Aluminum mold construction method, system, equipment and medium based on BIM point cloud

Abstract

The application relates to an aluminum formwork construction method, system, equipment and medium based on a BIM point cloud. The method comprises the following steps: aligning BIM design model data of an aluminum formwork with laser scanning point cloud data to obtain a BIM design model and a field point cloud model in the same coordinate system, and performing feature extraction and feature matching processing on the BIM design model and the field point cloud model respectively to obtain a matching point pair set; calculating the spatial deviation of the matching points according to the matching point pair set, and combining a preset kinematic model of a hydraulic leveling system to generate a hydraulic cylinder correction parameter sequence; and the sequence is used to drive the hydraulic leveling system to act and correct the aluminum formwork. The method can accurately align the BIM design model with the laser scanning point cloud data of the construction site, extract and match features, and automatically generate correction parameters in combination with the kinematic model of the hydraulic leveling system, so that the correction accuracy and the automation level of the aluminum formwork construction are significantly improved, the construction efficiency is effectively improved, and human errors are reduced.

Description

Technical Field

[0001] This invention belongs to the field of computer technology, and in particular relates to aluminum formwork construction methods, systems, equipment and media based on BIM point clouds. Background Technology

[0002] Aluminum formwork, with its advantages of light weight, high turnover rate, high construction efficiency, and good forming quality, has been widely used in the construction industry and occupies an important position in the standardized construction of high-rise buildings. With the development of building industrialization and digitalization, Building Information Modeling (BIM) technology, due to its visualization, parametric, and collaborative management characteristics, is gradually being deeply integrated with aluminum formwork construction, providing technical support for the refined management and control of aluminum formwork construction. At the same time, the maturity of laser scanning technology has made it possible to quickly acquire 3D point cloud data from the construction site. This 3D point cloud data can then realistically reflect the actual working conditions of the construction site, providing a data foundation for the accurate correction of aluminum formwork construction.

[0003] However, in existing aluminum formwork installation technologies, the calibration of aluminum formwork construction largely relies on manual measurement and adjustment using equipment such as levels and total stations. This method is not only inefficient but also highly susceptible to human error, making it difficult to meet the demands of high-precision construction. Although some improvement solutions attempt to introduce BIM technology to guide aluminum formwork installation through BIM design models, most solutions only remain at the model visualization level, failing to achieve a precise correlation between the BIM design model and the actual working conditions on the construction site. This results in a disconnect between design and construction, making it impossible to make targeted corrections based on actual deviations. On the other hand, some solutions attempt to combine point cloud data with BIM models, but there are still shortcomings in areas such as the alignment accuracy between the model and the point cloud, the reliability of feature point extraction and matching, and the automated conversion from deviation calculation to correction parameter generation. For example, during the alignment process, coordinate system transformation errors can easily lead to distortion in subsequent deviation calculations; the high mismatch rate during feature matching affects the accuracy of deviation assessment; and the generation of correction parameters relies heavily on empirical formulas, lacking precise adaptation to hydraulic leveling systems. Ultimately, it is difficult to achieve both accuracy and efficiency in aluminum formwork calibration. Summary of the Invention

[0004] Therefore, it is necessary to provide BIM point cloud-based aluminum formwork construction methods, systems, equipment, and media to address the aforementioned technical issues, aiming to improve the automation and accuracy of aluminum formwork construction.

[0005] Firstly, this application provides a BIM point cloud-based aluminum formwork construction method, including:

[0006] The BIM design model data of the aluminum formwork and the laser scanning point cloud data of the construction site are obtained, and the BIM design model data and the laser scanning point cloud data are aligned to obtain the BIM design model and the site point cloud model under the same preset construction coordinate system.

[0007] Based on the BIM design model and the site point cloud model, feature extraction is performed to obtain the BIM feature point set and the site feature point set. The BIM feature point set and the site feature point set are then subjected to feature matching processing to obtain a set of matching point pairs used to represent the correspondence between design points and actual points.

[0008] Based on the set of matching point pairs, the spatial deviation of the matching points is calculated, and combined with the kinematic model of the preset hydraulic leveling system, a hydraulic cylinder correction parameter sequence is generated; the hydraulic cylinder correction parameter sequence is used to drive the hydraulic leveling system to correct the aluminum template.

[0009] In one embodiment, the BIM design model data and laser scanning point cloud data are aligned to obtain a BIM design model and a site point cloud model in the same construction coordinate system, including:

[0010] The BIM design model data is processed to extract the design coordinates of the key feature points of the aluminum formwork. The design coordinates are then transformed to the preset construction coordinate system to obtain the set of BIM design feature points in the preset construction coordinate system.

[0011] The laser scanning point cloud data is processed by performing noise reduction, filtering and multi-station registration operations in sequence to obtain the field point cloud. The field point cloud is then converted to the preset construction coordinate system to obtain the field point cloud model under the preset construction coordinate system.

[0012] The point cloud corresponding to the aluminum template is segmented from the on-site point cloud model to obtain a subset of the on-site point cloud of the aluminum template;

[0013] Spatially align the BIM design feature point set under the preset construction coordinate system with the aluminum formwork site point cloud subset, and calculate the rigid body transformation matrix that minimizes the spatial position difference between the BIM design feature point set and the aluminum formwork site point cloud subset.

[0014] The coordinate transformation of the BIM design model data is performed using a rigid body transformation matrix to obtain the BIM design model in the preset construction coordinate system.

[0015] In one embodiment, feature extraction is performed based on the BIM design model and the site point cloud model to obtain a BIM feature point set and a site feature point set, respectively. Feature matching processing is then performed on the BIM feature point set and the site feature point set to obtain a set of matching point pairs representing the correspondence between design points and actual points, including:

[0016] Extract key feature points of aluminum formwork from BIM design model, calculate local geometric feature descriptors of each key feature point, and obtain BIM feature point set with descriptors;

[0017] Perform key point detection from the on-site point cloud model, extract candidate feature points, calculate the local geometric feature descriptor of each candidate feature point, and obtain the on-site feature point set with descriptors;

[0018] Based on the local geometric feature descriptors of each feature point in the BIM feature point set and the site feature point set, the nearest neighbor search is used to match each BIM feature point in the BIM feature point set with the site feature points in the site feature point set, forming an initial set of matching pairs.

[0019] Calculate the descriptor distance ratio of each matching pair in the initial matching pair set, and remove matching pairs whose descriptor distance ratio is greater than a preset ratio threshold to obtain a preliminary filtered matching pair set;

[0020] A spatial transformation model is established based on the rigid body transformation matrix. The geometric consistency of the preliminary set of matching pairs is verified by combining the random sampling consensus algorithm with the spatial transformation model. Mismatched pairs that do not conform to the same spatial transformation model are eliminated, and a set of matching point pairs is obtained.

[0021] In one embodiment, based on the set of matching point pairs, the spatial deviation of the matching points is calculated, and combined with a preset kinematic model of the hydraulic leveling system, a sequence of hydraulic cylinder correction parameters is generated, including:

[0022] Based on the design coordinates and actual coordinates of each point pair in the matching point pair set, calculate the three-dimensional position deviation vector of each matching point. Combine the BIM feature point set and the on-site feature point set to calculate the attitude deviation angle of each matching point, and obtain the six-degree-of-freedom pose deviation vector of each matching point. Use the six-degree-of-freedom pose deviation vector as the spatial deviation.

[0023] Information on aluminum formwork components is extracted from the BIM design model. Cluster analysis is performed on the six-degree-of-freedom pose deviation vectors of each matching point to identify the deviation patterns of the aluminum formwork and assess the impact of the deviation patterns on construction, thus obtaining the impact assessment results.

[0024] The priority of components requiring correction was determined by combining information on aluminum formwork components and the results of the impact assessment.

[0025] Based on the deviation mode, impact assessment results and priorities, and combined with the pre-set kinematic model of the hydraulic leveling system, an optimization problem is constructed with the goal of minimizing residual deviation and with hydraulic cylinder stroke and thrust as constraints.

[0026] The optimization problem is solved by using a sequential quadratic programming algorithm to obtain the extension and retraction of each hydraulic cylinder. The extension and retraction are then arranged into a sequence of hydraulic cylinder correction parameters according to the correction order.

[0027] In one embodiment, the method further includes:

[0028] S41. Decode the hydraulic cylinder calibration parameter sequence into control commands and send them to the hydraulic leveling system to drive the hydraulic cylinder to adjust the position of the aluminum template.

[0029] S42. During the adjustment process, real-time point cloud data of the aluminum template is collected in real time, and noise reduction processing is performed on the real-time point cloud data to obtain the processed real-time point cloud data.

[0030] S43. Extract the current coordinates of key points of the aluminum template from the processed real-time point cloud data;

[0031] S44. Compare the current coordinates with the design coordinates of the corresponding key points of the aluminum formwork obtained from the BIM design model, and calculate the current residual deviation of each key point of the aluminum formwork.

[0032] S45. Determine whether the current residual deviation is lower than the preset accuracy threshold and obtain the determination result; if the determination result is no, adjust the hydraulic cylinder calibration parameter sequence based on the current residual deviation to obtain the adjusted hydraulic cylinder calibration parameter sequence, and return the adjusted hydraulic cylinder calibration parameter sequence to execute S41; if the determination result is yes, stop the calibration.

[0033] In one embodiment, the adjusted hydraulic cylinder calibration parameter sequence is calculated using the following formula:

[0034]

[0035]

[0036] in, This represents the adjustment value of the extension / retraction amount of the j-th hydraulic cylinder in the (k+1)-th correction; Indicates the proportionality coefficient; Indicates the integral coefficient; Represents the differential coefficient; This represents the residual deviation of the key point of the aluminum template corresponding to the j-th hydraulic cylinder after the k-th correction. ,in The coordinates are after the k-th correction. For design coordinates; Indicates the time interval between two corrections; Indicates the gradient step size coefficient; Represents the residual deviation after the k-th correction. The gradient vector represents the rate of change of the deviation with respect to the extension and retraction of the hydraulic cylinder. ; The sequence of adjusted hydraulic cylinder calibration parameters after the k-th calibration is denoted as . The vector, This refers to the number of hydraulic cylinders. ; This represents the sequence of hydraulic cylinder extension / retraction adjustments during the (k+1)th correction. The vector, .

[0037] In one embodiment, a keypoint detection operation is performed from the on-site point cloud model to extract candidate feature points, including:

[0038] The ISS key point detection algorithm is used to perform key point detection on the on-site point cloud model. Based on the preset first and second scaling factors, the neighborhood radius is three times the average spacing of the points in the on-site point cloud model.

[0039] The scattering matrix of each point cloud in the on-site point cloud model is calculated based on the neighborhood radius. The eigenvalues ​​of the scattering matrix are solved, and the point clouds whose eigenvalues ​​meet the preset conditions are selected as the initial key points. The preset conditions are set based on the first and second scaling factors.

[0040] Non-maximum suppression is performed on the initial key points. Based on the preset suppression radius, duplicate initial key points within the preset suppression radius are removed to obtain candidate feature points.

[0041] Secondly, this application also provides an aluminum formwork construction system based on BIM point clouds, including:

[0042] The data alignment and model fusion module is used to acquire the BIM design model data of aluminum formwork and the laser scanning point cloud data of the construction site, and to align the BIM design model data and the laser scanning point cloud data to obtain the BIM design model and the site point cloud model under the same preset construction coordinate system.

[0043] The feature extraction and matching module is used to extract features based on the BIM design model and the site point cloud model respectively, to obtain the BIM feature point set and the site feature point set, and to perform feature matching processing on the BIM feature point set and the site feature point set to obtain a set of matching point pairs that represent the correspondence between design points and actual points.

[0044] The calibration parameter generation module is used to calculate the spatial deviation of the matching points based on the set of matching point pairs, and generate a hydraulic cylinder calibration parameter sequence in combination with the preset kinematic model of the hydraulic leveling system. The hydraulic cylinder calibration parameter sequence is used to drive the hydraulic leveling system to correct the aluminum template.

[0045] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the first aspect.

[0046] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the first aspect.

[0047] The aforementioned BIM point cloud-based aluminum formwork construction method, system, equipment, and medium first acquire and align the BIM design model data of the aluminum formwork with the laser-scanned point cloud data of the construction site, resolving the issues of data disconnect and coordinate system inconsistency between traditional design and site data, thus improving the accuracy of data fusion. Secondly, feature extraction and matching are performed on the aligned model, addressing the high mismatch rate of traditional feature matching and improving the reliability of the correspondence between design points and actual points. Finally, spatial deviation is calculated based on the matching point pairs, and a correction parameter sequence is generated using the kinematic model of the hydraulic leveling system, enhancing the automation and accuracy of aluminum formwork correction. Attached Figure Description

[0048] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 A flowchart of an aluminum formwork construction method based on BIM point cloud is provided as an exemplary embodiment of the present invention;

[0050] Figure 2 A flowchart of a method for generating a hydraulic cylinder calibration parameter sequence is provided as an exemplary embodiment of the present invention;

[0051] Figure 3 This is a schematic diagram of an aluminum formwork construction system based on BIM point cloud, provided as an exemplary embodiment of the present invention. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0053] In one embodiment, such as Figure 1 As shown, a BIM point cloud-based aluminum formwork construction method is provided. This embodiment illustrates the method's application to a terminal. It is understood that this method can also be applied to a server, and to a system including both a terminal and a server, and implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0054] S101: Obtain the BIM design model data of the aluminum formwork and the laser scanning point cloud data of the construction site, and align the BIM design model data and the laser scanning point cloud data to obtain the BIM design model and the site point cloud model under the same preset construction coordinate system.

[0055] Specifically, BIM design model data can be exported using mainstream BIM modeling software such as Revit and Bentley. This data can include information such as the 3D geometric coordinates, component numbers, and material properties of each component of the aluminum formwork. The 3D geometric coordinates must be stored in the native coordinate system of the modeling software to reserve an interface for subsequent coordinate transformation. Laser scanning point cloud data from the construction site can be acquired using a terrestrial 3D laser scanner. During the scanning process, the scanning range must cover the aluminum formwork installation area and at least three fixed reference points around it. Furthermore, the point cloud density must be set to be compatible with the construction area to ensure complete coverage of key details on the aluminum formwork surface. After scanning, the point cloud data can be exported using the scanner's accompanying software. By aligning these two types of data—that is, transforming the BIM design model from its original design coordinate system to the construction site coordinate system—the BIM design model and the site point cloud model are placed under the same preset construction coordinate system, thus providing a unified reference framework for subsequent feature extraction and correction.

[0056] S102: Based on the BIM design model and the site point cloud model, feature extraction is performed to obtain the BIM feature point set and the site feature point set. The BIM feature point set and the site feature point set are then subjected to feature matching processing to obtain a set of matching point pairs used to represent the correspondence between design points and actual points.

[0057] Specifically, after aligning the BIM design model and the on-site point cloud model, points with significant geometric features can be identified by analyzing the model's geometry and topology. This simplifies the complex model into a set of key points, facilitating subsequent matching. For example, in the BIM design model, the location of feature points can be determined by analyzing the model's geometric parameters. In point cloud data, corresponding feature points can be identified using point cloud analysis algorithms, such as curvature-based or normal-vector-based analysis. The extracted feature point sets can be referred to as the BIM feature point set and the on-site feature point set, respectively. Subsequently, the BIM feature point set and the on-site feature point set can be matched. Similarity metrics between the two point sets can be calculated using Euclidean distance or other geometric similarity indices to find the correspondence between design points and actual points, forming a set of matched point pairs.

[0058] S103: Based on the set of matching point pairs, calculate the spatial deviation of the matching points, and combine it with the preset kinematic model of the hydraulic leveling system to generate a hydraulic cylinder correction parameter sequence; the hydraulic cylinder correction parameter sequence is used to drive the hydraulic leveling system to correct the aluminum template.

[0059] Specifically, the hydraulic cylinders can be extended and retracted through a hydraulic leveling system to adjust the spatial pose of the aluminum template, making the actual pose approximate the design pose. Therefore, after obtaining the set of matching point pairs, the spatial deviation—the difference between the design and actual points—can be calculated to quantify the degree of difference between the design and reality. This spatial deviation can then be used as the basis for generating a sequence of hydraulic cylinder correction parameters. Furthermore, the kinematic model of the hydraulic leveling system describes the relationship between the movement of the hydraulic cylinders and the template position. By combining the deviation information with the kinematic model of the hydraulic leveling system, a series of hydraulic cylinder correction parameters can be generated by comprehensively considering factors such as the dynamic characteristics of the hydraulic system, the structural characteristics of the template, and construction requirements. Finally, the generated sequence of hydraulic cylinder correction parameters can be sent to the hydraulic leveling system to drive the hydraulic cylinders to perform corresponding extension and retraction actions, thereby adjusting the position of the aluminum template to meet the design requirements and achieving the correction of the aluminum template.

[0060] The above method first achieves precise integration of design and construction site data through alignment processing of the BIM design model and laser scanning point cloud data, enhancing the accuracy of construction correction. Secondly, based on feature extraction and matching, it accurately calculates the spatial deviation of matching points, effectively solving the problem of large human measurement errors in traditional methods and improving the reliability of correction. Finally, by combining the kinematic model of the hydraulic leveling system to generate correction parameters, it automates the correction process, improving construction efficiency and reducing labor costs.

[0061] In one embodiment, BIM design model data and laser scan point cloud data are aligned to obtain a BIM design model and a site point cloud model in the same construction coordinate system, including:

[0062] The BIM design model data is processed to extract the design coordinates of the key feature points of the aluminum formwork. The design coordinates are then transformed to the preset construction coordinate system to obtain the set of BIM design feature points in the preset construction coordinate system.

[0063] The laser scanning point cloud data is processed by performing noise reduction, filtering and multi-station registration operations in sequence to obtain the field point cloud. The field point cloud is then converted to the preset construction coordinate system to obtain the field point cloud model under the preset construction coordinate system.

[0064] The point cloud corresponding to the aluminum template is segmented from the on-site point cloud model to obtain a subset of the on-site point cloud of the aluminum template;

[0065] Spatially align the BIM design feature point set under the preset construction coordinate system with the aluminum formwork site point cloud subset, and calculate the rigid body transformation matrix that minimizes the spatial position difference between the BIM design feature point set and the aluminum formwork site point cloud subset.

[0066] The coordinate transformation of the BIM design model data is performed using a rigid body transformation matrix to obtain the BIM design model in the preset construction coordinate system.

[0067] Specifically, since BIM design model data contains non-spatial information such as component materials and construction techniques, directly using it for coordinate alignment would increase computational redundancy. Therefore, it's advisable to first extract key feature points that characterize the spatial orientation of the aluminum formwork. For example, a data extraction script can be written using the secondary development interface of the BIM modeling software. This script can automatically traverse all components of the aluminum formwork in the BIM design model, identifying and extracting key feature points such as corner points, joint midpoints, and bolt hole centers. These feature points must be spatially stable during aluminum formwork installation, resistant to deformation due to construction operations, and accurately correspond to similar features in the on-site laser scanning point cloud data. After extraction, the design coordinates of each key feature point in the native coordinate system of the BIM modeling software are obtained. Since the native coordinate system differs from the actual coordinate system on the construction site, the design coordinates can be transformed to a preset construction coordinate system. This coordinate system typically uses the geodetic coordinate system commonly used in building engineering, established with a preset benchmark point on the construction site as the origin. The coordinate transformation process can employ a seven-parameter coordinate transformation method, based on a transformation formula, using three translation parameters, three rotation parameters, and one scale parameter to establish a linear mapping relationship between different coordinate systems. The parameters of this formula can be obtained by using three or more known reference points in the preset construction coordinate system at the construction site. Specifically, the design coordinates of these reference points in the native BIM coordinate system and the actual measured coordinates in the preset construction coordinate system are obtained, substituted into the transformation formula to construct a system of equations, and then solved using the least squares method to obtain the optimal solution for the seven parameters. Substituting the native design coordinates of all key feature points into the transformation formula yields a set of BIM design feature points in the preset construction coordinate system. This set is stored in a table format, containing information such as feature point number, preset construction coordinate system coordinates, and feature point type.

[0068] Specifically, laser scanning point cloud data is easily affected by dust, lighting, and inherent scanner errors during the acquisition process, resulting in isolated noise points and uneven point cloud density. Therefore, preprocessing is necessary. Statistical filtering can be used for denoising, utilizing the statistical distribution characteristics of the point cloud to remove noise points that deviate from the normal distribution. Gaussian filtering can be used for further filtering, smoothing the denoised point cloud by constructing a Gaussian kernel function. This eliminates high-frequency jitter in the point cloud data while preserving the geometric features of the aluminum template surface. The weights of the Gaussian kernel function are determined by the distance between the point and the center pixel; the closer the point, the greater the weight, ensuring that the spatial morphology of the filtered point cloud matches the actual aluminum template.

[0069] Because there may be scanning blind spots in the aluminum formwork installation area at the construction site, single-station laser scanning cannot obtain complete point cloud data. Therefore, multi-station registration can be performed. During multi-station registration, an iterative nearest neighbor algorithm can be used to find the optimal matching relationship between point clouds from different stations. For example, the point cloud scanned by one station is selected as the target point cloud, and the point clouds from the remaining stations are used as source point clouds. The nearest neighbor of each point in the source point cloud in the target point cloud is calculated to construct initial matching point pairs. The rigid body transformation matrix that minimizes the average distance between the matching point pairs is calculated using the least squares method to transform the source point cloud to the target point cloud coordinate system. This iterative process is repeated until the average distance error between two adjacent iterations is less than a preset convergence threshold. This completes the stitching of the multi-station point clouds, resulting in a complete on-site point cloud covering the aluminum formwork installation area. In addition, after preprocessing, the complete site point cloud can be converted to a preset construction coordinate system. The conversion method is consistent with the coordinate conversion of BIM design feature points. That is, three reference points with known preset construction coordinate system coordinates are selected at the construction site. The coordinates of these reference points in the original coordinate system of the point cloud are obtained through the scanner software. The conversion parameters are substituted into the conversion formula to solve for the conversion parameters. The coordinate transformation of the complete site point cloud is performed using these parameters to obtain the site point cloud model in the preset construction coordinate system. This model fully presents the spatial form of the aluminum formwork and surrounding environment of the construction site in the form of a discrete point set.

[0070] Specifically, the site point cloud model under the preset construction coordinate system contains various scene elements such as aluminum formwork, scaffolding, and construction equipment. If it is directly used for alignment with the BIM design model, the alignment accuracy will decrease due to interference from irrelevant point clouds. Therefore, a region-growing-based segmentation algorithm can be used. By utilizing the consistency of the normal vector direction and curvature value of the point cloud on the aluminum formwork surface, points with similar geometric features are clustered into the same region, and the point cloud subset corresponding to the aluminum formwork is segmented from the site point cloud model. For example, the normal vector and curvature value of each point in the site point cloud model are first calculated using principal component analysis, that is, a neighborhood is defined for each point, and the covariance matrix of the points in the neighborhood is constructed. By solving the eigenvalues ​​and eigenvectors of the covariance matrix, the normal vector (the eigenvector corresponding to the smallest eigenvalue) and curvature value (calculated from the eigenvalue) of that point are obtained. Subsequently, normal vector angle thresholds and curvature thresholds can be set as constraints for region growth. A seed point on the aluminum template surface is selected as the starting point for growth. This involves including all points in the seed point's neighborhood whose normal vectors have an angle less than the seed point's normal vector and whose curvature values ​​are less than the curvature threshold. These points are then used as new seed points, and the above process is repeated until no new points can be included in the region. After growth, the resulting region is the point cloud region corresponding to the aluminum template. Removing other region point clouds from the on-site point cloud model yields a subset of the aluminum template's on-site point cloud. This subset contains only the point cloud data of the aluminum template surface, effectively reducing the computational load of subsequent alignment operations.

[0071] Specifically, although the BIM design feature point set and the aluminum formwork site point cloud subset are in the same preset construction coordinate system, slight deviations still exist in their spatial positions due to measurement and modeling errors. Therefore, a rigid body transformation matrix that minimizes the spatial positional difference between the two point sets can be calculated. This rigid body transformation only changes the spatial position and orientation of the point sets, without altering the relative distances within the point sets, thus meeting the requirements for model alignment. Specifically, spatial alignment uses a least-squares method based on feature point matching to solve for the rigid body transformation matrix. Let the BIM design feature point set be... Let i = 1, 2, ..., n, and the feature point set corresponding to the subset of the point cloud at the aluminum template site be... (Find the site feature points corresponding to the design feature points through nearest neighbor search), the rigid body transformation matrix contains translation vectors. With rotation matrix R, the transformation formula is:

[0072]

[0073] The goal of spatial alignment is to minimize the sum of squared positional errors of all matching feature point pairs. The error function can be:

[0074]

[0075] To solve this problem, first calculate the centroid coordinates of the two types of feature point sets. and The decentralized coordinates are obtained by subtracting the centroid from the coordinates of the feature points. and Then the covariance matrix can be constructed. ,right Singular value decomposition yields Rotation matrix Translation vector Thus, the error function is obtained. The smallest rigid body transformation matrix.

[0076] Finally, the batch coordinate transformation function of the BIM modeling software can be used to import the solved rigid body transformation matrix into the software. The software automatically traverses all components in the BIM design model and performs rotation and translation operations on the three-dimensional geometric coordinates of each component, that is, the original coordinates of each point on the component. Through formula The coordinates are converted to a preset construction coordinate system. After the coordinate transformation, the accuracy of the BIM design model and the on-site point cloud model can be verified. This involves randomly selecting 10-20 key feature points of the aluminum formwork and obtaining their coordinates in both models. The absolute value of the coordinate difference is calculated to ensure that the coordinate difference of all feature points is less than a preset accuracy threshold. Once the verification is successful, a BIM design model in the same preset construction coordinate system as the on-site point cloud model is obtained. The spatial positional deviation between the two models is controlled within a very small range, providing reliable basic data for subsequent deviation analysis based on feature matching.

[0077] In one embodiment, feature extraction is performed based on the BIM design model and the site point cloud model to obtain a BIM feature point set and a site feature point set, respectively. Feature matching processing is then performed on the BIM feature point set and the site feature point set to obtain a set of matching point pairs representing the correspondence between design points and actual points, including:

[0078] Extract key feature points of aluminum formwork from BIM design model, calculate local geometric feature descriptors of each key feature point, and obtain BIM feature point set with descriptors;

[0079] Perform key point detection from the on-site point cloud model, extract candidate feature points, calculate the local geometric feature descriptor of each candidate feature point, and obtain the on-site feature point set with descriptors;

[0080] Based on the local geometric feature descriptors of each feature point in the BIM feature point set and the site feature point set, the nearest neighbor search is used to match each BIM feature point in the BIM feature point set with the site feature points in the site feature point set, forming an initial set of matching pairs.

[0081] Calculate the descriptor distance ratio of each matching pair in the initial matching pair set, and remove matching pairs whose descriptor distance ratio is greater than a preset ratio threshold to obtain a preliminary filtered matching pair set;

[0082] A spatial transformation model is established based on the rigid body transformation matrix. The geometric consistency of the preliminary set of matching pairs is verified by combining the random sampling consensus algorithm with the spatial transformation model. Mismatched pairs that do not conform to the same spatial transformation model are eliminated, and a set of matching point pairs is obtained.

[0083] Specifically, the BIM design model is a structured 3D model. Directly comparing the entire model would lead to computational redundancy due to the high data dimensionality. Therefore, it is advisable to first extract key feature points that characterize the spatial pose of the aluminum formwork. For example, an automated extraction script can be written using the secondary development interface of the BIM modeling software. After traversing all components of the aluminum formwork, the script automatically identifies and extracts feature points such as corner points, joint midpoints, bolt hole centers, and formwork edge endpoints. These feature points have fixed spatial positions during the aluminum formwork processing and installation, and their geometric shapes are easily identifiable in the laser scanning point cloud. After extraction, the 3D coordinates of each feature point in the preset construction coordinate system are obtained, and the component number and feature type to which the feature point belongs are recorded. Subsequently, to achieve feature point similarity comparison, the FPFH (Fast Point Feature Histogram) descriptor can be used to calculate the local geometric feature descriptor of each key feature point. The technical principle is to construct a histogram vector that reflects the local geometric shape by statistically analyzing the angle between the normal vectors of the points in the neighborhood of the feature point and the spatial distance distribution. It has rotation invariance and scale robustness, and adapts to the posture changes and measurement errors during aluminum formwork construction. During calculation, a neighborhood radius is set centered on the feature point, and the size of the neighborhood radius can be determined according to the dimensions of the aluminum formwork component. The angle between the normal vector of all points within the neighborhood and the normal vector of the feature point is calculated. The angle range is divided into several intervals, and the number of points in each interval is counted. This is then weighted based on the distance distribution between the points and the feature point to generate a fixed-dimensional histogram vector, which serves as the FPFH descriptor for that feature point. Associating and storing the coordinates of each key feature point with its corresponding descriptor yields a set of BIM feature points with descriptors.

[0084] Specifically, the on-site point cloud model is a discrete set of points. First, a region-growing algorithm can be used to filter the aluminum template region from this set by considering the consistency of normal vectors and curvature, segmenting the point cloud corresponding to the aluminum template. Then, candidate feature points are extracted from this subset. During key point detection, a geometric saliency-based detection method can be used to ensure that the extracted candidate feature points are consistent with the key feature point types of the BIM design model. After detection, a local geometric feature descriptor can be calculated for each candidate feature point. The descriptor type is consistent with the descriptor of the BIM feature points, i.e., the FPFH descriptor, and the calculation parameters, such as neighborhood radius and histogram interval division, are consistent to ensure the fairness of feature similarity comparison and improve matching accuracy. By associating and storing the coordinates of the candidate feature points with their corresponding descriptors, a set of on-site feature points with descriptors can be obtained.

[0085] Furthermore, a nearest neighbor search algorithm can be used to find the corresponding feature points in the design and actual locations based on the similarity of descriptors. For example, taking each feature point in the BIM feature point set as the target point, all feature points in the site feature point set are traversed, and the Euclidean distance between the target point descriptor and the site feature point descriptor is calculated. The smaller the Euclidean distance, the more similar the local geometry of the two feature points. Subsequently, a K-nearest neighbor search (K=2) can be used to initially screen reliable matching pairs. That is, for each target point, the two site feature points with the smallest Euclidean distance in their descriptors (the first nearest neighbor and the second nearest neighbor) are found, and the target point and the first nearest neighbor are formed into an initial matching pair. After summing all the initial matching pairs, an initial matching pair set can be formed. However, the initial matching pair set may contain false matches, such as noise points in the site point cloud that are spuriously similar to design feature points. Therefore, a distance ratio method can also be used for preliminary screening. This distance ratio is the ratio of the "Euclidean distance between the target point and the first nearest neighbor" to the "Euclidean distance between the target point and the second nearest neighbor" in each initial matching pair. If the ratio is small, it indicates that the similarity between the first nearest neighbor and the target point is significantly higher than that between the second nearest neighbor, indicating high matching reliability. If the ratio is large, it indicates the existence of multiple similar feature points, indicating low matching reliability. As an illustration, a preset ratio threshold can be set, and the distance ratio of each matching pair can be calculated by traversing the initial matching pair set. Matching pairs with ratios greater than the threshold are removed, and the remaining matching pairs form a preliminary filtered matching pair set. This can further eliminate false matches and reduce the computational load of subsequent verification.

[0086] Specifically, a small number of mismatches may still exist in the initial set of matched pairs. Therefore, further elimination can be achieved through geometric consistency verification. This involves leveraging the fact that both types of models are already in the same preset construction coordinate system, and establishing a spatial transformation model based on the rigid body transformation matrix obtained in the alignment steps described above to perform this elimination process. This rigid body transformation matrix reflects the spatial transformation law from BIM design feature points to actual site feature points, conforming to the characteristics of rigid body transformation (only translation and rotation, no scaling). During the verification process, a random sampling consistency algorithm can be used. This involves randomly sampling some matched pairs to fit the spatial transformation model, counting the number of matched pairs (number of interior points) that conform to the model, iterating multiple times, and selecting the model with the most interior points as the optimal model. Finally, all matched pairs corresponding to interior points are retained. Illustratively, in each iteration, three matched pairs can be randomly selected from the initial set of matched pairs (three points can determine the rigid body transformation), and a temporary rigid body transformation matrix can be solved based on the coordinates of these three matched pairs. By traversing all matching pairs, the Euclidean distance between the coordinates of the BIM feature points in each matching pair after transformation by a temporary matrix and the coordinates of the on-site feature points is calculated. If the distance is less than a preset interior point threshold, reflecting the allowable measurement error during construction, it is determined to be an interior point, and the number of interior points and the temporary matrix for that iteration are recorded. After reaching a preset number of iterations, the temporary matrix with the most interior points is selected as the optimal spatial transformation model. Finally, all matching pairs (exterior points) that do not satisfy the optimal model are removed, and the matching pairs corresponding to the remaining interior points form a set of matching point pairs.

[0087] In one embodiment, performing keypoint detection from a point cloud model to extract candidate feature points includes:

[0088] The ISS key point detection algorithm is used to perform key point detection on the on-site point cloud model. Based on the preset first and second scaling factors, the neighborhood radius is three times the average spacing of the points in the on-site point cloud model.

[0089] The scattering matrix of each point cloud in the on-site point cloud model is calculated based on the neighborhood radius. The eigenvalues ​​of the scattering matrix are solved, and the point clouds whose eigenvalues ​​meet the preset conditions are selected as the initial key points. The preset conditions are set based on the first and second scaling factors.

[0090] Non-maximum suppression is performed on the initial key points. Based on the preset suppression radius, duplicate initial key points within the preset suppression radius are removed to obtain candidate feature points.

[0091] Specifically, the ISS (Intrinsic Shape Signatures) key point detection algorithm is used. This algorithm extracts on-site feature points that accurately match BIM design feature points by quantifying the local geometric saliency and uniqueness of feature points, avoiding the problems of feature point redundancy or missed detection in traditional detection methods. The core of the ISS algorithm is to analyze the scattering matrix eigenvalues ​​of the local neighborhood of the point cloud and select points with intrinsic geometric saliency as key points. Using a neighborhood radius of 3 times the average spacing of the point cloud ensures that the neighborhood contains a sufficient number of points (usually no less than 50), thus accurately reflecting the local geometric shape while avoiding feature blurring caused by an excessively large neighborhood. Preset first and second scaling factors are used for subsequent feature value selection to quantify the "saliency" of feature points; the closer the scaling factor is to 1, the higher the geometric saliency of the selected feature points. For example, all points can be traversed first, the distance between each point and its nearest neighbor can be calculated, and the average distance can be taken to obtain the average spacing of the point cloud model on-site. This average spacing is then multiplied by 3 to obtain the neighborhood radius, while simultaneously importing the preset first and second scaling factors.

[0092] Specifically, the scattering matrix is ​​used to describe the spatial distribution characteristics of points within the neighborhood of a feature point. During calculation, each point cloud can be considered as a candidate point, and all points within a neighborhood radius can be defined as neighborhood points. The formula for calculating the scattering matrix is ​​as follows:

[0093]

[0094] in, Let k be the scattering matrix, and k be the number of points in the neighborhood. For the neighboring region The three-dimensional coordinates of the points Let the coordinates be the centroid coordinates of all points in the neighborhood. This indicates the transpose. The eigenvalues ​​of this matrix reflect the dispersion of neighborhood points in different directions. The larger the eigenvalue, the more dispersed the neighborhood points are in the direction of the corresponding eigenvector.

[0095] After solving for the three eigenvalues ​​of the above scattering matrix, they can be sorted in ascending order to obtain... , , ,in The direction in which the distribution of neighboring points is most dispersed. The direction corresponding to the highest concentration. The preset condition can be set to " / >Second proportional coefficient and / "First proportional coefficient". Based on this condition, candidate points only possess geometric significance when the three feature values ​​show significant gradient differences (e.g., the neighborhood distribution of corner points and edge points will show obvious gradient characteristics). Traverse all point clouds in the on-site point cloud model, calculate the scattering matrix and feature values ​​of each point cloud, and select point clouds that meet the preset conditions as initial key points.

[0096] Furthermore, the initial keypoints may contain duplicate keypoints that are too close together in space. For example, a dense point cloud on the flat surface of an aluminum template may detect multiple neighboring initial keypoints. These duplicate points increase the redundancy of subsequent descriptor calculations and matching. Therefore, non-maximum suppression can be used to ensure that the final candidate feature points are spatially uniformly distributed by setting a preset suppression radius. For example, all initial keypoints are first divided into grids according to their three-dimensional coordinates, with the grid size consistent with the preset suppression radius. Then, for each initial keypoint within a grid, the maximum eigenvalue of its scattering matrix is ​​calculated. ,reserve The largest keypoint (with the highest geometric saliency) is selected, and all other initial keypoints within the raster are removed. After all raster data is processed, the remaining keypoints are the candidate feature points. This process reduces the number of feature points while retaining highly saliency, thus improving the efficiency of subsequent processing.

[0097] In one embodiment, such as Figure 2 As shown, based on the set of matching point pairs, the spatial deviation of the matching points is calculated, and combined with the preset kinematic model of the hydraulic leveling system, a sequence of hydraulic cylinder correction parameters is generated, including:

[0098] S201: Based on the design coordinates and actual coordinates of each point pair in the matching point pair set, calculate the three-dimensional position deviation vector of each matching point, combine the BIM feature point set and the on-site feature point set to calculate the attitude deviation angle of each matching point, and obtain the six-degree-of-freedom pose deviation vector of each matching point. Use the six-degree-of-freedom pose deviation vector as the spatial deviation.

[0099] S202: Extract aluminum formwork component information from the BIM design model, perform cluster analysis on the six-degree-of-freedom pose deviation vectors of each matching point, identify the deviation patterns of the aluminum formwork, evaluate the impact of the deviation patterns on construction, and obtain the impact assessment results.

[0100] S203: Determine the priority of components requiring correction by combining information on aluminum formwork components and impact assessment results; Based on deviation patterns, impact assessment results and priorities, and combined with the pre-set kinematic model of the hydraulic leveling system, construct an optimization problem with the goal of minimizing residual deviation and constraints of hydraulic cylinder stroke and thrust.

[0101] S204: The optimization problem is solved by using a sequential quadratic programming algorithm to obtain the extension and retraction of each hydraulic cylinder. The extension and retraction are then arranged into a hydraulic cylinder calibration parameter sequence according to the calibration order.

[0102] Specifically, each matching point pair in the matching point pair set contains the design coordinates of the BIM design feature points and the actual coordinates of the site feature points. The three-dimensional position deviation vector can be calculated using the vector subtraction rule, as shown in the formula. ,in It is a three-dimensional position deviation vector, reflecting the deviation of the matching point in the three translational directions of X, Y, and Z under the preset construction coordinate system; To match the 3D design coordinates of BIM design feature points in a point pair, This provides the actual 3D coordinates of the corresponding site feature points. The attitude deviation angle can then be calculated using the normal vector information of the feature points. This normal vector can be obtained through principal component analysis (PCA), which involves defining neighborhoods for both the BIM design feature points and the site feature points, constructing the covariance matrix of the points within each neighborhood, and finding the eigenvector corresponding to the smallest eigenvalue of the matrix. Based on the normal vectors of the BIM design feature points and the site feature points, the angle between the two normal vectors can be calculated using their dot product as the attitude deviation angle. Combining the 3D position deviation vector with the attitude deviation angle yields a six-degree-of-freedom pose deviation vector for each matching point. This vector includes X, Y, and Z translational deviations and rotational deviations around the three axes, quantifying the design-actual difference of the aluminum formwork in space.

[0103] Specifically, when extracting aluminum formwork component information from the BIM design model, key information such as component number, dimensional parameters, material strength, connection relationships, and load-bearing capacity can be automatically obtained through the interface script of the BIM modeling software. This information provides a foundation for deviation pattern analysis and correction priority determination. Subsequently, the K-means clustering algorithm can be used to perform cluster analysis on the six-degree-of-freedom pose deviation vectors of each matching point. That is, the deviation vectors are divided into several clusters using the Euclidean distance of the deviation vectors as the similarity metric. Matching points within the same cluster correspond to the same aluminum formwork component or adjacent components. The distribution characteristics of the deviation vectors within the cluster can reflect the deviation pattern. For example, if the Z-axis translational deviation of all points in a cluster is positive and the values ​​are close, it indicates that the corresponding component has an overall upward offset. If the rotational deviation around the X-axis of points in a cluster is concentrated, it indicates that there is a tilt deviation in the corresponding area. Furthermore, by combining the construction technical requirements of aluminum formwork, the impact of deviation patterns on construction can be assessed. For example, if the offset deviation of load-bearing components exceeds the limit, it will affect the structural stability; if the posture deviation at the joints leads to grout leakage, the assessment can be based on the ratio of the deviation value to the specification limit and the functional importance of the component (e.g., load-bearing components have higher priority than decorative components). This assessment result can be divided into three levels: severe impact, moderate impact, and minor impact. In addition, by combining the information of aluminum formwork components with the impact assessment results, the priority of components requiring correction can be obtained. For example, components marked as load-bearing components with a severe impact assessment result have the highest priority, while components marked as "non-load-bearing infill components" with a minor impact assessment result have the lowest priority. This ensures that the correction process addresses key issues first and avoids wasting resources.

[0104] Specifically, when constructing the optimization problem based on the deviation pattern, impact assessment results, and priorities, the core is to establish a kinematic model of the hydraulic leveling system. This model can describe the mapping relationship between the hydraulic cylinder extension / retraction and the aluminum template pose change using a Jacobian matrix, denoted as the hydraulic cylinder extension / retraction vector. (Including the extension and retraction values ​​of each hydraulic cylinder), the six-degree-of-freedom pose adjustment vector of the aluminum template is... Then the kinematic model expression is ,in The Jacobian matrix represents the change in the aluminum template's orientation direction when a single hydraulic cylinder extends or retracts by a unit length. Its value is calculated using the hydraulic cylinder's installation coordinates, the connection node parameters with the aluminum template, and the template's stiffness characteristics. The optimization objective can be set as minimizing the residual deviation of the corrected aluminum template, where the residual deviation is the sum of the magnitudes of the six-degree-of-freedom orientation deviation vectors at the corrected matching points. Constraints can include stroke and thrust constraints on the hydraulic cylinder. Stroke constraints are determined by the cylinder's physical structure, such as ensuring the extension / retraction amount does not exceed the equipment's rated stroke range. Thrust constraints are determined by the hydraulic system's rated output power, such as ensuring the thrust does not exceed the equipment's rated thrust to prevent equipment damage or template deformation. The resulting optimization problem is a constrained nonlinear programming problem.

[0105] The optimization problem can then be solved using a sequential quadratic programming algorithm. This algorithm transforms the nonlinear optimization problem into a series of quadratic programming subproblems for iterative solving, efficiently handling constrained optimization scenarios and adapting to the kinematic model characteristics of the hydraulic leveling system. During the solution process, the six-degree-of-freedom pose deviation vectors of each matching point can be used as the target adjustment amount. An initial quadratic programming subproblem can be constructed using the Jacobian matrix to obtain the initial hydraulic cylinder extension / retraction amount. Then, the objective function weights are adjusted based on the residual deviations corresponding to the initial extension / retraction amount, iteratively optimizing until the residual deviations meet the construction accuracy requirements, ultimately obtaining the optimal extension / retraction amount for each hydraulic cylinder. Finally, when organizing the extension / retraction amounts into a hydraulic cylinder correction parameter sequence, the correction order can be determined by combining the correction priority with the mechanical characteristics of the aluminum formwork. For example, the hydraulic cylinders corresponding to high-priority load-bearing components can be adjusted first, followed by non-load-bearing components. Furthermore, the hydraulic cylinders that achieve overall pose correction can be adjusted first, followed by fine-tuning the hydraulic cylinders corresponding to local deviations. In addition, the number, extension / retraction value, and execution order of each hydraulic cylinder can be labeled in the parameter sequence to ensure that the aluminum formwork accurately approximates the design pose after the hydraulic leveling system executes the actions in sequence.

[0106] In one embodiment, the method further includes:

[0107] S41. Decode the hydraulic cylinder calibration parameter sequence into control commands and send them to the hydraulic leveling system to drive the hydraulic cylinder to adjust the position of the aluminum template.

[0108] S42. During the adjustment process, real-time point cloud data of the aluminum template is collected in real time, and noise reduction processing is performed on the real-time point cloud data to obtain the processed real-time point cloud data.

[0109] S43. Extract the current coordinates of key points of the aluminum template from the processed real-time point cloud data;

[0110] S44. Compare the current coordinates with the design coordinates of the corresponding key points of the aluminum formwork obtained from the BIM design model, and calculate the current residual deviation of each key point of the aluminum formwork.

[0111] S45. Determine whether the current residual deviation is lower than the preset accuracy threshold and obtain the determination result; if the determination result is no, adjust the hydraulic cylinder calibration parameter sequence based on the current residual deviation to obtain the adjusted hydraulic cylinder calibration parameter sequence, and return the adjusted hydraulic cylinder calibration parameter sequence to execute S41; if the determination result is yes, stop the calibration.

[0112] Specifically, the hydraulic cylinder calibration parameter sequence can be stored in digital vector form, containing information such as the cylinder number, target extension / retraction amount, and execution speed. Decoding this digital information converts it into control commands recognizable by the hydraulic leveling system controller. For example, industrial control software can call a communication protocol interface to convert the parameter sequence into a 4-20mA analog electrical signal or an Ethernet digital signal, which is then transmitted via industrial Ethernet. Upon receiving the commands, the hydraulic leveling system controller adjusts the hydraulic pump's output flow based on the target extension / retraction amount, controls the oil circuit of the corresponding hydraulic cylinder by controlling the on / off state of the solenoid valve, and adjusts the solenoid valve opening based on the execution speed parameter to achieve smooth extension / retraction of the hydraulic cylinder, preventing structural deformation or overshooting of the aluminum formwork due to excessively rapid movement. During the adjustment process, a portable laser scanner mounted on a fixed bracket in the construction area can collect real-time point cloud data. The acquisition frequency matches the hydraulic cylinder's movement frequency to ensure synchronous capture of dynamic changes in the aluminum formwork's pose. The scanning range can focus on the key calibration area of ​​the aluminum formwork, avoiding computational redundancy caused by point clouds in irrelevant areas. When preprocessing the real-time point cloud data, a statistical filtering method can be used for noise reduction. This involves defining a neighborhood around each point in the real-time point cloud, calculating the average distance between the points in the neighborhood and the point, calculating the mean and standard deviation of the average distances of all points, removing isolated points whose average distance from the neighborhood exceeds the mean plus a preset multiple of the standard deviation, and retaining the effective point cloud on the surface of the aluminum template. This process yields the processed real-time point cloud data, eliminating the interference of noise introduced by construction dust and equipment vibration on subsequent coordinate extraction.

[0113] Specifically, the current coordinates of key points of the aluminum formwork can be extracted from the processed real-time point cloud data based on feature matching algorithms. For example, a key point detection program (with the same feature extraction logic as in the above embodiment) can be called first to identify the key points of the aluminum formwork corresponding to the BIM design model from the processed real-time point cloud. This involves calculating the curvature and normal vector of the point cloud, filtering points with abrupt curvature changes and normal vector directions that conform to preset features, determining the correspondence of each key point through coordinate matching, and finally outputting the current coordinates of each key point in the preset construction coordinate system to ensure that the extracted coordinates correspond one-to-one with the key points of the BIM design. Subsequently, the three-dimensional Euclidean distance between the current coordinates and the corresponding BIM design coordinates can be calculated to obtain the current residual deviation of a single key point. After summing the residual deviations of all key points to obtain the current residual deviation, this can be used as the core input data for subsequent parameter adjustments. This data can directly reflect the degree of difference between the current pose and the design pose of the aluminum formwork.

[0114] Specifically, it can be determined whether the current residual deviation is lower than a preset accuracy threshold. This preset accuracy threshold can be determined according to industry standards for aluminum formwork construction and is used to define whether the correction meets the standards. If the current residual deviation does not reach the threshold, the adjusted hydraulic cylinder correction parameter sequence can be calculated using a composite control algorithm. This adjusted hydraulic cylinder correction parameter sequence is then returned to execution S41 to continue adjusting the aluminum formwork. The calculation formula can be:

[0115]

[0116]

[0117] in, This represents the adjustment value of the extension / retraction amount of the j-th hydraulic cylinder in the (k+1)-th correction; This represents the proportionality coefficient, used for rapid response to the current residual deviation; This represents the integral coefficient, used to eliminate cumulative bias; This represents the differential coefficient, used to suppress correction overshoot; This represents the residual deviation of the key point of the aluminum template corresponding to the j-th hydraulic cylinder after the k-th correction. ,in The current coordinates after the k-th correction. Design its BIM coordinates; Indicates the time interval between two corrections; Indicates the gradient step size coefficient; Represents the residual deviation after the k-th correction. The gradient vector represents the rate of change of the deviation with respect to the extension and retraction of the hydraulic cylinder. ; The sequence of adjusted hydraulic cylinder calibration parameters after the k-th calibration is denoted as . The vector, Let be the number of hydraulic cylinders, and let the element be the extension / retraction amount of each hydraulic cylinder in the kth iteration, i.e. ; This represents the sequence of hydraulic cylinder extension / retraction adjustments during the (k+1)th correction. The vector, whose elements are the adjustment values ​​of each hydraulic cylinder. .

[0118] Indicative, the calculation is obtained through the above formula. Then, it can be returned to S41 to repeat the calibration process. If the current residual deviation is lower than the preset accuracy threshold, it means that the current construction has met the construction requirements, and the calibration can be stopped. That is, a stop command can be sent to the hydraulic leveling system to shut down the hydraulic pump and solenoid valve.

[0119] Based on the same inventive concept, this application also provides a BIM point cloud-based aluminum formwork construction system for implementing the aforementioned BIM point cloud-based aluminum formwork construction method. The solution provided by this system is similar to the solution described in the above method. Therefore, the specific limitations of one or more BIM point cloud-based aluminum formwork construction system embodiments provided below can be found in the limitations of the BIM point cloud-based aluminum formwork construction method described above, and will not be repeated here.

[0120] In one exemplary embodiment, such as Figure 3 As shown, a BIM point cloud-based aluminum formwork construction system 300 is provided, including:

[0121] The data alignment and model fusion module 301 is used to acquire the BIM design model data of the aluminum formwork and the laser scanning point cloud data of the construction site, and to align the BIM design model data and the laser scanning point cloud data to obtain the BIM design model and the site point cloud model under the same preset construction coordinate system.

[0122] The feature extraction and matching module 302 is used to extract features based on the BIM design model and the site point cloud model respectively, to obtain the BIM feature point set and the site feature point set, and to perform feature matching processing on the BIM feature point set and the site feature point set to obtain a set of matching point pairs used to represent the correspondence between design points and actual points.

[0123] The calibration parameter generation module 303 is used to calculate the spatial deviation of the matching points based on the set of matching point pairs, and generate a hydraulic cylinder calibration parameter sequence in combination with the preset kinematic model of the hydraulic leveling system; the hydraulic cylinder calibration parameter sequence is used to drive the hydraulic leveling system to correct the aluminum template.

[0124] In one exemplary embodiment, the present invention also provides a computer device, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the aluminum formwork construction method based on BIM point clouds of this application. A multi-core processor is preferred to improve the system's parallel processing capability. The memory provides sufficient temporary storage space to support program execution and data processing. The memory capacity should be large enough to accommodate a large amount of data and computational tasks.

[0125] In one exemplary embodiment, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the aluminum formwork construction method based on BIM point clouds of this application. The computer-readable storage medium may include: a read-only memory, a random access memory (RAM), a solid-state drive (SSD), or an optical disc, etc.

[0126] The above-described embodiments are merely illustrative of several implementation methods of the embodiments of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the embodiments of this application, and these modifications and improvements all fall within the protection scope of the embodiments of this application.

Claims

1. A construction method for aluminum formwork based on BIM point clouds, characterized in that, The method includes: The BIM design model data of the aluminum formwork and the laser scanning point cloud data of the construction site are obtained, and the BIM design model data and the laser scanning point cloud data are aligned to obtain the BIM design model and the site point cloud model under the same preset construction coordinate system. Based on the BIM design model and the site point cloud model, feature extraction is performed to obtain the BIM feature point set and the site feature point set. The BIM feature point set and the site feature point set are then subjected to feature matching processing to obtain a set of matching point pairs used to characterize the correspondence between design points and actual points. Based on the set of matching point pairs, the spatial deviation of the matching points is calculated, and combined with the kinematic model of the preset hydraulic leveling system, a hydraulic cylinder correction parameter sequence is generated; the hydraulic cylinder correction parameter sequence is used to drive the hydraulic leveling system to correct the aluminum template.

2. The method according to claim 1, characterized in that, The process of aligning the BIM design model data with the laser-scanned point cloud data to obtain a BIM design model and a site point cloud model in the same construction coordinate system includes: The BIM design model data is processed to extract the design coordinates of the key feature points of the aluminum formwork, and the design coordinates are transformed to the preset construction coordinate system to obtain the BIM design feature point set under the preset construction coordinate system. The laser scanning point cloud data is processed by performing noise reduction, filtering and multi-station registration operations in sequence to obtain the field point cloud. The field point cloud is then converted to the preset construction coordinate system to obtain the field point cloud model under the preset construction coordinate system. The point cloud corresponding to the aluminum template is segmented from the on-site point cloud model to obtain a subset of the on-site point cloud of the aluminum template; Spatially align the BIM design feature point set under the preset construction coordinate system with the aluminum formwork site point cloud subset, and calculate the rigid body transformation matrix that minimizes the spatial position difference between the BIM design feature point set and the aluminum formwork site point cloud subset. The BIM design model data is transformed using the rigid body transformation matrix to obtain the BIM design model in the preset construction coordinate system.

3. The method according to claim 2, characterized in that, Based on the BIM design model and the site point cloud model, feature extraction is performed to obtain a BIM feature point set and a site feature point set, respectively. Feature matching processing is then performed on the BIM feature point set and the site feature point set to obtain a set of matching point pairs representing the correspondence between design points and actual points, including: Extract the key feature points of the aluminum template from the BIM design model, calculate the local geometric feature descriptor of each key feature point, and obtain the BIM feature point set with descriptors; Perform key point detection operation from the field point cloud model, extract candidate feature points, calculate local geometric feature descriptors for each candidate feature point, and obtain the field feature point set with descriptors; Based on the local geometric feature descriptors of each feature point in the BIM feature point set and the site feature point set, nearest neighbor search is used to match each BIM feature point in the BIM feature point set with the site feature points in the site feature point set, forming an initial set of matching pairs. Calculate the descriptor distance ratio of each matching pair in the initial matching pair set, and remove matching pairs whose descriptor distance ratio is greater than a preset ratio threshold to obtain a preliminary filtered matching pair set; A spatial transformation model is established based on the rigid body transformation matrix. The geometric consistency of the preliminary screening matching pair set is verified by using a random sampling consensus algorithm in conjunction with the spatial transformation model. Mismatched pairs that do not conform to the same spatial transformation model are eliminated to obtain the matching point pair set.

4. The method according to claim 1, characterized in that, The step involves calculating the spatial deviation of the matching points based on the set of matching point pairs, and generating a sequence of hydraulic cylinder correction parameters by combining this with a preset kinematic model of the hydraulic leveling system. Based on the design coordinates and actual coordinates of each point pair in the matching point pair set, calculate the three-dimensional position deviation vector of each matching point, combine the BIM feature point set and the site feature point set to calculate the attitude deviation angle of each matching point, and obtain the six-degree-of-freedom pose deviation vector of each matching point. Use the six-degree-of-freedom pose deviation vector as the spatial deviation. Information on aluminum formwork components is extracted from the BIM design model. Cluster analysis is performed on the six-degree-of-freedom pose deviation vectors of each matching point to identify the deviation patterns of the aluminum formwork and evaluate the impact of the deviation patterns on construction to obtain the impact assessment results. Based on the information on the aluminum formwork components and the impact assessment results, the priority of the components requiring correction is determined; Based on the deviation mode, the impact assessment results, and the priority, and combined with the preset kinematic model of the hydraulic leveling system, an optimization problem is constructed with the goal of minimizing residual deviation and with hydraulic cylinder stroke and thrust as constraints. The optimization problem is solved using a sequential quadratic programming algorithm to obtain the extension and retraction of each hydraulic cylinder. The extension and retraction are then arranged in a correction order to form a sequence of correction parameters for the hydraulic cylinders.

5. The method according to claim 1, characterized in that, The method further includes: S41. Decode the hydraulic cylinder calibration parameter sequence into a control command and send it to the hydraulic leveling system to drive the hydraulic cylinder to adjust the position of the aluminum template. S42. During the adjustment process, real-time point cloud data of the aluminum template is collected in real time, and noise reduction processing is performed on the real-time point cloud data to obtain processed real-time point cloud data. S43. Extract the current coordinates of key points of the aluminum template from the processed real-time point cloud data; S44. Compare the current coordinates with the design coordinates of the corresponding aluminum formwork key points obtained from the BIM design model, and calculate the current residual deviation of each aluminum formwork key point; S45. Determine whether the current residual deviation is lower than a preset accuracy threshold and obtain a determination result; if the determination result is negative, adjust the hydraulic cylinder calibration parameter sequence based on the current residual deviation to obtain an adjusted hydraulic cylinder calibration parameter sequence, and return the adjusted hydraulic cylinder calibration parameter sequence to execute S41; if the determination result is positive, stop calibration.

6. The method according to claim 5, characterized in that, The adjusted hydraulic cylinder calibration parameter sequence is calculated using the following formula: in, This represents the adjustment value of the extension / retraction amount of the j-th hydraulic cylinder in the (k+1)-th correction; Indicates the proportionality coefficient; Indicates the integral coefficient; Represents the differential coefficient; This represents the residual deviation of the key point of the aluminum template corresponding to the j-th hydraulic cylinder after the k-th correction. ,in The coordinates are after the k-th correction. The design coordinates; Indicates the time interval between two corrections; Indicates the gradient step size coefficient; Represents the residual deviation after the k-th correction. The gradient vector represents the rate of change of the deviation with respect to the extension and retraction of the hydraulic cylinder. ; The sequence of adjusted hydraulic cylinder calibration parameters for the k-th calibration is denoted as . The vector, This refers to the number of hydraulic cylinders. ; This represents the sequence of adjustments to the extension / retraction of the hydraulic cylinder during the (k+1)th correction. The vector, .

7. The method according to claim 3, characterized in that, The step of performing key point detection operation from the on-site point cloud model and extracting candidate feature points includes: The ISS key point detection algorithm is used to perform key point detection operation on the on-site point cloud model. Based on the preset first and second scaling factors, the neighborhood radius is three times the average spacing of the points in the on-site point cloud model. The scattering matrix of each point cloud in the on-site point cloud model is calculated based on the neighborhood radius, and the eigenvalues ​​of the scattering matrix are solved. Point clouds whose eigenvalues ​​satisfy preset conditions are selected as initial key points; wherein the preset conditions are obtained based on the first scaling factor and the second scaling factor. A non-maximum suppression operation is performed on the initial key points. Based on a preset suppression radius, duplicate initial key points within the preset suppression radius are removed to obtain the candidate feature points.

8. A BIM point cloud-based aluminum formwork construction system, characterized in that, The system includes: The data alignment and model fusion module is used to acquire the BIM design model data of aluminum formwork and the laser scanning point cloud data of the construction site, and to align the BIM design model data and the laser scanning point cloud data to obtain the BIM design model and the site point cloud model under the same preset construction coordinate system. The feature extraction and matching module is used to extract features based on the BIM design model and the site point cloud model respectively to obtain the BIM feature point set and the site feature point set, and to perform feature matching processing on the BIM feature point set and the site feature point set to obtain a set of matching point pairs used to characterize the correspondence between design points and actual points. The calibration parameter generation module is used to calculate the spatial deviation of the matching points based on the set of matching point pairs, and generate a hydraulic cylinder calibration parameter sequence in combination with the preset kinematic model of the hydraulic leveling system; the hydraulic cylinder calibration parameter sequence is used to drive the hydraulic leveling system to correct the aluminum template.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images