolated points within the radius with fewer than a threshold number of neighboring points can be deleted to improve the signal-to-noise ratio of the data.

[0052] Meanwhile, point cloud data of non-target objects (such as construction machinery, temporary facilities, vegetation, etc.) that may exist in the scanning field of view can interfere with subsequent entity contour modeling. During processing, the three-dimensional spatial coordinates, reflection intensity, and RGB color information of the point cloud in this embodiment can be combined with manual selection, automatic region growing and segmentation, or feature-based classification methods to identify and remove the point cloud data of these irrelevant regions, ensuring that the dataset focuses on the target structural entity itself.

[0053] It should be noted that the density of the original point cloud may be unevenly distributed due to differences in scanning distance and angle. To facilitate subsequent modeling and control the amount of data, this embodiment employs a uniform sampling algorithm or a voxelized mesh filtering algorithm. For example, the voxelized mesh method divides the 3D space into tiny cubes (voxels) of equal volume, retaining only one representative point (such as the centroid or center point of all points) within each voxel. In this way, while maintaining the main geometric features, the total number of points is significantly reduced, achieving data lightweighting, while making the point cloud more uniformly distributed in space, avoiding data redundancy and computational burden during modeling.

[0054] S22, analyze the clean point cloud dataset to identify and distinguish between regular geometric features and freeform surface features.

[0055] It should be noted that intelligent geometric analysis is performed on the clean point cloud dataset. By calculating geometric feature parameters such as the local curvature, normal vector distribution, and neighborhood topological relationships of the point cloud, and combining them with preset regular feature templates (such as planes, cylinders, spheres, etc.), the algorithm automatically identifies and distinguishes regular geometric feature elements (such as straight beams and columns of buildings, regular opening edges, etc.) and free-form surface feature elements (such as hyperbolic eaves, gradually twisted surfaces, etc.) contained in the point cloud.

[0056] Specifically, for the identified regular geometric features, point cloud slices are generated by creating cutting planes, and two-dimensional contour lines are drawn based on the vectorized point cloud slices. According to the principle of parametric modeling, three-dimensional operations such as stretching, rotating, and lofting are applied to the two-dimensional contour lines to generate parametric solid models (such as cubes, cylinders, extruded bodies, etc.) with dimensional constraints and geometric relationships. This parametric solid model is clearly defined, easy to modify, and its edge cloud engineering quantities are calculated.

[0057] For the identified freeform surface features, the point cloud of the corresponding region is directly converted into a triangular mesh model (i.e., approximating the surface with numerous triangular patches). A surface fitting algorithm (such as NURBS surface fitting) is applied to reconstruct a smooth, continuous high-order mathematical surface based on the triangular mesh. This process includes surface editing (adjusting control points to optimize the shape) and surface stitching (smoothly connecting multiple surface patches into a whole), generating a high-precision continuous surface model.

[0058] S24 performs Boolean operations and splices the parametric solid model and the continuous surface model to form a three-dimensional model of the structural solid contour.

[0059] It should be noted that the various parametric solid models (such as beams, columns, and walls) and continuous curved surface models (such as hyperbolic eaves) generated above are imported into the same 3D design environment (such as Rhinoceros, AutoCAD Civil 3D, or professional reverse engineering software). By performing Boolean operations (including union, difference, and intersection) and geometric splicing operations, the spatial boundary relationships between various models are precisely handled. For example, a curved eaves model can be seamlessly spliced ​​with a regular wall model below, or pipe openings can be precisely created on the wall model (difference operation). This ensures that components generated by different modeling methods are geometrically continuous and topologically correct at shared boundaries, integrating into a complete, interference-free, and watertight 3D structural solid outline model.

[0060] To improve the accuracy of the 3D model of the structural entity outline, this implementation compares the integrated 3D model of the structural entity outline with the original clean point cloud dataset under the same global construction coordinate system.

[0061] Specifically, a 3D deviation analysis technique is used to calculate the minimum spatial normal distance from each sampling point in the point cloud dataset to the surface of the 3D model. Based on the distance data of all sampling points, a color deviation chromatogram is generated (usually green to represent zero deviation or within tolerance range, and red / blue gradient to represent positive / negative deviations), thus intuitively and globally displaying the fit between the model and the point cloud.

[0062] Meanwhile, quantitative accuracy evaluation software is used to output key statistical indicators, mainly including: root mean square error (RMSE), used to comprehensively evaluate the overall fitting accuracy; maximum positive / negative deviation, used to locate extreme error positions; and the percentage of point clouds within a preset tolerance range (for example, requiring more than 95% of points to have a deviation within ±5mm). These quantitative indicators provide objective acceptance criteria for whether the model accuracy meets the requirements of subsequent engineering applications.

[0063] Furthermore, if the deviation analysis results show that the error in a local area exceeds the limit, the process returns to the modeling stage (S23) to fine-tune and optimize the surface control points or parametric dimensions of the corresponding area, and then performs integration and deviation analysis again. This iterative process continues until the overall accuracy of the model meets the preset threshold requirements.

[0064] In a preferred embodiment, S3 includes the following steps: S31: Obtain the original curved curtain wall design data and build or import a 3D model of the design appearance.

[0065] It should be noted that the original curved curtain wall design data includes, but is not limited to: construction drawings of the curtain wall system, detailed node drawings, 3D BIM model files (such as .rvt, .ifc, .skp formats), and mathematical parameters or geometric definition files describing the curved surface shape (such as NURBS curve and surface data).

[0066] If the designer provides a complete, ready-to-use 3D BIM model or digital model, it can be directly imported into the modeling and analysis software environment (such as Rhinoceros, AutoCAD, etc.) used in this method through a data interface.

[0067] If the design data is mainly based on two-dimensional drawings or the provided three-dimensional model does not meet the requirements, it is necessary to reconstruct it in the software based on the design parameters. According to the geometric definition of the curved curtain wall (such as lofting curves, control points, and generation logic), use three-dimensional modeling tools to create the appearance curved surface, panel joints, and necessary three-dimensional details that reflect the design intent.

[0068] Understandably, the models obtained through either of the above methods should be continuous, closed, or three-dimensional geometries (or shells) with clearly defined inner and outer surfaces. Simultaneously, they should accurately represent the forming target of the outer surface (visible surface) of the curtain wall cladding panels. Their geometric accuracy and level of detail should at least meet the requirements for reverse positioning of the keel system. The resulting three-dimensional model of the appearance transforms abstract design drawings into a digital entity that can be used by computers for precise spatial calculations and collision analysis.

[0069] S32 imports the 3D model of the design appearance into the global construction coordinate system defined by the spatial reference control network, where the 3D model of the structural entity outline is located.

[0070] Specifically, in 3D modeling or point cloud processing software, a 3D model of the structural entity outline, already in the global construction coordinate system, is loaded as a reference background. Then, the 3D model of the design appearance is imported as a new object into the same software scene. Based on the confirmed theoretical alignment benchmark, an initial spatial transformation is applied to the design model using the software's coordinate transformation functions (including translation, rotation, and scaling if necessary). This ensures that the key control elements are initially aligned in space with their corresponding theoretical positions in the global construction coordinate system (these positions may be represented by reference points or lines).

[0071] S33. Select no fewer than three corresponding feature alignment points in the 3D model of the structural entity outline, and perform spatial transformation on the 3D model of the design appearance effect so that it is aligned with the 3D model of the structural entity outline in terms of spatial position and orientation.

[0072] It should be noted that the feature alignment points preferably include one or more combinations of the following types: Embedded component center point: The measured center point of the pre-embedded connectors, anchor plates, and post-installed embedded plates in the on-site structure, which has a corresponding theoretical installation position in the design model. This point reflects the absolute positioning reference.

[0073] Structural axis intersections: such as the intersections of column grid axes and main facade control axes at identifiable features on the site structure (obtained by fitting a point cloud to a plane), corresponding to the theoretical axis intersections in the design model. This point reflects the overall orientation and layout benchmark.

[0074] Control points or endpoints of two-dimensional and / or three-dimensional feature curves used to define the curved shape of a curved curtain wall: For complex curved curtain walls, the endpoints or control vertices (for B-spline / NURBS curves) of key feature curves (such as boundary lines, ridge lines, and generation generatrices) used to generate or control the main curved surface are selected in the design model. Then, through point cloud analysis or geometric fitting, the corresponding physical locations of these curves on the actual structure are found on the three-dimensional model of the structural entity outline. These points are directly related to the geometric definition of the curved surface shape and are crucial for achieving accurate matching of complex forms.

[0075] Spatial alignment mainly includes the following steps: First, in the 3D software, simultaneously display the 3D model of the structural entity outline (as a reference) and the 3D model of the design appearance effect that has been initially imported and aligned. Then, locate and mark the selected feature alignment points on the 3D model of the structural entity outline. Finally, find the theoretical position points on the 3D model of the design appearance effect that correspond one-to-one with the aforementioned feature points.

[0076] Then, the software's best-fit alignment or constraint-based coordinate transformation function is invoked. Point-to-point constraints are established using the feature alignment points on the structural model as "target points" and the corresponding theoretical points on the design model as "source points." Through optimization algorithms such as least squares, an optimal spatial transformation matrix (including 3D translation and rotation, and uniform scaling if necessary) is automatically calculated to minimize the overall distance error between all source points on the design model and the target points on the structural model after transformation.

[0077] Finally, the transformation matrix is ​​applied to move and / or rotate and / or scale the 3D model of the design appearance effect as a whole, so as to complete the precise positioning and attitude calibration of the 3D model of the appearance effect in the global construction coordinate system.

[0078] This step involves selecting feature alignment points that possess information on absolute position (such as embedded parts), macroscopic orientation (such as axes), and microscopic morphology (such as characteristic curves) to achieve multi-level, high-precision spatial matching between the 3D model of the appearance effect and the 3D model of the structural entity's outline, from the overall to the local, and from flat to curved surfaces. This ensures that the reverse-derived curtain wall keel system can adapt to the actual positional deviations of the structural entity while strictly adhering to the design intent of the curved surfaces. It is the spatial registration step that achieves the goal of "reproducing the design based on reality" in reverse adaptation.

[0079] In a preferred embodiment, S4 includes the following steps: S41, Pre-defined layout rules for the curtain wall keel system, including keel spacing, cross-sectional specifications and connection standards.

[0080] Specifically, this step aims to transform the general design requirements and engineering specifications of the curtain wall keel system into computer-recognizable parametric layout rules. These layout rules constitute a rule base, which includes at least: Keel spacing: The theoretical center distance between the main keel and the secondary keel in the horizontal and vertical directions. Gradient or variable spacing rules can be set for different areas.

[0081] Cross-section specifications: Standard cross-section shape, dimensions and material parameters of different types of keels (such as main keel, secondary keel and transition keel).

[0082] Connection standards: Standard connection methods, node construction and tolerance requirements between keels and between keels and the main structure (such as with embedded parts).

[0083] This rule base provides design basis and constraints for the subsequent automated and standardized generation of keel models.

[0084] S42, based on the structural entity contour model, generates the spatial starting point of the keel according to the layout rules.

[0085] This step addresses the issue of anchoring the keel to the main structure. Specifically, based on the aligned 3D model of the structural entity outline and combined with preset connection standards, the system automatically or semi-automatically identifies connection features on the structural surface that can be used to fix the keel. These locations are typically the centers of embedded parts, post-installed anchor points, or load-bearing areas approved by the structural design.

[0086] Based on the preset keel spacing rules and using the connection feature positions as a reference, a series of three-dimensional spatial points are generated on the surface of the structural entity model through algorithms such as equidistant projection, extension, or arraying, serving as the spatial starting points of the keels. Each starting point corresponds to a fixed position at the structural end of a keel, and its coordinates accurately reflect the actual three-dimensional shape and construction errors of the structural surface.

[0087] S43 generates spatial target points or target surfaces for the keel based on the 3D model of the design appearance.

[0088] It should be noted that this step solves the problem of positioning the inner surface of the decorative panel that the keel needs to support. Based on the aligned 3D model of the design appearance, this model accurately represents the outer surface (visible surface) of the curtain wall decorative panel.

[0089] Based on the structural layers of the curtain wall system (such as panel thickness, insulation layer, and air layer), the inner surface of the decorative panel (i.e. the supporting surface of the keel) is calculated from the outer surface of the design appearance model using an inward normal offset algorithm.

[0090] Based on the preset keel spacing rules and the grid layout of the decorative panel, a series of three-dimensional spatial points are generated on the inner surface. These points are defined as the spatial target points of the keel; or, the inner surface itself is directly defined as the spatial target surface of the keel. Each target point or corresponding point on the target surface represents the theoretical end position that a keel needs to reach to support the decorative panel.

[0091] Understandably, steps S41 to S43 collectively construct the parametric framework and spatial boundary conditions for the reverse design of the keel. The problem of locating each keel in space is transformed into a geometric problem of finding the optimal connection path between its starting point and its target point / surface. This provides a unique and precise input condition for calculating the spatial data of the keel centerline, and is a crucial transformation step from "dual-model alignment" to "precise positioning of a single keel."

[0092] S44 calculates and generates the centerline spatial data of each keel by connecting the spatial starting point with the spatial target point or target surface.

[0093] Specifically, for the "point-to-point" scenario (starting point corresponds to target point): theoretically, the straight line segment in three-dimensional space connecting the two corresponding points is the simplest centerline of the keel. However, in engineering practice, it may be necessary to consider structural avoidance of the keel (such as avoiding other pipelines), standardized length, or angle limitations. Therefore, the algorithm may introduce optimization conditions, such as ensuring that the centerline is at a specific angle (such as perpendicular) to the structural surface or target surface, or using segmented polylines for connection in complex areas.

[0094] For "point-to-surface" scenarios (starting point corresponds to the target surface): the algorithm needs to determine the optimal connection path from the starting point to the target surface. Specifically, this is achieved by calculating the shortest normal distance from the starting point to the target surface, i.e., finding the point on the target surface closest to the starting point (the perpendicular point), and then using the straight line segment connecting the starting point and the perpendicular point as the keel centerline. This method ensures that the keel supports the decorative panel in the most direct way, with a clear force path. For complex curved surfaces, the algorithm considers the local normals of the surface to ensure the rationality of the connection.

[0095] It should be noted that during the calculation process, the algorithm strictly follows the preset layout rules, such as checking the keel spacing, ensuring that the center line does not interfere with the main structure or other keel models (preliminary collision detection can be performed), and meeting the preset connection angle requirements.

[0096] Finally, the algorithm outputs the three-dimensional centerline data for each keel. This data is typically represented as a three-dimensional vector, line segment, or spatially parametric curve (for curved keels), containing complete spatial coordinate information of the centerline in the global construction coordinate system.

[0097] This step completes the crucial transformation from "control points" to "positioning lines." Specifically, through intelligent spatial geometric calculations, the structural fixing points and the support requirements of the decorative panels are concretized into the precise orientation and length of each keel in three-dimensional space. These centerline spatial data constitute the spatial skeleton of the keel system. This ensures that the reverse-designed keel system is geometrically adaptable to the actual conditions of the site structure (through the starting point) and accurately achieves the curved surface shape of the designed appearance (through the target points / surfaces).

[0098] S45 generates a three-dimensional keel model based on centerline spatial data and cross-sectional specifications.

[0099] Specifically, the construction of the 3D keel model can employ a core algorithm of path-scanning modeling. Specifically, a specified 2D cross-sectional profile (defined according to cross-sectional specifications) is used as the "profile graphic," and the centerline (3D line segment or curve) obtained in step S44 is used as the "scanning path." The 2D cross-sectional profile is translated and rotated along its associated centerline path. During the movement, the normal of the cross-sectional profile typically maintains a specific preset relationship (e.g., perpendicular) with the tangent direction of the path, thereby "sweeping" out a continuous 3D solid volume. For a straight centerline, a straight profile solid is generated; for a curved centerline, a corresponding curved profile solid is generated.

[0100] During the generation process, the software automatically processes the sealing of the cross section at the beginning and end of the path, and generates the necessary three-dimensional geometry of the end connectors (such as connecting plates and bolt holes) according to the connection standards, and merges them with the main keel entity through Boolean operations to form a complete three-dimensional keel model.

[0101] S46 verifies the 3D keel model. Based on the verified 3D keel model, automatically generate curtain wall keel construction drawings and component processing and cutting lists. The 3D keel model can be divided into several unit truss models to guide factory prefabrication.

[0102] Specifically, the above verification of the 3D keel model includes at least one of the following verification methods: (1) Structural interference verification: In a 3D design environment (such as Navisworks, Revit, or professional collision detection software), collision detection analysis is performed on the 3D keel model and the 3D outline model of the structural entity to ensure that there is no physical interference and that there is no mutual interference between the components of the keel model. This eliminates physical interference and avoids rework and modifications caused by inability to install on site.

[0103] (2) Construction space verification: In the three-dimensional design environment, check the minimum distance between the outer contour surface of the three-dimensional keel model and the inner surface of the three-dimensional model of the design appearance effect to ensure that it is not less than the minimum construction space thickness required by the curtain wall system, so as to ensure that the complete physical structure of the curtain wall system can be realized and to prevent the keel from "hitting" the decorative panel, which would prevent construction or affect the performance.

[0104] (3) Design rule compliance verification: Check whether the arrangement spacing, cross-section selection and connection method of each component in the three-dimensional keel model are consistent with the arrangement rules of the preset curtain wall keel system. This ensures the standardization and consistency of the design, meets the requirements of the specifications, and facilitates large-scale prefabrication.

[0105] (4) Preliminary verification of mechanical properties: Import the three-dimensional keel model into structural analysis software (such as SAP2000, Midas or finite element analysis software), apply preset loads to conduct preliminary stress analysis, and verify whether its strength, stiffness and stability meet the design requirements. Among them, the preset loads include but are not limited to: self-weight, wind load (positive and negative wind pressure), snow load, seismic action and temperature load.

[0106] In a preferred embodiment, S6 includes the following steps: S61, based on the curtain wall keel construction drawings, uses the benchmark points and benchmark lines as a joint benchmark, and uses a total station to measure and set the keel installation control lines and control points.

[0107] Specifically, surveyors use a total station, with the unified construction surveying benchmarks and baselines established and marked by S12 as the station and backsight benchmarks, to set up the station and orient. This ensures that the on-site layout coordinate system is completely consistent with the global construction coordinate system on which the design model and scanned point cloud depend, thus realizing the transfer from virtual coordinates to the actual location.

[0108] Under instrument control, spatial points corresponding to the design coordinates are measured and established on the corresponding structural surfaces or spatial locations on site using either polar coordinates or free stationing methods. For each measured point, clear, indelible markings (such as ink lines, punch marks, or adhesive targets) are used for on-site identification. Control points typically include: the theoretical installation center point of each unit truss or main keel end, axis control points, elevation control points, and key points of the outline boundary.

[0109] Simultaneously, based on the keel layout direction, the main keel installation control lines (such as center lines and edge lines) are measured and marked out. These control lines and control points together form a three-dimensional spatial control network covering the installation area, serving as a direct visual and measurement reference for subsequent hoisting and positioning.

[0110] S62, according to the keel number, the prefabricated keel is hoisted to the designated area, and positioned, leveled and fixed according to the installation control line and control point.

[0111] Specifically, based on the unique serial number of each keel unit (such as a truss panel) and its associated component information sheet, the corresponding prefabricated component to be installed is accurately identified in the yard. Suitable hoisting equipment (such as a 25T truck crane or tower crane) is used to smoothly transport the component to the vicinity of the installation area. Protective measures must be taken during the hoisting process to prevent component deformation or coating damage.

[0112] During the hoisting process, the installation personnel guide the components into initial position based on the installation control lines and three-dimensional control points measured and marked on site according to S61. For example, aligning the end center mark of the truss keel with the control point, or aligning the edge line of the component with the pop-up control ink line.

[0113] By utilizing the correspondence between "number-drawing-control point", alignment can be achieved by "following the map", significantly reducing the complex on-site measurement and trial-and-error process.

[0114] After initial positioning, a high-precision level, laser level, or total station is used for fine-tuning. The three-dimensional coordinates, levelness, verticality, and elevation of key locations of the component (such as both ends and the midpoint) are measured and compared with the requirements of the design drawings.

[0115] By using fine-tuning devices (such as jacks, adjustable supports, and fine-tuning bolts), the position and orientation of the components are adjusted with micron-level precision until all installation deviations meet the allowable error range of the design and specifications (such as elevation ±2mm, axis position ±3mm).

[0116] After the calibration is completed, temporary fixing measures (such as positioning welding, special clamps, temporary bolts) are used to stabilize the components on the main structure to prevent displacement during subsequent operations or under the influence of external forces.

[0117] After verification that all keel units have been accurately aligned and temporarily fixed according to the design positions, the final fixing work is carried out according to the construction drawings and connection standards. This usually includes, but is not limited to: structural welding (the weld grade must meet the requirements of S7.3), installing high-strength bolts and tightening them to the specified torque, or completing other permanent mechanical connections specified in the design.

[0118] This step optimizes the traditional keel installation process, which relies heavily on on-site measurement, layout, cutting, and welding, into a standardized assembly process of "identification-hoisting-alignment-correction-fixing." By addressing complex spatial positioning issues in advance during the digital design and factory prefabrication stages, and solidifying the results in component numbering and the on-site control network, on-site installation operations are simplified, efficient, and precision-controlled. This not only significantly improves installation efficiency and quality stability, reduces the risks of working at heights and reliance on skilled workers, but also ensures that the final keel system can achieve the adaptation intent of reverse engineering.

[0119] S63. According to the panel number and layout design drawing, the decorative panels are aligned and installed on the corrected keel.

[0120] Specifically, specialized lifting tools or vacuum suction cups are used to safely and smoothly transport the decorative panels to the installation location. Installers then perform preliminary alignment of the decorative panels based on the spatial relative positions marked on the layout design drawings and the positioning aids already set on the keel (such as the panel boundary lines transferred via S61 layout).

[0121] Initial connection or attachment is made using the pre-installed mounting system (such as corner brackets or hangers) on the decorative panel and the corresponding connection points on the keel. This process allows for minor adjustments within a small degree of freedom to avoid stress or deformation caused by forced fixing.

[0122] After the decorative panels are initially attached, their spatial orientation is finely adjusted using three-dimensional measurement tools (such as total station, 3D laser scanner or portable 3D photogrammetry system).

[0123] The actual three-dimensional coordinates of several feature control points (such as corner points, center points, or pre-labeled points) on the surface of the decorative panel are measured and compared in real time with the theoretical coordinates on the layout design drawing.

[0124] Based on the deviation data from the measurement feedback, the position, flatness, and joint width of the decorative panel are finely adjusted by the three-dimensional adjustable mechanism on the fine-tuning bracket (which usually has adjustment functions in three dimensions: front and back, left and right, and up and down) until its spatial position, the uniformity of the joints between adjacent panels, and the smoothness of the overall curved surface all meet the design accuracy requirements (such as surface flatness ±1.5mm and joint width difference ±1mm).

[0125] After fine-tuning to meet the standards and verifying that there are no errors, according to the design requirements, all adjustment mechanisms of the connecting brackets are finally locked (such as tightening the anti-loosening nuts and applying shear pins), and necessary secondary connections are completed (such as welding and bolt tightening) to ensure that the decorative panel and the keel system form a stable connection.

[0126] S64, sealing and surface cleaning of the decorative panel.

[0127] Specifically, insert Φ10mm foam strips into the gaps of the decorative panel, then apply weather-resistant adhesive evenly, and finally remove and clean the protective film from the aluminum decorative panel packaging to complete the installation of the aluminum decorative panel.

[0128] Example 2 Please refer to Figure 2 As shown, a curved curtain wall keel construction system based on three-dimensional laser scanning is used to implement the construction method as described in Embodiment 1, including: The 3D laser scanning and point cloud fusion module 101 is used to perform 3D laser scanning on the completed structural entity, acquire point cloud data, and stitch them together to form a complete point cloud. The entity contour construction module 102 is used to construct a three-dimensional model of the structural entity contour based on the complete point cloud; The spatial alignment module 103 is used to construct a three-dimensional model of appearance effect based on the three-dimensional model of structural entity outline, and to spatially align the three-dimensional model of appearance effect with the three-dimensional model of structural entity outline. The reverse derivation module 104 is used to reverse derive the curtain wall keel construction drawings that fit the completed structural entity based on the spatially aligned 3D model of the structural entity outline and the 3D model of the appearance effect. Prefabricated module 105 is used for the factory prefabrication, numbering, and zoning and packaging of the keel and decorative panels based on the curtain wall keel construction drawings; The positioning and installation module 106 is used to position and install the keel and decorative panels based on the curtain wall keel construction drawings.

[0129] The construction method provided by this invention has been successfully applied in the exterior curtain wall project of the newly built Shangyu South Station on the Hangzhou-Shaoxing-Taizhou Railway. The station building's front and rear facades feature a double-curved aluminum panel curtain wall system, with an overall three-tiered stepped layout to embody the architectural image of a new-era gateway. The curtain wall's longitudinal cantilever width varies from 1.2 to 8.45 meters, with a total transverse length of 148.8 meters and a maximum installation height of 23.5 meters from the ±0.000 elevation of the drop-off platform. To achieve the complex double-curved design effect of the eaves, the entire project employed precise measurement and reverse construction techniques based on three-dimensional laser scanning technology.

[0130] Currently, the mainstream construction technology for irregularly curved aluminum panel curtain walls in domestic railway passenger stations still relies on ideal BIM models for component cutting and on-site installation. This traditional method does not fully consider the actual construction errors of the on-site structural entity and the secondary errors generated during on-site welding of the keel, which easily leads to discrepancies between the components and the actual site, resulting in material waste and rework. In contrast, the construction method adopted by Shangyu South Station uses 3D laser scanning to realistically capture the on-site structural morphology and performs reverse adaptation design based on a digital model that accurately reflects the errors. This systematically eliminates the impact of the above two types of errors, significantly reduces material waste and labor rework, and achieves significant economic benefits.

[0131] The intelligent scanning equipment used in this method can efficiently collect on-site physical data and upload it to the cloud, significantly reducing the workload of traditional manual measurement and data processing, and saving labor costs. By integrating a design appearance model with a 3D model of the structural entity outline that accurately reflects the on-site conditions, precise dimensions for keel processing and installation are derived through reverse engineering, ultimately generating construction drawings that perfectly match the physical structure. The construction method provided by this invention can be widely applied to the keel processing of various dry-hanging curtain walls, and is especially suitable for areas with complex shapes and difficult positioning measurements, such as eaves.

[0132] This method is applicable to the construction of irregular curved exterior curtain walls of public buildings such as railway passenger stations. It can effectively overcome the shortcomings of traditional processes and minimize the waste of materials and labor caused by structural and welding errors. It has universal applicability value in improving the construction accuracy, efficiency and economy of irregular curved curtain walls.

[0133] The above description is a specific implementation of the embodiments of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A construction method for curved curtain wall keel based on three-dimensional laser scanning, characterized in that, include: S1, Perform 3D laser scanning on the completed structural entity to acquire point cloud data and stitch it together to form a complete point cloud; S2, Based on the complete point cloud, construct a three-dimensional model of the structural entity outline; S3, Based on the three-dimensional model of the structural entity outline, construct a three-dimensional model of the appearance effect, and spatially align the three-dimensional model of the appearance effect with the three-dimensional model of the structural entity outline; S4. Based on the three-dimensional model of the structural entity outline and the three-dimensional model of the appearance effect after spatial alignment, the construction drawings of the curtain wall keel that fit the completed structural entity are derived in reverse. S5. Based on the curtain wall keel construction drawings, the keel and decorative panels are prefabricated in the factory, numbered, and packaged in sections. S6. Based on the curtain wall keel construction drawings, position and install the keel and decorative panels.

2. The construction method for curved curtain wall keel based on three-dimensional laser scanning according to claim 1, characterized in that, S1 includes the following steps: S11, Confirm that there are no obstructions on the surface of the structure to be scanned that would affect the scanning accuracy; S12, Use a total station to set up and mark uniform construction survey benchmark points and baselines on site; S13, Based on the reference points and reference lines, a spatial reference control network consisting of multiple control targets with known coordinates is deployed in the scanning area; S14, Based on the spatial reference control network, plan multiple scanning stations with overlapping areas, so that adjacent scanning stations cover at least one of the control targets; S15, set up a three-dimensional laser scanner at each scanning station to scan the on-site structural entity and simultaneously collect point cloud data of the control target; S16, Based on the known coordinates of the control target, the point cloud data obtained by scanning at each station are stitched together to generate a complete point cloud.

3. The construction method for curved curtain wall keel based on three-dimensional laser scanning according to claim 2, characterized in that, S2 includes the following steps: S21, under the global construction coordinate system defined by the spatial reference control network, the complete point cloud is denoised and filtered to remove interference points and optimize data density, so as to obtain a clean point cloud dataset that characterizes the surface of the structural entity. S22, Analyze the clean point cloud dataset to identify and distinguish between regular geometric features and free-form surface features; S23, for the identified regular geometric feature elements, point cloud slices are generated by creating a cutting plane, two-dimensional contour lines are drawn based on the vectorization of the slice point cloud, and a parametric solid model is generated through extrusion and lofting three-dimensional operations; For the identified freeform surface features, the corresponding point cloud data is converted into a triangular mesh model, and surface fitting, editing and stitching are performed to generate a high-precision continuous surface model; S24 performs Boolean operations and splices the parametric solid model and the continuous surface model to form a three-dimensional model of the structural solid contour.

4. The construction method for curved curtain wall keel based on three-dimensional laser scanning according to claim 3, characterized in that, The point cloud data includes the three-dimensional spatial coordinates of the sampling points, laser reflection intensity information, and / or true color information.

5. The construction method for curved curtain wall keel based on three-dimensional laser scanning according to claim 4, characterized in that, S3 includes the following steps: S31, Obtain the original curved curtain wall design data and build or import a 3D model of the design appearance; S32, import the three-dimensional model of the design appearance effect into the global construction coordinate system defined by the spatial reference control network where the three-dimensional model of the structural entity outline is located; S33, select no fewer than three corresponding feature alignment points in the three-dimensional model of the structural entity outline, and perform spatial transformation on the three-dimensional model of the design appearance effect so that it is aligned with the three-dimensional model of the structural entity outline in terms of spatial position and orientation.

6. The construction method for curved curtain wall keel based on three-dimensional laser scanning according to claim 5, characterized in that, The feature alignment points include the center point of the embedded part and / or the intersection of the structural axis and / or the control points or endpoints of the two-dimensional and / or three-dimensional feature curves used to define the curved shape of the curved curtain wall; the spatial transformation includes rotation and / or movement and / or scaling.

7. The construction method for curved curtain wall keel based on three-dimensional laser scanning according to claim 6, characterized in that, S4 includes the following steps: S41, Pre-set the layout rules of the curtain wall keel system, the layout rules including keel spacing, cross-sectional specifications and connection standards; S42, Based on the structural entity outline model, generate the spatial starting point of the keel according to the arrangement rules; S43, Based on the three-dimensional model of the design appearance effect, generate spatial target points or target surfaces for the keel; S44, by connecting the spatial starting point and the spatial target point, the centerline spatial data of each keel is calculated and generated; S45, Based on the centerline spatial data and cross-sectional specifications, generate a three-dimensional keel model; S46, The three-dimensional keel model is verified. Based on the verified three-dimensional keel model, the curtain wall keel construction drawings and component processing and cutting list are automatically generated. The three-dimensional keel model can be divided into several unit truss models to guide factory prefabrication.

8. The construction method for curved curtain wall keel based on three-dimensional laser scanning according to claim 7, characterized in that, The verification of the three-dimensional keel model in step S46 includes at least one of the following verification methods: Structural interference verification: In the 3D design environment, collision detection analysis is performed on the 3D keel model and the 3D outline model of the structural entity to ensure that there is no physical interference and that there is no mutual interference between the components of the keel model; Construction space verification: In the 3D design environment, check the minimum distance between the outer contour surface of the 3D keel model and the inner surface of the 3D model of the design appearance effect to ensure that it is not less than the minimum construction space thickness required by the curtain wall system. Design rule compliance verification: Check whether the arrangement spacing, cross-section selection and connection method of each component in the three-dimensional keel model are consistent with the arrangement rules of the preset curtain wall keel system; Preliminary verification of mechanical properties: The three-dimensional keel model was imported into structural analysis software, and a preset load was applied to conduct a preliminary stress analysis to verify whether its strength, stiffness and stability meet the design requirements.

9. The construction method for curved curtain wall keel based on three-dimensional laser scanning according to claim 8, characterized in that, S6 includes the following steps: S61. Based on the curtain wall keel construction drawings, using the benchmark points and benchmark lines as a joint benchmark, use a total station to measure and set out the keel installation control lines and control points. S62, according to the keel number, the prefabricated keel is hoisted to the designated area, and positioned, leveled and fixed according to the installation control line and control point; S63. According to the panel number and layout design drawing, the decorative panels are aligned and installed on the corrected keel. S64, sealing and surface cleaning of the decorative panel.

10. A curved curtain wall keel construction system based on three-dimensional laser scanning, used to implement the construction method as described in any one of claims 1-9, characterized in that, include: The 3D laser scanning and point cloud fusion module (101) is used to perform 3D laser scanning on the completed structural entity, acquire point cloud data and stitch it together to form a complete point cloud; The entity contour construction module (102) is used to construct a three-dimensional model of the structural entity contour based on the complete point cloud; The spatial alignment module (103) is used to construct a three-dimensional model of appearance effect based on the three-dimensional model of the structural entity outline, and to spatially align the three-dimensional model of appearance effect with the three-dimensional model of the structural entity outline. The reverse derivation module (104) is used to reverse derive the curtain wall keel construction drawings that fit the completed structural entity based on the three-dimensional model of the outline of the spatially aligned structural entity and the three-dimensional model of the appearance effect. The prefabrication module (105) is used to prefabricate, number, and package the keel and decorative panels in the factory based on the curtain wall keel construction drawings. The positioning and installation module (106) is used to position and install the keel and the decorative panel based on the curtain wall keel construction drawings.

Images