Aluminum template installation precision control method, system, device and medium

Abstract

The application relates to an aluminum template installation precision control method, system, device and medium. The method comprises the following steps: extracting the actual local characteristics of the installed aluminum template, matching the actual local characteristics with the predefined key local characteristics, and giving corresponding semantic labels; calculating six-degree-of-freedom deviation parameters and weight coefficients to construct a multi-objective optimization function; based on the function, based on the physical connection relationship and the physical stroke limit value, a mathematical optimization problem is constructed; the problem is numerically solved to obtain the optimal adjustment amount of each aluminum template unit, generate and execute the adjustment instruction, obtain the verification point cloud data of the adjusted aluminum template, and verify and iteratively optimize until the new structured deviation data set meets the preset precision requirement. The method constructs a multi-objective optimization model, combines the physical connection relationship and the stroke limit value, realizes high-precision and intelligent adjustment of the aluminum template installation, significantly improves the precision and efficiency of the aluminum template installation, and enhances the global optimization capability and intelligent level of the adjustment process.

Description

Technical Field

[0001] This invention belongs to the field of computer technology, and in particular relates to methods, systems, equipment and media for controlling the installation accuracy of aluminum formwork. Background Technology

[0002] As a core temporary support structure for the molding of concrete components, the installation accuracy of aluminum formwork directly determines the dimensional accuracy, surface flatness, and verticality of the concrete components, making it a crucial factor affecting the quality and efficiency of construction projects. With the widespread application of intelligent construction and digital twin technologies in the construction field, traditional methods of controlling the installation accuracy of aluminum formwork, relying on manual experience and total station sampling, are no longer sufficient to meet the high-precision, high-efficiency, and intelligent requirements of modern engineering. Traditional methods for controlling the installation accuracy of aluminum formwork have significant limitations. For example, on the one hand, manual adjustment relies on operational experience, resulting in low sensitivity to minor deviations, and discrete point sampling cannot comprehensively reflect the overall installation status of the formwork system, easily leading to problems of inaccurate measurement and imprecise adjustment. On the other hand, the adjustment process lacks systematic planning and does not consider the linkage constraints formed between formwork units through pins and bolts, often resulting in oscillating adjustments by "robbing Peter to pay Paul," making it difficult to achieve optimal global accuracy.

[0003] To overcome the aforementioned technical bottlenecks, the industry has gradually developed digital precision control technology based on 3D scanning. This technology can detect deviations by comparing the 3D point cloud data of the template system with the BIM design model. However, existing digitization methods still have many technical shortcomings: First, using full point cloud registration (such as the ICP algorithm) for deviation calculation not only brings a huge computational burden to millions of point cloud data, resulting in low processing efficiency, but also dilutes the information of key connection parts, boundaries and other core features by a large number of irrelevant surface points in the point cloud, making the system less sensitive to early local minor deviations, and easily forming a dilemma of "data flood" and "information scarcity" coexisting; Second, existing digitization methods can only output the overall deviation field based on point-to-surface distance. This deviation result is only a description of the template installation status and cannot be directly converted into executable adjustment decisions. It cannot answer practical engineering questions such as "which template to adjust, in what order to adjust, and by how much to adjust," and requires secondary interpretation by engineers, which is highly subjective and makes it difficult to guarantee the global optimality of the adjustment plan; Third, the multi-body coupling characteristics of the aluminum template system are not explicitly modeled, and the geometric linkage relationship between template units is ignored. The generated adjustment commands are prone to conflict, resulting in an unstable adjustment process and failure to achieve stable convergence of accuracy. Summary of the Invention

[0004] Therefore, it is necessary to provide methods, systems, equipment, and media for controlling the installation accuracy of aluminum formwork to address the aforementioned technical problems, aiming to improve the accuracy and efficiency of aluminum formwork installation and enhance the intelligence level and overall optimization capability of the aluminum formwork installation process.

[0005] Firstly, this application provides a method for controlling the installation accuracy of aluminum formwork, including:

[0006] S1. Extract key local features from the design model of the aluminum formwork system and assign semantic labels to them to obtain predefined key local features; obtain the original point cloud data of the installed aluminum formwork, extract actual local features based on the original point cloud data, match the actual local features with the predefined key local features to obtain the matched actual local features, assign corresponding semantic labels to the matched actual local features, record the actual geometric parameters of the matched actual local features and the design theoretical geometric parameters of the corresponding matched predefined key local features, and obtain the set of actual features with semantic labels.

[0007] S2. Based on the actual geometric parameters of each feature in the actual feature set with semantic labels and the corresponding theoretical geometric parameters of the design, calculate the six-degree-of-freedom deviation parameters of each aluminum template unit relative to the design pose of the aluminum template unit; based on the semantic labels, query the preset feature importance weight rules and assign corresponding weight coefficients to each six-degree-of-freedom deviation parameter; based on the six-degree-of-freedom deviation parameters of each aluminum template unit and the corresponding weight coefficients, generate a structured deviation dataset.

[0008] S3. Obtain the physical connection relationships of the aluminum formwork system and the physical travel limits of each adjustable support device from the digital twin knowledge base of the aluminum formwork system; construct a multi-objective optimization function based on the adjustment amount of each aluminum formwork unit based on the six-degree-of-freedom deviation parameters and weight coefficients in the structured deviation dataset; generate geometric coordination equality constraints based on the physical connection relationships; generate boundary inequality constraints based on the physical travel limits; combine the multi-objective optimization function, geometric coordination equality constraints, and boundary inequality constraints to obtain the mathematical optimization problem.

[0009] S4. Solve the mathematical optimization problem numerically to obtain the optimal adjustment amount for each aluminum template unit; generate adjustment instructions based on the optimal adjustment amount to instruct the adjustable support equipment to perform corresponding operations; determine the execution order of the adjustment instructions based on the physical connection relationship to obtain the adjustment instruction sequence.

[0010] S5. Send the adjustment instruction sequence to the adjustable support device, execute the adjustment instruction sequence, and obtain the verification point cloud data of the adjusted aluminum template; repeat S1 to S2 based on the verification point cloud data to generate a new structured deviation dataset, determine whether the new structured deviation dataset meets the preset accuracy requirements, and if the new structured deviation dataset does not meet the preset accuracy requirements, repeat S3 to S5 based on the new structured deviation dataset.

[0011] In one embodiment, actual local features are extracted based on the original point cloud data. These actual local features are then matched with predefined key local features to obtain successfully matched actual local features. Semantic labels are assigned to the successfully matched actual local features. The actual geometric parameters of the successfully matched actual local features and the design theoretical geometric parameters of the corresponding predefined key local features are recorded to obtain a set of actual features with semantic labels, including:

[0012] The original point cloud data is subjected to statistical outlier removal to obtain a denoised point cloud; the voxel grid size is set based on the point cloud density of the denoised point cloud, and the voxel grid downsampling process is performed on the denoised point cloud to obtain a preprocessed point cloud.

[0013] Multi-scale geometric analysis is performed on the preprocessed point cloud to calculate the normal vector and curvature of each point. Candidate feature points are selected based on a preset curvature threshold and an angle threshold between the normal vectors of adjacent points.

[0014] Euclidean clustering is performed on the candidate feature points to obtain multiple candidate feature clusters, and the candidate feature clusters are determined as actual local features.

[0015] Calculate the geometric descriptor for each candidate feature cluster and extract the corresponding baseline feature descriptor from the predefined key local features;

[0016] The geometric descriptor of each candidate feature cluster is matched with the baseline feature descriptor for similarity. Stable matching pairs are selected by random sampling consensus algorithm to obtain the successfully matched candidate feature clusters.

[0017] The successfully matched candidate feature clusters are used as the actual local features of the successful match, and the successfully matched actual local features are assigned semantic labels of the corresponding predefined key local features.

[0018] For each successfully matched actual local feature, perform geometric fitting to obtain the corresponding fitting parameters. Extract the centroid coordinates, normal vectors, and fitting geometric parameters of the successfully matched candidate feature clusters as actual geometric parameters. Retrieve the design theoretical geometric parameters of the predefined key local features that are matched.

[0019] By integrating semantic labels, actual geometric parameters, and design theoretical geometric parameters, a set of actual features with semantic labels is obtained.

[0020] In one embodiment, based on the actual geometric parameters of each feature in the semantically labeled actual feature set and the corresponding theoretical geometric parameters of the design, the six-degree-of-freedom deviation parameters of each aluminum template unit relative to the design pose of the aluminum template unit are calculated; based on the semantically labeled query, a preset feature importance weighting rule is used to assign a corresponding weight coefficient to each six-degree-of-freedom deviation parameter; based on the six-degree-of-freedom deviation parameters of each aluminum template unit and the corresponding weight coefficients, a structured deviation dataset is generated, including:

[0021] The topological relationships of the aluminum formwork system are obtained from the digital twin knowledge base of the aluminum formwork system, and the affiliation of the aluminum formwork unit corresponding to each semantic tag is determined based on the topological relationships;

[0022] The actual geometric parameters and theoretical geometric parameters corresponding to the semantic tags of each aluminum template unit are summarized to construct a least squares optimization model;

[0023] The least squares optimization model is solved by singular value decomposition algorithm to obtain the translational and rotational deviation parameters of the aluminum template unit relative to the design pose of the aluminum template unit, which are then combined to form six degrees of freedom deviation parameters.

[0024] Calculate the solution residual of the least squares optimization model and query the preset feature type importance weights based on semantic tags;

[0025] The preset feature type importance weights are weighted and fused with the solution residuals to obtain the weight coefficients corresponding to each six-degree-of-freedom deviation parameter;

[0026] By associating the corresponding six-degree-of-freedom deviation parameters and weighting coefficients with the aluminum template unit identifier, a structured deviation dataset is generated.

[0027] In one embodiment, a multi-objective optimization function based on the six-degree-of-freedom deviation parameters and weighting coefficients in the structured deviation dataset is constructed, based on the adjustment amounts of each aluminum template unit; geometrical compatibility constraints are generated based on physical connection relationships; boundary inequality constraints are generated based on physical travel limits; and the multi-objective optimization function, geometrical compatibility constraints, and boundary inequality constraints are combined to obtain a mathematical optimization problem, including:

[0028] Define decision variables, which consist of translation and rotation adjustment components of each aluminum template unit;

[0029] Based on the linear mapping relationship between the six-degree-of-freedom bias parameters and decision variables, a first objective is constructed. The first objective is the weighted sum of squares of the adjusted residual biases, and the weight matrix of the first objective is composed of the weight coefficients in the structured bias dataset.

[0030] Construct a second objective, which is the sum of the L1 norms of all decision variables;

[0031] By fusing the first and second objectives using preset target weight coefficients, a multi-objective optimization function is obtained;

[0032] Based on the pin connection point information in the physical connection relationship, the displacement continuity condition of the connected aluminum template unit at the connection point is constructed, and the displacement continuity condition is transformed into a linear equation system based on decision variables. The linear equation system is used as a geometric compatibility equality constraint.

[0033] Based on the physical travel limit of the adjustable support equipment, the upper and lower boundary value ranges of each decision variable are determined, forming boundary inequality constraints;

[0034] By integrating multi-objective optimization functions, geometric compatibility equality constraints, and boundary inequality constraints, a mathematical optimization problem is obtained.

[0035] In one embodiment, a mathematical optimization problem is numerically solved to obtain the optimal adjustment amount for each aluminum template unit; adjustment instructions are generated based on the optimal adjustment amounts to instruct the adjustable support equipment to perform corresponding operations; the execution order of the adjustment instructions is determined based on the physical connection relationship to obtain an adjustment instruction sequence, including:

[0036] The interior point method or the alternating direction multiplier method is used to numerically solve the mathematical optimization problem, output the optimal solution of the decision variables, and use the optimal solution of the decision variables as the optimal adjustment amount of each aluminum template element.

[0037] The minimum movement step of the adjustable support equipment is obtained, and the optimal adjustment amount is discretized and rounded to obtain the basic adjustment amount that adapts to the adjustable support equipment.

[0038] Based on the basic adjustment amount, an adjustment instruction is generated that includes the adjustment direction, adjustment magnitude, and target unit identifier;

[0039] The support dependencies of aluminum formwork units are extracted from the digital twin knowledge base of the aluminum formwork system, and the adjustment priority is determined based on the support dependencies;

[0040] The adjustment instructions are sorted according to their adjustment priority to obtain the adjustment instruction sequence.

[0041] In one embodiment, it is determined whether the new structured bias dataset meets a preset accuracy requirement. If the new structured bias dataset does not meet the preset accuracy requirement, then S3 to S5 are repeated based on the new structured bias dataset, including:

[0042] Obtain the preset accuracy threshold, which includes the critical feature deviation threshold and the ordinary feature deviation threshold;

[0043] Extract the six-degree-of-freedom deviation parameters of each aluminum template unit in the new structured deviation dataset, and compare the six-degree-of-freedom deviation parameters with the key feature deviation threshold and the ordinary feature deviation threshold respectively to obtain the comparison results;

[0044] If the comparison results show that the six degrees of freedom deviation parameters corresponding to the preset key features of each aluminum template unit are all less than the key feature deviation threshold, and the six degrees of freedom deviation parameters corresponding to the non-preset key features are all less than the ordinary feature deviation threshold, then the new structured deviation dataset is determined to meet the preset accuracy requirements.

[0045] If the comparison result shows that the six degrees of freedom deviation parameter corresponding to the preset key feature of the aluminum template unit is not less than the key feature deviation threshold, or the six degrees of freedom deviation parameter corresponding to the non-preset key feature is not less than the ordinary feature deviation threshold, then the new structured deviation dataset does not meet the preset accuracy requirements.

[0046] Using the new structured bias dataset as input, re-execute S3 to S5 until the generated real-time structured bias dataset meets the preset accuracy requirements.

[0047] In one embodiment, the mathematical expression of the multi-objective optimization function is:

[0048]

[0049] in, This is the adjustment vector for the aluminum template element. The total number of units; This represents the six-degree-of-freedom adjustment amount for the k-th aluminum template element; This is the concatenated vector of the initial six-degree-of-freedom deviation parameters; This is a diagonal weight matrix, where the diagonal elements are weight coefficients; All are target weight coefficients; This refers to the physical connection edge set of the aluminum formwork unit. and Index of adjacent cells; This represents the coupling coefficient between adjacent units.

[0050] Secondly, this application also provides an aluminum formwork installation accuracy control system, including:

[0051] The feature extraction and matching module is used to perform step 1, which includes: pre-extracting key local features from the design model of the aluminum formwork system and assigning semantic labels to them to obtain predefined key local features; acquiring the original point cloud data of the installed aluminum formwork, extracting actual local features based on the original point cloud data, matching the actual local features with the predefined key local features to obtain successfully matched actual local features, assigning corresponding semantic labels to the successfully matched actual local features, recording the actual geometric parameters of the successfully matched actual local features and the design theoretical geometric parameters of the corresponding matched predefined key local features, and obtaining a set of actual features with semantic labels.

[0052] The deviation calculation and data generation module is used to execute step 2, which includes: calculating the six-degree-of-freedom deviation parameters of each aluminum template unit relative to the design pose of the aluminum template unit based on the actual geometric parameters of each feature in the actual feature set with semantic labels and the corresponding design theoretical geometric parameters; querying the preset feature importance weight rules based on semantic labels and assigning corresponding weight coefficients to each six-degree-of-freedom deviation parameter; and generating a structured deviation dataset based on the six-degree-of-freedom deviation parameters of each aluminum template unit and the corresponding weight coefficients.

[0053] The optimization problem construction module is used to execute step 3, which includes: obtaining the physical connection relationship of the aluminum formwork system and the physical travel limit of each adjustable support device from the digital twin knowledge base of the aluminum formwork system; constructing a multi-objective optimization function based on the adjustment amount of each aluminum formwork unit based on the six-degree-of-freedom deviation parameters and weight coefficients in the structured deviation dataset; generating geometric coordination equality constraints based on the physical connection relationship; generating boundary inequality constraints based on the physical travel limit; and combining the multi-objective optimization function, geometric coordination equality constraints, and boundary inequality constraints to obtain the mathematical optimization problem.

[0054] The adjustment amount calculation and instruction generation module is used to execute step 4, which includes: numerically solving the mathematical optimization problem to obtain the optimal adjustment amount of each aluminum template unit; generating adjustment instructions based on the optimal adjustment amount to instruct the adjustable support equipment to perform corresponding operations; and determining the execution order of the adjustment instructions based on the physical connection relationship to obtain the adjustment instruction sequence.

[0055] The adjustment execution and verification module is used to execute step 5, which includes: sending the adjustment instruction sequence to the adjustable support device, executing the adjustment instruction sequence, and obtaining the verification point cloud data of the adjusted aluminum template; repeating steps 1 to 2 based on the verification point cloud data to generate a new structured deviation dataset, determining whether the new structured deviation dataset meets the preset accuracy requirements, and if the new structured deviation dataset does not meet the preset accuracy requirements, repeating steps 3 to 5 based on the new structured deviation dataset.

[0056] 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.

[0057] 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.

[0058] The aforementioned method, system, equipment, and medium for controlling the installation accuracy of aluminum formwork first extracts predefined key local features from the design model. These features are then combined with raw point cloud data to extract and match actual local features, forming a feature set with semantic labels. This solves the problem of redundancy in massive point cloud information and lays the foundation for accurate deviation calculation. Secondly, six-degree-of-freedom deviation parameters are calculated based on feature geometric parameters and assigned weights to generate a structured deviation dataset. This overcomes the lack of priority differentiation in traditional deviation quantification, improving the engineering practicality of deviation description. Furthermore, by integrating physical connection relationships and equipment travel limits, a multi-objective optimization function and constraints are constructed, forming a mathematical optimization problem. This overcomes the deficiency of neglecting system linkage constraints in adjustment schemes, ensuring the global optimality of adjustment decisions. Finally, an ordered sequence of adjustment instructions is generated through numerical solution. Combined with closed-loop verification and iterative optimization, this solves the problems of traditional adjustment lacking feedback mechanisms and unstable accuracy convergence, significantly improving the installation accuracy and construction efficiency of aluminum formwork. Attached Figure Description

[0059] 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.

[0060] Figure 1 A flowchart of an aluminum template installation accuracy control method provided as an exemplary embodiment of the present invention;

[0061] Figure 2 A flowchart of a method for obtaining an adjustment instruction sequence is provided as an exemplary embodiment of the present invention;

[0062] Figure 3 A schematic diagram of an aluminum template installation accuracy control system provided as an exemplary embodiment of the present invention. Detailed Implementation

[0063] 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.

[0064] In one embodiment, such as Figure 1 As shown, a method for controlling the installation accuracy of aluminum formwork is provided. This embodiment illustrates the application of this method 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:

[0065] S101: Extract key local features from the design model of the aluminum formwork system and assign semantic labels to them to obtain predefined key local features; acquire the original point cloud data of the installed aluminum formwork, extract actual local features based on the original point cloud data, match the actual local features with the predefined key local features to obtain successfully matched actual local features, assign corresponding semantic labels to the successfully matched actual local features, record the actual geometric parameters of the successfully matched actual local features and the design theoretical geometric parameters of the corresponding matched predefined key local features, and obtain a set of actual features with semantic labels.

[0066] Specifically, based on the 3D geometric information of the design model and combined with the precision requirements of components in building construction codes, irrelevant surface details are eliminated, retaining only key local features with clear engineering functions. Each key local feature is given a unique identifier, establishing a binding relationship between the feature and engineering semantics, ensuring that subsequent deviation analysis can directly relate to the quality requirements of specific engineering parts. Then, by acquiring the original point cloud data of the installed aluminum formwork, the 3D spatial state of the formwork after actual installation can be comprehensively captured. Extracting actual local features from the original point cloud data allows for the selection of geometric entities corresponding to predefined key local features in the design model, avoiding computational redundancy caused by full point cloud processing. Matching the actual local features with the predefined key local features establishes a correspondence through geometric feature consistency checks (such as shape, size, and relative positional relationships), ensuring that each actual feature is accurately associated with the design benchmark, thereby achieving a direct comparison of "actual state - design intent". Subsequently, corresponding semantic labels can be assigned to the successfully matched local features, transferring the engineering semantics from the design phase to the actual state data. This transforms subsequent deviation analysis from a simple geometric difference calculation into an engineering-oriented quality diagnosis, allowing for the recording of actual geometric parameters and theoretical design geometric parameters to quantify their differences. By integrating the above data, a set of semantically labeled actual features can be obtained, laying the foundation for subsequent accurate deviation quantification.

[0067] S102: Based on the actual geometric parameters of each feature in the actual feature set with semantic labels and the corresponding theoretical geometric parameters of the design, calculate the six-degree-of-freedom deviation parameters of each aluminum template unit relative to the design pose of the aluminum template unit; based on the semantic labels, query the preset feature importance weight rules and assign corresponding weight coefficients to each six-degree-of-freedom deviation parameter; based on the six-degree-of-freedom deviation parameters of each aluminum template unit and the corresponding weight coefficients, generate a structured deviation dataset.

[0068] Specifically, since deviations during aluminum formwork installation include not only positional offsets but also rotational deviations caused by support tilting and splicing misalignment, and positional deviations alone cannot fully reflect the actual installation state of the formwork, a rigid body transformation model can be used based on the actual feature set with semantic tags. By minimizing the sum of squares of feature-level geometric differences, the translational deviations (in the X, Y, and Z directions) and rotational deviations (angles around the X, Y, and Z axes) of each aluminum formwork unit relative to its designed pose can be solved, yielding six-degree-of-freedom deviation parameters. These six-degree-of-freedom parameters can completely characterize the pose differences of the formwork, ensuring the comprehensiveness and accuracy of the deviation description. Furthermore, the impact of features corresponding to different semantic tags on installation accuracy varies in engineering practice. For example, deviations in load-bearing boundary features directly affect the load-bearing capacity of concrete components, and their weight should be higher than that of decorative feature deviations in non-load-bearing areas. Therefore, preset feature importance weight rules can be queried based on semantic tags, and corresponding weight coefficients can be assigned to the six-degree-of-freedom deviation parameters. The pre-defined feature importance weighting rules are based on building construction specifications, concrete forming quality requirements, and engineering practice experience. By associating semantic tags with the weighting rules, deviation parameters can acquire engineering priority attributes, avoiding indiscriminate treatment of all deviations during subsequent optimization, thereby improving the targeting and engineering applicability of optimization decisions. Dispersed feature-level deviations, element-level pose information, and engineering priority weights are integrated to generate a structured deviation dataset. This dataset provides a unified and structured input for subsequent optimization models, enabling optimization algorithms to directly identify the priority and impact of different element deviations.

[0069] S103: Obtain the physical connection relationships of the aluminum formwork system and the physical travel limits of each adjustable support device from the digital twin knowledge base of the aluminum formwork system; construct a multi-objective optimization function based on the adjustment amount of each aluminum formwork unit based on the six-degree-of-freedom deviation parameters and weight coefficients in the structured deviation dataset; generate geometric coordination equality constraints based on the physical connection relationships; generate boundary inequality constraints based on the physical travel limits; combine the multi-objective optimization function, geometric coordination equality constraints, and boundary inequality constraints to obtain the mathematical optimization problem.

[0070] Specifically, the digital twin knowledge base stores detailed information about the aluminum formwork system, including the physical connections between formwork units (such as connections via pins and bolts) and the physical travel limits of each adjustable support device. This further restricts the adjustment range and interrelationships of the aluminum formwork units. If the physical connections of pins and bolts are ignored, the generated adjustment commands may lead to problems such as misalignment and stress concentration at the connection points. Furthermore, the physical travel limits of the adjustable support devices are objective constraints on the adjustment actions; adjustment commands exceeding these limits cannot be executed in engineering practice. Subsequently, a multi-objective optimization function based on the six-degree-of-freedom deviation parameters and weighting coefficients in the structured deviation dataset can be constructed. The objective of this optimization function is to minimize the deviation of the aluminum formwork units while satisfying the physical connection relationships and physical travel limits. Simultaneously, geometrical compatibility constraints can be generated based on the physical connection relationships to ensure that the relative positional relationships between formwork units meet design requirements, and boundary inequality constraints can be generated based on the physical travel limits to restrict the adjustment range of the adjustable support devices. By combining multi-objective optimization functions, geometric compatibility equality constraints, and boundary inequality constraints, a mathematical optimization problem is formed. This problem can transform the engineering requirements (accuracy, efficiency, and feasibility) for the precision control of aluminum formwork installation into a standardized mathematical model. This model includes both the optimization objectives of quality and efficiency and the objective constraints of system coupling and equipment performance, providing a clear and rigorous computational framework for subsequent numerical solutions.

[0071] S104: Numerically solve the mathematical optimization problem to obtain the optimal adjustment amount for each aluminum template unit; generate adjustment instructions based on the optimal adjustment amount to instruct the adjustable support equipment to perform corresponding operations; determine the execution order of the adjustment instructions based on the physical connection relationship to obtain the adjustment instruction sequence.

[0072] Specifically, by solving mathematical optimization problems using algorithms such as the interior-point method, the alternating direction multiplier method, and genetic algorithms, it is possible to find the adjustment amount that minimizes the objective function under given constraints. Once the optimal adjustment amount is obtained, adjustment instructions can be generated to instruct the adjustable support equipment to perform operations. These instructions can include the adjustment direction and amount for each adjustable support equipment. Furthermore, based on the physical connection relationships of the aluminum formwork system, the execution order of the adjustment instructions can be determined to form an adjustment instruction sequence, ensuring the coordination and stability of the adjustment process and avoiding local conflicts or overall accuracy degradation caused by improper adjustment order.

[0073] S105: Send the adjustment instruction sequence to the adjustable support device, execute the adjustment instruction sequence, and obtain the verification point cloud data of the adjusted aluminum template; repeat S101 to S102 based on the verification point cloud data to generate a new structured deviation dataset, determine whether the new structured deviation dataset meets the preset accuracy requirements, and if the new structured deviation dataset does not meet the preset accuracy requirements, repeat S103 to S105 based on the new structured deviation dataset.

[0074] Specifically, sending the adjustment command sequence to the adjustable support equipment and executing it translates optimization decisions into actual construction operations. Then, focusing on the adjusted formwork units and related areas, verification point cloud data of the adjusted aluminum formwork is obtained. Based on the verification point cloud data, S101 to S102 are repeated, using the same feature extraction and deviation quantification logic to accurately diagnose the adjusted state and generate a new structured deviation dataset that is comparable to the dataset before adjustment. The preset accuracy requirements are set based on the aluminum formwork installation accuracy acceptance standards specified in building construction codes, clearly defining the deviation thresholds for key and ordinary features. If the new structured deviation dataset does not meet the preset accuracy requirements, S103 to S105 can be repeated based on this new dataset for iterative optimization. Through a feedback mechanism, deviations are continuously corrected, thereby improving the stability and reliability of installation accuracy and ultimately achieving precise control of aluminum formwork installation accuracy.

[0075] The aforementioned method first achieves accurate identification and matching of aluminum formwork system features by extracting key local features and assigning semantic labels. Secondly, it generates a structured deviation dataset based on semantic labels and weighting rules, addressing the problem that existing digital methods cannot directly translate deviation results into adjustment decisions and suffer from strong subjectivity. Furthermore, by constructing a multi-objective optimization function and constraints, it generates a globally optimal adjustment scheme, resolving the instability in accuracy caused by the lack of systematic planning and failure to consider linkage constraints in traditional methods. Finally, through repeated verification and optimization of the adjustment process, it ensures that the installation accuracy meets requirements, improving the intelligence level and global optimization capability of aluminum formwork installation accuracy control.

[0076] In one embodiment, actual local features are extracted based on the original point cloud data. These actual local features are then matched with predefined key local features to obtain successfully matched actual local features. Semantic labels are assigned to the successfully matched actual local features. The actual geometric parameters of the successfully matched actual local features and the design theoretical geometric parameters of the corresponding predefined key local features are recorded to obtain a set of semantically labeled actual features, including:

[0077] The original point cloud data is subjected to statistical outlier removal to obtain a denoised point cloud; the voxel grid size is set based on the point cloud density of the denoised point cloud, and the voxel grid downsampling process is performed on the denoised point cloud to obtain a preprocessed point cloud.

[0078] Multi-scale geometric analysis is performed on the preprocessed point cloud to calculate the normal vector and curvature of each point. Candidate feature points are selected based on a preset curvature threshold and an angle threshold between the normal vectors of adjacent points.

[0079] Euclidean clustering is performed on the candidate feature points to obtain multiple candidate feature clusters, and the candidate feature clusters are determined as actual local features; the geometric descriptor of each candidate feature cluster is calculated, and the corresponding baseline feature descriptor is extracted from the predefined key local features;

[0080] The geometric descriptor of each candidate feature cluster is matched with the baseline feature descriptor for similarity. Stable matching pairs are selected by random sampling consensus algorithm to obtain the successfully matched candidate feature clusters. The successfully matched candidate feature clusters are used as the actual local features of the successful match, and the corresponding predefined key local features are assigned semantic labels to the actual local features of the successful match.

[0081] For each successfully matched actual local feature, perform geometric fitting to obtain the corresponding fitting parameters. Extract the centroid coordinates, normal vectors, and fitting geometric parameters of the successfully matched candidate feature clusters as actual geometric parameters. Retrieve the design theoretical geometric parameters of the predefined key local features that are matched.

[0082] By integrating semantic labels, actual geometric parameters, and design theoretical geometric parameters, a set of actual features with semantic labels is obtained.

[0083] Specifically, when performing statistical outlier removal on raw point cloud data, isolated points introduced by factors such as measurement noise and environmental interference can be identified and eliminated based on the spatial distribution characteristics of the point cloud data. For example, first, select the K nearest neighbors for each point in the raw point cloud data, calculate the average distance from the point to these K nearest neighbors, and then calculate the mean and standard deviation of the average distances of all points. Points whose average distance is greater than the sum of the mean and a preset multiple of the standard deviation are identified as outliers and removed. This denoising method can effectively preserve the overall geometric structure of the point cloud, eliminate the impact of random noise on the point cloud quality, and provide a reliable data foundation for subsequent feature extraction. Furthermore, the voxel grid size can be set based on the point cloud density of the denoised point cloud to balance the needs of data reduction and geometric feature preservation during downsampling. For example, the voxel grid size can be positively correlated with the average point spacing of the denoised point cloud. Subsequently, when performing voxel grid downsampling on the denoised point cloud, the space of the denoised point cloud can be divided into multiple regular voxel grids. The centroid coordinates of all points within each voxel are calculated, and these centroid coordinates are used to replace all the original points within the voxel. This results in a preprocessed point cloud with a significantly reduced data volume while retaining key geometric features, thereby reducing the complexity of subsequent calculations and improving overall processing efficiency.

[0084] Specifically, preprocessed point clouds can capture the geometric feature responses of point clouds at different scales, avoiding feature omissions or mis-extractions caused by single-scale analysis. For example, when calculating the normal vector of each point, a local neighborhood is first constructed for each point in the preprocessed point cloud. The covariance matrix is ​​calculated based on the coordinates of points within the local neighborhood. Eigenvalue decomposition is then performed on the covariance matrix, and the eigenvector corresponding to the smallest eigenvalue is the normal vector of that point. This method ensures the accuracy of normal vector calculation through statistical analysis of local geometric information. Curvature calculation can be derived based on the eigenvalues ​​of the covariance matrix. For example, if the three eigenvalues ​​of the covariance matrix are... Then curvature ,in It reflects the degree of dispersion of the local neighborhood in the direction of minimum change. and Corresponding to the dispersion in the other two directions, the curvature value characterizes the geometric convexity of the point's location. A larger curvature indicates that the point is more likely to be located in a feature region such as an edge or corner. After calculation, candidate feature points can be screened based on a preset curvature threshold and a threshold for the angle between the normal vectors of adjacent points. Specifically, points with curvature greater than the preset curvature threshold are initially identified as potential feature points. Then, the angle between the normal vectors of these potential feature points and their adjacent points is calculated, and points with angles greater than the preset angle threshold are retained as candidate feature points. Through this dual screening of curvature and normal vector angles, feature points with significant geometric changes can be accurately located, providing core seed points for the subsequent formation of feature clusters.

[0085] Furthermore, Euclidean clustering can be applied to candidate feature points. This grouping is based on the Euclidean distance between spatial points, resulting in smaller spatial distances within the same cluster and larger spatial distances between points in different clusters. For example, an Euclidean distance threshold can be set first. An unlabeled point can be selected as the initial seed point from the candidate feature point set. All adjacent candidate feature points whose Euclidean distance to the seed point is less than the threshold can be included in the same cluster. This process is repeated with the newly included point as the new seed point until no new points can be included in the cluster. Then, the next unlabeled point is selected as the new seed point to start a new clustering. This process is repeated until all candidate feature points are labeled, resulting in multiple candidate feature clusters. Since each candidate feature cluster consists of a set of spatially closely related points with significant geometric features, it can completely correspond to the geometric shape of predefined key local features in the design model, such as the circular feature cluster corresponding to pin holes and the linear feature cluster corresponding to splicing seams. Therefore, the candidate feature clusters can be identified as actual local features to ensure the consistency of the actual local features with the predefined key local features in terms of geometric structure.

[0086] Subsequently, Fast Point Feature Histogram (FPFH) can be used to calculate the geometric descriptor for each candidate feature cluster, transforming the geometric information of the feature cluster into a quantifiable and comparable vector form to achieve accurate matching with predefined key local features. For example, firstly, based on the normal vector and local neighborhood information of each point in the candidate feature cluster, the geometric relationships such as azimuth, polar angle, and distance between point pairs are calculated. Then, these geometric relationships are statistically analyzed using a histogram and normalized to finally form a high-dimensional feature vector. This descriptor has good rotation invariance and scale robustness, effectively resisting slight deformations and noise interference in point cloud data. Furthermore, the same calculation method as for the candidate feature cluster geometric descriptor can be used to extract the corresponding baseline feature descriptor from the predefined key local features, thus ensuring the comparability of the two in the feature space. For example, when the predefined key local feature is a circular pin hole in a design model, the FPFH descriptor can also be calculated based on its geometric data as the baseline feature descriptor. When performing similarity matching between the geometric descriptors of each candidate feature cluster and the baseline feature descriptors, cosine similarity can be used as the matching metric. The value of cosine similarity ranges from -1 to 1, with a value closer to 1 indicating a higher similarity between the two descriptors. For illustration, a similarity threshold can be set, and descriptor pairs with a cosine similarity greater than this threshold can be considered as initial matching pairs.

[0087] Specifically, stable matching pairs are screened using a random sampling consensus algorithm. This involves eliminating erroneous matches based on a geometric transformation model. For example, assuming a rigid transformation relationship between the candidate feature cluster and predefined key local features, a small number of sample points are randomly selected from the initial matching pairs to estimate the transformation matrix parameters. Then, the reprojection error of all initial matching pairs under this transformation matrix is ​​calculated. Matching pairs with reprojection errors less than a preset error threshold are identified as interior points. After multiple iterations, the interior point corresponding to the transformation matrix with the most interior points is selected as the stable matching pair, ultimately yielding successfully matched candidate feature clusters. This screening process effectively eliminates erroneous matches caused by descriptor similarity, ensuring the stability and accuracy of the matching results. A stable matching pair indicates that the candidate feature cluster and the predefined key local features are highly consistent in geometric structure and feature representation, accurately reflecting the actual installation state of the predefined key local features. Therefore, a successfully matched candidate feature cluster can be considered as a successfully matched actual local feature. Subsequently, the association between actual features and design semantics can be established through matching relationships. For example, if the semantic label of a successfully matched predefined key local feature is "exterior wall panel A - group 2 pin holes", then the semantic label is assigned to the corresponding successfully matched actual local feature, so that the actual local feature has clear engineering semantic attributes, providing specific engineering directions for subsequent deviation analysis.

[0088] Furthermore, geometric fitting can be performed on each successfully matched local feature to transform the discrete point cloud feature clusters into a regular geometric model, thereby accurately extracting their geometric parameters. The fitting process can employ the least squares method; for example, for a circular feature cluster (such as a pin hole), a circular equation can be fitted. ,in Let r be the coordinates of the circle's center and r be the radius. The parameters are solved by minimizing the sum of the squared distances from all points in the feature cluster to the circle. , And r. For planar feature clusters (such as template splicing surfaces), the fitted plane equation is ax + by + cz + d = 0, where (a, b, c) are the plane normal vector components, and d is the distance from the plane to the origin. The parameters a, b, c, and d are solved by minimizing the sum of the squared distances from all points in the feature cluster to the plane, and these parameters are the fitted parameters. In addition, the centroid coordinates, normal vectors, and fitted parameters of successfully matched candidate feature clusters can be extracted as actual geometric parameters. The centroid coordinates are obtained by calculating the average of the coordinates of all points in the feature cluster, and the normal vector is the normal vector of the fitted geometric model (such as the normal vector of the fitted plane, the axial direction vector of the fitted cylinder). These parameters completely characterize the geometric state of the actual local features from different dimensions. Furthermore, the design theoretical geometric parameters of the corresponding matched predefined key local features can be retrieved, that is, parameters of the same type as the actual geometric parameters are extracted from the design model. For example, the design theoretical geometric parameters include the design centroid coordinates, design normal vectors, and design fitted parameters (such as design radius, design plane equation parameters), ensuring the direct comparability between the actual geometric parameters and the design theoretical geometric parameters.

[0089] By integrating semantic tags, actual geometric parameters, and theoretical geometric parameters, the three types of information corresponding to each successfully matched actual local feature are associated and stored to form a structured data unit, resulting in a set of actual features with semantic tags. This set achieves a deep binding between engineering semantics, actual geometric state, and design geometric benchmarks. This provides accurate geometric data support for subsequent pose deviation calculations and enables deviation analysis to be directly linked to specific engineering parts and quality requirements, ensuring the relevance and effectiveness of subsequent steps.

[0090] In one embodiment, based on the actual geometric parameters of each feature in the semantically labeled actual feature set and the corresponding theoretical geometric parameters of the design, the six-degree-of-freedom deviation parameters of each aluminum template unit relative to the design pose of the aluminum template unit are calculated; based on the semantically labeled query, a preset feature importance weighting rule is used to assign a corresponding weight coefficient to each six-degree-of-freedom deviation parameter; based on the six-degree-of-freedom deviation parameters of each aluminum template unit and the corresponding weight coefficients, a structured deviation dataset is generated, including:

[0091] The topological relationships of the aluminum formwork system are obtained from the digital twin knowledge base of the aluminum formwork system, and the affiliation of the aluminum formwork unit corresponding to each semantic tag is determined based on the topological relationships;

[0092] The actual geometric parameters and theoretical geometric parameters corresponding to the semantic tags of each aluminum template unit are summarized to construct a least squares optimization model;

[0093] The least squares optimization model is solved by singular value decomposition algorithm to obtain the translational and rotational deviation parameters of the aluminum template unit relative to the design pose of the aluminum template unit, which are then combined to form six degrees of freedom deviation parameters.

[0094] Calculate the solution residual of the least squares optimization model and query the preset feature type importance weights based on semantic tags;

[0095] The preset feature type importance weights are weighted and fused with the solution residuals to obtain the weight coefficients corresponding to each six-degree-of-freedom deviation parameter;

[0096] By associating the corresponding six-degree-of-freedom deviation parameters and weighting coefficients with the aluminum template unit identifier, a structured deviation dataset is generated.

[0097] Specifically, the topological relationships of the aluminum formwork system can be obtained from the digital twin knowledge base of the aluminum formwork system. These topological relationships include the unit division rules of the aluminum formwork system, the spatial distribution of each formwork unit, and the engineering area information covered by each formwork unit. They also record the correspondence between predefined key local features in the design model and the formwork units. Based on this topological relationship, the affiliation of each semantic tag to an aluminum formwork unit can be determined. That is, by matching the engineering part information associated with the semantic tag to the coverage area of ​​the formwork unit in the topological relationship, each actual feature with a semantic tag is mapped to its corresponding aluminum formwork unit, ensuring that feature data can be summarized by formwork unit. Subsequently, based on the rigid body transformation characteristics of the aluminum formwork unit—that is, the pose change of the formwork unit after installation can be equivalent to a rigid body transformation (containing only translation and rotation, without shape deformation)—the actual geometric parameters and design theoretical geometric parameters corresponding to the semantic tags of each aluminum formwork unit can be summarized. By solving for the optimal rigid body transformation, the difference between the geometric parameters of the actual features and the design theoretical geometric parameters can be further quantified. For example, let the rigid body transformation of the k-th aluminum formwork unit be... (Including translation vectors) With rotation matrix The location point in the actual geometric parameters of the i-th feature associated with this unit is... The corresponding position point in the design theoretical geometric parameters is Then the objective function of the least squares optimization model is:

[0098]

[0099] in Let be the total number of features associated with the k-th aluminum template unit. The core of this model is to find the translation and rotation parameters that minimize the sum of the squared distances between the actual feature points and the design feature points after rigid body transformation, thereby characterizing the pose deviation of the template unit.

[0100] The least squares optimization model can then be solved using the singular value decomposition algorithm, that is, first calculate the centroid of the actual feature point associated with the k-th template unit. and the center of gravity of the design feature points Subsequently, the actual feature points and the designed feature points were decentralized to obtain... , It can also calculate the covariance matrix of the decentralized point set. ,in This is a transpose operation. For the covariance matrix... Perform singular value decomposition to obtain ,in , It is an orthogonal matrix. It is a diagonal matrix. Rotation matrix. Depend on If Then for (Recalculate after inverting the last column), translation vector Depend on The rotation parameters around the X, Y, and Z axes corresponding to the rotation matrix are combined with the X, Y, and Z components of the translation vector to form the six-degree-of-freedom deviation parameters of the aluminum template element.

[0101] Specifically, the solution residuals of the least squares optimization model can also be calculated. These residuals are the Euclidean distances between each feature point after rigid body transformation and the designed feature points, i.e., the residuals corresponding to the i-th feature of the k-th element. The magnitude of the residual reflects the matching accuracy between the actual geometric parameters of the feature point and the theoretical geometric parameters of the design. Furthermore, it allows querying preset feature type importance weights based on semantic tags. These weights are preset coefficients corresponding to the feature type, denoted as... ,in This represents the feature type (such as pin hole, splicing surface, etc.) corresponding to the semantic label associated with the i-th six-degree-of-freedom deviation parameter. Key feature types corresponding to The values ​​are larger, while the values ​​corresponding to ordinary feature types are smaller.

[0102] By weighting and fusing the preset feature type importance weights with the solved residuals, the weight coefficients corresponding to each six-degree-of-freedom deviation parameter can be obtained. The fusion process can be achieved using the following formula:

[0103]

[0104] in, The weighting coefficient corresponding to the i-th six-degree-of-freedom deviation parameter of the k-th aluminum template element; The feature type corresponding to the semantic label associated with the i-th six-degree-of-freedom deviation parameter, such as pin hole, splicing surface, inside and outside corner, etc. The preset feature type importance weights, and Key feature types The value can be ≥0.7, while for ordinary features it is ≤0.3; Solve the residual by least squares for the i-th feature of the k-th aluminum template element; For the first The aluminum formwork unit Solve the residuals using least squares for each feature; The residual normalization coefficient is set based on the design accuracy threshold of the aluminum template system. The feature reliability factor is determined by the ratio of the number of successfully matched point clouds to the total number of points for that feature. ; For the first The formula calculates the total number of features associated with each aluminum template unit. It penalizes the residuals with an exponential term; the smaller the residual, the closer the exponential term is to 1. Combined with the feature type importance weight and feature reliability factor, the final weight coefficients reflect both the engineering importance of the feature type and the matching accuracy of the residuals and the reliability of the features. This allows subsequent optimization processes to prioritize deviation parameters with high importance and good matching accuracy.

[0105] Finally, the aluminum template unit identifier can be used as an index, and each index can store the six degrees of freedom deviation parameters of the unit and the weight coefficients corresponding to each parameter. At the same time, the semantic label information of the unit can be associated to form a unified and structured data set, which provides standardized input data for the subsequent construction of multi-objective optimization models.

[0106] In one embodiment, a multi-objective optimization function based on the six-degree-of-freedom deviation parameters and weighting coefficients in the structured deviation dataset is constructed, based on the adjustment amounts of each aluminum template unit; geometrical compatibility constraints are generated based on physical connection relationships; boundary inequality constraints are generated based on physical travel limits; and the multi-objective optimization function, geometrical compatibility constraints, and boundary inequality constraints are combined to obtain a mathematical optimization problem, including:

[0107] Define decision variables, which consist of translation and rotation adjustment components of each aluminum template unit;

[0108] Based on the linear mapping relationship between the six-degree-of-freedom bias parameters and decision variables, a first objective is constructed. The first objective is the weighted sum of squares of the adjusted residual biases, and the weight matrix of the first objective is composed of the weight coefficients in the structured bias dataset.

[0109] Construct a second objective, which is the sum of the L1 norms of all decision variables;

[0110] By fusing the first and second objectives using preset target weight coefficients, a multi-objective optimization function is obtained;

[0111] Based on the pin connection point information in the physical connection relationship, the displacement continuity condition of the connected aluminum template unit at the connection point is constructed, and the displacement continuity condition is transformed into a linear equation system based on decision variables. The linear equation system is used as a geometric compatibility equality constraint.

[0112] Based on the physical travel limit of the adjustable support equipment, the upper and lower boundary value ranges of each decision variable are determined, forming boundary inequality constraints;

[0113] By integrating multi-objective optimization functions, geometric compatibility equality constraints, and boundary inequality constraints, a mathematical optimization problem is obtained.

[0114] Specifically, the decision variable consists of the translational and rotational adjustment components of each aluminum template element, which is a vector formed by concatenating the six degrees of freedom adjustment values ​​of all aluminum template elements, denoted as . ,Right now Where N is the total number of aluminum formwork units, and Then the translation adjustment component corresponding to the kth aluminum template unit ( With rotation adjustment component ( The dimensions of the decision variables mentioned above cover all degrees of freedom for the pose adjustment of each unit, which can ensure complete compensation for the six degrees of freedom deviation of the template unit.

[0115] Based on the linear mapping relationship between the six-degree-of-freedom deviation parameters and the decision variables, a first objective based on the weighted sum of squares of the adjusted residual deviations can be constructed. The underlying principle is that the adjustment amount compensates for the initial six-degree-of-freedom deviations of the template unit. Therefore, under the assumption of small deviations, the adjusted residual deviations can be approximated as the difference between the initial six-degree-of-freedom deviation parameters and the decision variables, i.e., by concatenating the residual deviation vectors of all units. ,in Let be the concatenated vector of the initial six-degree-of-freedom deviation parameters, then the expression for the first objective is: ,in This is a diagonal weight matrix, where the diagonal elements are the weight coefficients corresponding to the previously obtained six-degree-of-freedom deviation parameters. The larger the weighting coefficient, the stronger the penalty for residual bias, thus ensuring that biases with high project priority are compensated first. Furthermore, a second objective can be constructed, which is the sum of the L1 norms of all decision variables, i.e. (in (This refers to the d-th adjustment component of the k-th unit). In the above construction process, based on the sparsity-induced properties of the L1 norm, most adjustment components in the optimization result can be set to 0 or close to 0, while only a few adjustment quantities that are key to deviation compensation are retained, thereby reducing unnecessary adjustment actions and reducing equipment wear and time costs during construction.

[0116] Specifically, by fusing the first and second objectives using preset target weight coefficients, and combining this with the physical connection and coupling characteristics of the template unit, a multi-objective optimization function can be obtained, the mathematical expression of which is:

[0117]

[0118] in, All are target weight coefficients. Minimizing the dominant residual bias, The sparsity of the dominant adjustment quantity Consistency in the adjustment of adjacent units; This refers to the physical connection edge set of the aluminum formwork unit. and Index of adjacent cells; The coupling coefficient between adjacent units is determined by the connection type, such as pin connection. Bolted connection Overlap Used to quantify the constraint strength of adjacent unit adjustments. This is the square of the Euclidean distance between the adjustment vectors of adjacent elements. This term is used to constrain the difference in adjustment between adjacent elements and avoid misalignment or stress concentration caused by inconsistent adjustment at the connection points.

[0119] Specifically, based on the pin connection point information in the physical connection relationship, adjacent units can be set. and The connection point in the unit The coordinates in the local coordinate system are In the unit The coordinates in the local coordinate system are Unit before adjustment The pose is a rotation matrix With translation vector ,unit The position is and Under the assumption of a small angle, the increment of the adjusted rotation matrix It can be approximated as an antisymmetric matrix (in (For rotational adjustment components). The global coordinates after the connection point adjustment must satisfy... By rearranging the formula, we can obtain information about... (including) )and (including) A system of linear equations can be formed by combining all such equations from adjacent units. This set of equations constitutes the geometric compatibility equation constraint, ensuring the continuity of displacement at the connection points.

[0120] Schematic illustration: The upper and lower bounds of each decision variable can also be determined based on the physical travel limits of the adjustable support equipment. The translational adjustment stroke and rotational adjustment angle of the adjustable support equipment both have physical limits, such as the maximum extension and minimum contraction of the semi-shift adjustment, and the maximum rotation angle of the rotational adjustment. Therefore, the decision variables can be determined based on these limits. For each component, determine its lower limit. and upper limit ,in The component corresponds to the minimum adjustment amount of the equipment. The component corresponds to the maximum adjustment amount of the equipment, thus forming Boundary inequality constraints are used to ensure that the adjustment amount is within the physical feasibility range of the equipment. Finally, by integrating the multi-objective optimization function, geometric compatibility equality constraints, and boundary inequality constraints, the mathematical optimization problem can be obtained. This problem is based on... As decision variables, using a multi-objective optimization function The objective function satisfies both geometric compatibility equality constraints and boundary inequality constraints. The optimal adjustment amount can be obtained by solving this problem.

[0121] In one embodiment, such as Figure 2 As shown, the mathematical optimization problem is numerically solved to obtain the optimal adjustment amount for each aluminum template unit; adjustment instructions are generated based on the optimal adjustment amounts to instruct the adjustable support equipment to perform corresponding operations; the execution order of the adjustment instructions is determined based on the physical connection relationship to obtain the adjustment instruction sequence, including:

[0122] S201: The interior point method or the alternating direction multiplier method is used to numerically solve the mathematical optimization problem, output the optimal solution of the decision variables, and use the optimal solution of the decision variables as the optimal adjustment amount of each aluminum template unit;

[0123] S202: Obtain the minimum movement step of the adjustable support equipment, discretize and round the optimal adjustment amount to obtain the basic adjustment amount that adapts to the adjustable support equipment;

[0124] S203: Generate an adjustment instruction containing adjustment direction, adjustment amplitude, and target unit identifier based on the basic adjustment amount;

[0125] S204: Extract the support dependencies of aluminum formwork units from the digital twin knowledge base of the aluminum formwork system, determine the adjustment priority based on the support dependencies, sort the adjustment instructions according to the adjustment priority, and obtain the adjustment instruction sequence.

[0126] Specifically, the interior-point method is suitable for handling convex optimization problems with equality and inequality constraints. Its core principle is to introduce a barrier function, such as a logarithmic barrier function, within the feasible region, transforming the constrained optimization problem into an iterative solution of a series of unconstrained optimization problems. Each iteration updates the decision variables using Newton's method while gradually tightening the parameters of the barrier function, causing the solution to gradually approach the boundary of the feasible region. The alternating direction multiplier method, on the other hand, is suitable for large-scale optimization problems. It decomposes the original problem into multiple independently solvable subproblems, achieving distributed solution by alternately optimizing different parts of the decision variables and updating the dual variable, thus reducing computational complexity. Illustratively, the interior-point method or the alternating direction multiplier method can be used to numerically solve mathematical optimization problems. During the solution process, convergence conditions can be set first, such as the difference between the objective functions of two adjacent iterations being less than a preset threshold. When the convergence condition is met, the output decision variable value is the optimal solution, which corresponds to the six degrees of freedom adjustment of each aluminum template element, i.e., the optimal adjustment amount.

[0127] Specifically, since the execution accuracy of the adjustable support device is determined by the minimum movement step size, translation adjustment corresponds to a linear step size, and rotation adjustment corresponds to an angular step size. Adjustments that are not integer multiples of the step size cannot be executed. Therefore, the minimum movement step size of the adjustable support device can be obtained, and the optimal adjustment amount can be discretized and rounded down. That is, for each optimal adjustment component... (The d-th adjustment component of the k-th unit) is calculated, along with the minimum movement step size. ratio Integer coefficients are obtained through rounding operations (such as rounding to the nearest integer). round Then calculate the basic adjustment amount. This ensures that the basic adjustment amount is an integer multiple of the minimum movement step size, fully matching the device's execution capabilities. Based on the basic adjustment amount, adjustment instructions containing the adjustment direction, adjustment amplitude, and target unit identifier can be generated. The adjustment direction is determined by the sign of the basic adjustment amount. If >0, then the adjustment direction is the positive direction of that degree of freedom. If the value is less than 0, the adjustment direction is the opposite of that degree of freedom. The adjustment amplitude is the absolute value of the base adjustment amount. This is used to specify the range of motion of the equipment. The target unit identifier is a unique code for the aluminum template unit, such as "template unit-001", used to specify the object to which the adjustment action is applied. Each adjustment command corresponds to a single degree of freedom adjustment of an aluminum template unit, or integrates multiple degree of freedom adjustments of the same unit into a composite command, to ensure that the command information is complete and can be directly parsed by the equipment.

[0128] Furthermore, the support dependencies of aluminum formwork units can be extracted from the digital twin knowledge base of the aluminum formwork system. These dependencies describe the mechanical support logic between formwork units. For example, some formwork units are reference units (such as corner formwork), whose installation positions provide positioning references for other units; others are dependent units, whose poses are determined by the poses of the reference units. Therefore, based on these support dependencies, a topological sorting algorithm can be used to sort all formwork units. This involves representing the support dependencies as a directed graph (nodes are formwork units, and directed edges represent the "dependent → dependent" relationship). By traversing the directed graph, an acyclic sorting sequence is obtained, where units at the beginning of the sequence are reference units with higher priority, and units at the end are dependent units with lower priority. Finally, each adjustment instruction can be associated with the corresponding template unit. Based on the topological sorting result of the template unit, the adjustment instructions of the corresponding reference unit are arranged at the beginning of the adjustment instruction sequence, and the adjustment instructions of the corresponding dependent unit are arranged at the end of the adjustment instruction sequence. This ensures that the adjustment of the reference unit is completed first and its pose is locked before the adjustment of the dependent unit is performed, avoiding the adjustment deviation of the dependent unit caused by the change of the reference unit pose, and ensuring the stability and accuracy of the entire adjustment process.

[0129] In one embodiment, it is determined whether the new structured bias dataset meets a preset accuracy requirement. If the new structured bias dataset does not meet the preset accuracy requirement, then S3 to S5 are repeated based on the new structured bias dataset, including:

[0130] Obtain the preset accuracy threshold, which includes the critical feature deviation threshold and the ordinary feature deviation threshold;

[0131] Extract the six-degree-of-freedom deviation parameters of each aluminum template unit in the new structured deviation dataset, and compare the six-degree-of-freedom deviation parameters with the key feature deviation threshold and the ordinary feature deviation threshold respectively to obtain the comparison results;

[0132] If the comparison results show that the six degrees of freedom deviation parameters corresponding to the preset key features of each aluminum template unit are all less than the key feature deviation threshold, and the six degrees of freedom deviation parameters corresponding to the non-preset key features are all less than the ordinary feature deviation threshold, then the new structured deviation dataset is determined to meet the preset accuracy requirements.

[0133] If the comparison result shows that the six degrees of freedom deviation parameter corresponding to the preset key feature of the aluminum template unit is not less than the key feature deviation threshold, or the six degrees of freedom deviation parameter corresponding to the non-preset key feature is not less than the ordinary feature deviation threshold, then the new structured deviation dataset does not meet the preset accuracy requirements.

[0134] Using the new structured bias dataset as input, re-execute S3 to S5 until the generated real-time structured bias dataset meets the preset accuracy requirements.

[0135] Specifically, the preset accuracy thresholds include key feature deviation thresholds and ordinary feature deviation thresholds. Illustratively, based on the geometric tolerance requirements for aluminum formwork installation in the building construction quality acceptance specifications, and combined with the functional requirements of concrete components, the features in the aluminum formwork system can be divided into preset key features and non-preset key features. Preset key features are those that play a decisive role in the load-bearing capacity and sealing performance of concrete components, such as load-bearing boundaries and pin connection holes. Non-preset key features are those that have a smaller impact on the core quality of the component, such as decorative lines in non-load-bearing areas. Subsequently, the key feature deviation threshold can be set to be lower than the ordinary feature deviation threshold to match the higher engineering quality control standards of the preset key features. Furthermore, in the new structured deviation dataset, the six degrees of freedom deviation parameters of each aluminum formwork unit are associated with corresponding semantic tags. These semantic tags can be used to determine the feature type (preset key feature or non-preset key feature). Therefore, each six-degree-of-freedom deviation parameter can be bound to the corresponding feature type. Then, the corresponding key feature deviation threshold or ordinary feature deviation threshold can be retrieved respectively. The absolute value of each component (translation deviation component, rotation deviation component) of the six-degree-of-freedom deviation parameter can be calculated. The absolute value is compared with the corresponding threshold to obtain the comparison result of each deviation parameter (less than or not less than the threshold).

[0136] If the comparison results show that the absolute values ​​of all components of the six-degree-of-freedom deviation parameters corresponding to all preset key features are less than the key feature deviation threshold, and the absolute values ​​of all components of the six-degree-of-freedom deviation parameters corresponding to all non-preset key features are less than the ordinary feature deviation threshold, then the new structured deviation dataset can be determined to meet the preset accuracy requirements. This judgment logic ensures that the installation accuracy of the core parts of the project meets the standards while also taking into account the reasonable error range of non-core parts, which is in line with the actual quality control logic of building construction. If the comparison results show that the absolute value of any component of the six-degree-of-freedom deviation parameter corresponding to any preset key feature is not less than the key feature deviation threshold, or the absolute value of any component of the six-degree-of-freedom deviation parameter corresponding to any non-preset key feature is not less than the ordinary feature deviation threshold, then the new structured deviation dataset can be determined to not meet the preset accuracy requirements. This overcomes the defects that any deviation exceeding the threshold may lead to problems such as misalignment of aluminum formwork splicing and deviation of concrete forming dimensions, thus ensuring the final quality of the components. Furthermore, the new structured deviation dataset can be used as input to re-execute subsequent steps, namely, constructing a multi-objective optimization function, solving a mathematical optimization problem, generating adjustment instructions, generating a new structured deviation dataset and comparing it, etc., until the generated real-time structured deviation dataset meets the preset accuracy requirements, thereby ensuring that the installation accuracy of the aluminum formwork gradually converges to the range that meets the engineering requirements.

[0137] Based on the same inventive concept, this application also provides an aluminum formwork installation accuracy control system for implementing the aluminum formwork installation accuracy control method described above. The solution provided by this system is similar to the solution described in the above method; therefore, the specific limitations in one or more embodiments of the aluminum formwork installation accuracy control system provided below can be found in the limitations of the aluminum formwork installation accuracy control method described above, and will not be repeated here.

[0138] In one exemplary embodiment, such as Figure 3 As shown, an aluminum formwork installation accuracy control system 300 is provided, including:

[0139] The feature extraction and matching module 301 is used to perform step 1, which includes: pre-extracting key local features from the design model of the aluminum template system and assigning semantic labels to them to obtain predefined key local features; acquiring the original point cloud data of the installed aluminum template, extracting actual local features based on the original point cloud data, matching the actual local features with the predefined key local features to obtain successfully matched actual local features, assigning corresponding semantic labels to the successfully matched actual local features, recording the actual geometric parameters of the successfully matched actual local features and the design theoretical geometric parameters of the corresponding matched predefined key local features, and obtaining a set of actual features with semantic labels.

[0140] The deviation calculation and data generation module 302 is used to execute step 2, which includes: calculating the six-degree-of-freedom deviation parameters of each aluminum template unit relative to the design pose of the aluminum template unit based on the actual geometric parameters of each feature in the actual feature set with semantic labels and the corresponding design theoretical geometric parameters; querying the preset feature importance weight rules based on semantic labels and assigning corresponding weight coefficients to each six-degree-of-freedom deviation parameter; and generating a structured deviation dataset based on the six-degree-of-freedom deviation parameters of each aluminum template unit and the corresponding weight coefficients.

[0141] The optimization problem construction module 303 is used to execute step 3, which includes: obtaining the physical connection relationship of the aluminum formwork system and the physical travel limit of each adjustable support device from the digital twin knowledge base of the aluminum formwork system; constructing a multi-objective optimization function based on the adjustment amount of each aluminum formwork unit based on the six-degree-of-freedom deviation parameters and weight coefficients in the structured deviation dataset; generating geometric coordination equality constraints based on the physical connection relationship; generating boundary inequality constraints based on the physical travel limit; and combining the multi-objective optimization function, geometric coordination equality constraints, and boundary inequality constraints to obtain the mathematical optimization problem.

[0142] The adjustment amount calculation and instruction generation module 304 is used to execute step 4, which includes: numerically solving the mathematical optimization problem to obtain the optimal adjustment amount of each aluminum template unit; generating adjustment instructions based on the optimal adjustment amount to instruct the adjustable support equipment to perform corresponding operations; and determining the execution order of the adjustment instructions based on the physical connection relationship to obtain the adjustment instruction sequence.

[0143] The adjustment execution and verification module 305 is used to execute step 5, which includes: sending the adjustment instruction sequence to the adjustable support device, executing the adjustment instruction sequence, and obtaining the verification point cloud data of the adjusted aluminum template; repeating steps 1 to 2 based on the verification point cloud data to generate a new structured deviation dataset, determining whether the new structured deviation dataset meets the preset accuracy requirements, and if the new structured deviation dataset does not meet the preset accuracy requirements, repeating steps 3 to 5 based on the new structured deviation dataset.

[0144] 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 template installation accuracy control method 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 large amounts of data and computational tasks.

[0145] In one exemplary embodiment, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the aluminum formwork installation accuracy control method 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.

[0146] 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 method for controlling the installation accuracy of aluminum formwork, characterized in that, The method includes: S1. Extract key local features from the design model of the aluminum formwork system and assign semantic labels to them to obtain predefined key local features; obtain the original point cloud data of the installed aluminum formwork, extract actual local features based on the original point cloud data, match the actual local features with the predefined key local features to obtain successfully matched actual local features, assign corresponding semantic labels to the successfully matched actual local features, record the actual geometric parameters of the successfully matched actual local features and the design theoretical geometric parameters of the corresponding matched predefined key local features, and obtain a set of actual features with semantic labels. S2. Based on the actual geometric parameters of each feature in the actual feature set with semantic labels and the corresponding theoretical geometric parameters of the design, calculate the six-degree-of-freedom deviation parameters of each aluminum template unit relative to the design pose of the aluminum template unit; based on the semantic labels, query the preset feature importance weight rules and assign corresponding weight coefficients to each of the six-degree-of-freedom deviation parameters; based on the six-degree-of-freedom deviation parameters of each aluminum template unit and the corresponding weight coefficients, generate a structured deviation dataset. S3. Obtain the physical connection relationships of the aluminum formwork system and the physical travel limits of each adjustable support device from the digital twin knowledge base of the aluminum formwork system; construct a multi-objective optimization function based on the adjustment amount of each aluminum formwork unit based on the six-degree-of-freedom deviation parameters and weight coefficients in the structured deviation dataset; generate geometrical compatibility equality constraints based on the physical connection relationships; generate boundary inequality constraints based on the physical travel limits; combine the multi-objective optimization function, the geometrical compatibility equality constraints, and the boundary inequality constraints to obtain the mathematical optimization problem; S4. Solve the mathematical optimization problem numerically to obtain the optimal adjustment amount for each of the aluminum template units; generate adjustment instructions based on the optimal adjustment amounts to instruct the adjustable support equipment to perform corresponding operations; determine the execution order of the adjustment instructions based on the physical connection relationship to obtain an adjustment instruction sequence; S5. Send the adjustment instruction sequence to the adjustable support device, execute the adjustment instruction sequence, and obtain the verification point cloud data of the adjusted aluminum template; repeat S1 to S2 based on the verification point cloud data to generate a new structured deviation dataset, determine whether the new structured deviation dataset meets the preset accuracy requirements, and if the new structured deviation dataset does not meet the preset accuracy requirements, repeat S3 to S5 based on the new structured deviation dataset.

2. The method according to claim 1, characterized in that, The process involves extracting actual local features from the original point cloud data, matching these actual local features with predefined key local features to obtain successfully matched actual local features, assigning corresponding semantic labels to these successfully matched actual local features, and recording the actual geometric parameters of the successfully matched actual local features and the design theoretical geometric parameters of the corresponding matched predefined key local features. This results in a set of actual features with semantic labels, including: The original point cloud data is subjected to statistical outlier removal processing to obtain a denoised point cloud; the voxel grid size is set based on the point cloud density of the denoised point cloud, and the voxel grid downsampling processing is performed on the denoised point cloud to obtain a preprocessed point cloud. Multi-scale geometric analysis is performed on the preprocessed point cloud to calculate the normal vector and curvature of each point. Candidate feature points are selected based on a preset curvature threshold and an angle threshold between the normal vectors of adjacent points. The candidate feature points are subjected to Euclidean clustering to obtain multiple candidate feature clusters, and the candidate feature clusters are determined as the actual local features. Calculate the geometric descriptor for each candidate feature cluster, and extract the corresponding baseline feature descriptor from the predefined key local features; The geometric descriptor of each candidate feature cluster is matched with the baseline feature descriptor for similarity, and stable matching pairs are selected by random sampling consensus algorithm to obtain the successfully matched candidate feature clusters. The successfully matched candidate feature clusters are used as the successfully matched actual local features, and the successfully matched actual local features are assigned semantic labels corresponding to the predefined key local features. For each successfully matched actual local feature, perform geometric fitting to obtain the corresponding fitting parameters. Extract the centroid coordinates, normal vector, and fitting geometric parameters of the successfully matched candidate feature cluster as actual geometric parameters. Retrieve the design theoretical geometric parameters of the predefined key local features that are matched. By integrating the semantic tags, the actual geometric parameters, and the design theoretical geometric parameters, the actual feature set with semantic tags is obtained.

3. The method according to claim 1, characterized in that, Based on the actual geometric parameters of each feature in the actual feature set with semantic tags and the corresponding theoretical geometric parameters of the design, the six-degree-of-freedom deviation parameters of each aluminum template unit relative to the design pose of the aluminum template unit are calculated; based on the semantic tags, a preset feature importance weighting rule is queried, and a corresponding weighting coefficient is assigned to each of the six-degree-of-freedom deviation parameters; Based on the six-degree-of-freedom deviation parameters and corresponding weighting coefficients of each aluminum template unit, a structured deviation dataset is generated, including: The topological relationship of the aluminum formwork system is obtained from the digital twin knowledge base of the aluminum formwork system, and the affiliation of the aluminum formwork unit corresponding to each semantic tag is determined according to the topological relationship; The actual geometric parameters and theoretical geometric parameters corresponding to the semantic tags of each aluminum template unit are summarized to construct a least squares optimization model; The least squares optimization model is solved by singular value decomposition algorithm to obtain the translational deviation parameters and rotational deviation parameters of the aluminum template unit relative to the design pose of the aluminum template unit, which are then combined to form the six-degree-of-freedom deviation parameters. Calculate the solution residual of the least squares optimization model, and query the preset feature type importance weight based on the semantic label; The preset feature type importance weights are weighted and fused with the solution residuals to obtain the weight coefficients corresponding to each of the six degrees of freedom deviation parameters; The structured deviation dataset is generated by associating the corresponding six-degree-of-freedom deviation parameters and weighting coefficients with the aluminum template unit identifier.

4. The method according to claim 1, characterized in that, Based on the six-degree-of-freedom deviation parameters and weight coefficients in the structured deviation dataset, a multi-objective optimization function based on the adjustment amount of each aluminum template unit is constructed. Based on the physical connection relationship, generate geometric coordination equation constraints; Based on the physical travel limit, generate boundary inequality constraints; Combining the multi-objective optimization function, the geometric compatibility equality constraints, and the boundary inequality constraints yields a mathematical optimization problem, including: Define decision variables, which consist of translation adjustment components and rotation adjustment components of each aluminum template unit; Based on the linear mapping relationship between the six-degree-of-freedom bias parameters and the decision variables, a first objective is constructed. The first objective is the weighted sum of squares of the adjusted residual biases, and the weight matrix of the first objective is composed of the weight coefficients in the structured bias dataset. Construct a second objective, which is the sum of the L1 norms of all the aforementioned decision variables; The first objective and the second objective are fused by a preset objective weight coefficient to obtain the multi-objective optimization function; Based on the pin connection point information in the physical connection relationship, the displacement continuity condition of the connected aluminum template units at the connection point is constructed, the displacement continuity condition is transformed into a system of linear equations based on the decision variables, and the system of linear equations is used as the geometric compatibility equation constraint. Based on the physical travel limit of the adjustable support equipment, the upper and lower boundary value ranges of each decision variable are determined to form the boundary inequality constraints. By integrating the multi-objective optimization function, the geometric compatibility equality constraint, and the boundary inequality constraint, the mathematical optimization problem is obtained.

5. The method according to claim 1, characterized in that, The mathematical optimization problem is numerically solved to obtain the optimal adjustment amount for each of the aluminum template units; based on the optimal adjustment amount, an adjustment command is generated to instruct the adjustable support device to perform the corresponding operation. Based on the physical connection relationship, the execution order of the adjustment instructions is determined, resulting in an adjustment instruction sequence, including: The mathematical optimization problem is solved numerically using the interior point method or the alternating direction multiplier method, and the optimal solution of the decision variables is output. The optimal solution of the decision variables is used as the optimal adjustment amount of each aluminum template unit. The minimum movement step of the adjustable support device is obtained, and the optimal adjustment amount is discretized and rounded to obtain the basic adjustment amount that adapts to the adjustable support device. Based on the aforementioned basic adjustment amount, an adjustment instruction is generated that includes the adjustment direction, adjustment amplitude, and target unit identifier; Extract the support dependencies of the aluminum formwork units from the digital twin knowledge base of the aluminum formwork system, and determine the adjustment priority based on the support dependencies; The adjustment instructions are sorted according to the adjustment priority to obtain the adjustment instruction sequence.

6. The method according to claim 1, characterized in that, The step of determining whether the new structured bias dataset meets the preset accuracy requirement, and if the new structured bias dataset does not meet the preset accuracy requirement, then repeating steps S3 to S5 based on the new structured bias dataset, includes: Obtain a preset accuracy threshold, which includes a key feature deviation threshold and a common feature deviation threshold; Extract the six-degree-of-freedom deviation parameters of each aluminum template unit in the new structured deviation dataset, and compare the six-degree-of-freedom deviation parameters with the key feature deviation threshold and the ordinary feature deviation threshold respectively to obtain the comparison results; If the comparison result shows that the six degrees of freedom deviation parameters corresponding to the preset key features of each aluminum template unit are all less than the key feature deviation threshold, and the six degrees of freedom deviation parameters corresponding to the non-preset key features are all less than the ordinary feature deviation threshold, then it is determined that the new structured deviation dataset meets the preset accuracy requirement. If the comparison result shows that the six-degree-of-freedom deviation parameter corresponding to the preset key feature of the aluminum template unit is not less than the key feature deviation threshold, or the six-degree-of-freedom deviation parameter corresponding to the non-preset key feature is not less than the ordinary feature deviation threshold, then it is determined that the new structured deviation dataset does not meet the preset accuracy requirement. Using the new structured bias dataset as input, S3 to S5 are re-executed until the generated real-time structured bias dataset meets the preset accuracy requirements.

7. According to claim 4, the mathematical expression of the multi-objective optimization function is: in, This is the adjustment vector for the aluminum template unit. The total number of units; This represents the six-degree-of-freedom adjustment amount for the k-th aluminum template element; This is the concatenated vector of the initial six-degree-of-freedom deviation parameters; This is a diagonal weight matrix, where the diagonal elements are the weight coefficients. All are target weight coefficients; This refers to the physical connection edge set of the aluminum formwork unit. and Index of adjacent cells; This represents the coupling coefficient between adjacent units.

8. A precision control system for aluminum formwork installation, characterized in that, The system includes: The feature extraction and matching module is used to perform step 1, which includes: pre-extracting key local features from the design model of the aluminum formwork system and assigning semantic labels to them to obtain predefined key local features; acquiring the original point cloud data of the installed aluminum formwork, extracting actual local features based on the original point cloud data, matching the actual local features with the predefined key local features to obtain successfully matched actual local features, assigning corresponding semantic labels to the successfully matched actual local features, and recording the actual geometric parameters of the successfully matched actual local features and the design theoretical geometric parameters of the corresponding matched predefined key local features to obtain a set of actual features with semantic labels. The deviation calculation and data generation module is used to execute step 2, which includes: calculating the six-degree-of-freedom deviation parameters of each aluminum template unit relative to the design pose of the aluminum template unit based on the actual geometric parameters of each feature in the actual feature set with semantic labels and the corresponding design theoretical geometric parameters; querying the preset feature importance weight rules based on the semantic labels and assigning corresponding weight coefficients to each of the six-degree-of-freedom deviation parameters; and generating a structured deviation dataset based on the six-degree-of-freedom deviation parameters and corresponding weight coefficients of each aluminum template unit. The optimization problem construction module is used to execute step 3, which includes: obtaining the physical connection relationship of the aluminum formwork system and the physical travel limit of each adjustable support device from the digital twin knowledge base of the aluminum formwork system; constructing a multi-objective optimization function based on the adjustment amount of each aluminum formwork unit based on the six-degree-of-freedom deviation parameters and weight coefficients in the structured deviation dataset; generating geometrical compatibility equality constraints based on the physical connection relationship; generating boundary inequality constraints based on the physical travel limit; and combining the multi-objective optimization function, the geometrical compatibility equality constraints, and the boundary inequality constraints to obtain the mathematical optimization problem. The adjustment amount calculation and instruction generation module is used to execute step 4, which includes: numerically solving the mathematical optimization problem to obtain the optimal adjustment amount of each of the aluminum template units; generating adjustment instructions based on the optimal adjustment amounts to instruct the adjustable support equipment to perform corresponding operations; and determining the execution order of the adjustment instructions based on the physical connection relationship to obtain an adjustment instruction sequence. The adjustment execution and verification module is used to execute step 5, which includes: sending the adjustment instruction sequence to the adjustable support device, executing the adjustment instruction sequence, and obtaining verification point cloud data of the adjusted aluminum template; repeating steps 1 to 2 based on the verification point cloud data to generate a new structured deviation dataset; determining whether the new structured deviation dataset meets the preset accuracy requirements; if the new structured deviation dataset does not meet the preset accuracy requirements, then repeating steps 3 to 5 based on the new structured deviation dataset.

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