A method and system for processing installation positioning data of a curved curtain wall

Abstract

The application discloses a kind of installation positioning data processing method and system for curved curtain wall, specifically relates to big data analysis and data processing technical field, for solving the problem of low installation precision and insufficient adaptive decision-making ability caused by insufficient multi-source data fusion in the installation positioning process of existing curved curtain wall;By acquiring design model data and real-time scanning point cloud data, space-time alignment is generated to generate a synchronized dataset, based on error propagation analysis, a global position uncertainty distribution is quantitatively generated, a nonlinear tightening installation point adjustment threshold is dynamically generated according to the position uncertainty distribution, the actual installation point coordinates are calculated by interpolation algorithm and the installation instruction set is generated, the installation instruction set is executed and the feedback data is collected, the installation error is traced back to the source to distinguish the systematic error and the random error and the installation point adjustment threshold is classified and corrected until the installation error is less than the preset tolerance, the continuous optimization and intelligent improvement of installation positioning accuracy are realized.

Description

Technical Field

[0001] This invention relates to the field of big data analysis and data processing technology, and more specifically, to a method and system for processing installation positioning data for curved curtain walls. Background Technology

[0002] In the digital installation and construction of curved curtain walls, the positioning accuracy and efficiency highly depend on the data processing capabilities of the design model and the data collected on-site. Existing technologies typically use a combination of computer-aided design models and discrete on-site measurement data for installation point planning. Specifically, the data processing system acquires the curtain wall's design model and attempts to match it with the on-site point cloud data obtained through scanning equipment, and then calculates the theoretical installation points using predefined algorithms. However, due to the complexity of the curved curtain wall's form, there are inherent data barriers and spatiotemporal reference differences between the two types of heterogeneous and massive data: the design model and the on-site environmental point cloud.

[0003] The existing data processing methods described above struggle to achieve high-precision data fusion and intelligent decision-making based on data characteristics when faced with massive, multi-source, and unstructured engineering data. This results in the system-generated installation instructions failing to adapt to the dynamic changes in surface curvature, and the decision results failing to match the actual state of the engineering entity, thus hindering the improvement of installation accuracy and intelligence. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides a method and system for processing installation positioning data for curved curtain walls to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for processing installation positioning data for curved curtain walls includes the following steps:

[0007] S1. Obtain the design model data and real-time scanned point cloud data of the curved curtain wall, perform spatiotemporal alignment of the design model data and real-time scanned point cloud data, and generate a synchronous dataset.

[0008] S2. Based on the synchronous dataset, the error propagation between the design model data and the real-time scanned point cloud data during the fusion process is analyzed to quantify the distribution of the position uncertainty across the entire domain.

[0009] S3. Dynamically generate the installation point adjustment threshold based on the location uncertainty distribution, wherein the installation point adjustment threshold tightens non-linearly as the location uncertainty increases;

[0010] S4. Based on the installation point adjustment threshold and the synchronization dataset, calculate the actual installation point coordinates using an interpolation algorithm and generate an installation instruction set;

[0011] S5. Execute the installation instruction set and collect feedback data during the installation process;

[0012] S6. Perform source analysis on the installation errors in the feedback data to distinguish between systematic errors and random errors, and adjust the installation point adjustment threshold according to the error type until the installation error is less than the preset tolerance.

[0013] Furthermore, design model data and real-time scanned point cloud data of the curved curtain wall are acquired, and spatiotemporal alignment of the design model data and real-time scanned point cloud data is performed to generate a synchronized dataset, including:

[0014] Transform the design model data to a preset global coordinate system;

[0015] Denoising is performed on real-time scanned point cloud data to remove noise points;

[0016] The iterative nearest point algorithm is used to register the denoised real-time scan point cloud data with the converted design model data;

[0017] And generate a synchronous dataset based on the registration results, which includes the registered design model data and real-time scanned point cloud data.

[0018] Furthermore, based on the synchronous dataset, by analyzing the error propagation during the fusion process of the design model data and real-time scanned point cloud data, the global position uncertainty distribution is quantified, including:

[0019] Extract the point cloud density distribution and registration residual between the design model data and the real-time scanned point cloud data from the synchronous dataset;

[0020] The local uncertainty of each region of the surface is calculated based on the point cloud density distribution and registration residual.

[0021] By combining sensor accuracy indicators and using the error propagation law, local uncertainties are fused to generate a position uncertainty distribution covering the entire curved curtain wall.

[0022] Furthermore, the calculation of local uncertainty in each region of the surface based on point cloud density distribution and registration residual includes: normalizing the point cloud density distribution to obtain density uncertainty components; normalizing the registration residual to obtain residual uncertainty components; and weighting and fusing the density uncertainty components and residual uncertainty components based on preset weighting coefficients to calculate the local uncertainty in each region of the surface.

[0023] Furthermore, an installation point adjustment threshold is dynamically generated based on the location uncertainty distribution, wherein the installation point adjustment threshold tightens non-linearly with increasing location uncertainty, including:

[0024] Obtain the position uncertainty values ​​for each region from the position uncertainty distribution;

[0025] The position uncertainty value is mapped to the corresponding installation point adjustment threshold based on a preset nonlinear function;

[0026] The nonlinear function is configured such that when the position uncertainty increases, the installation point adjustment threshold decreases nonlinearly.

[0027] Output the set of installation point adjustment thresholds corresponding to each area of ​​the curved curtain wall.

[0028] Furthermore, the method of mapping the position uncertainty value to the corresponding installation point adjustment threshold based on a preset nonlinear function includes: using an exponential function as the preset nonlinear function, inputting the position uncertainty value into the exponential function to calculate the installation point adjustment threshold; wherein the exponential function is configured such that the installation point adjustment threshold decreases exponentially as the position uncertainty value increases.

[0029] Furthermore, based on the installation point adjustment threshold and the synchronization dataset, the actual installation point coordinates are calculated using an interpolation algorithm, and an installation instruction set is generated, including:

[0030] Extract the coordinates of the theoretical installation point from the design model data in the synchronized dataset;

[0031] The allowable adjustment range of the theoretical installation point coordinates is determined based on the installation point adjustment threshold;

[0032] Within the allowable adjustment range, the coordinates of the theoretical installation point are smoothly interpolated using a bicubic spline interpolation algorithm to obtain the coordinates of the actual installation point.

[0033] The actual installation point coordinates are combined with the installation attitude parameters to generate an installation instruction set.

[0034] Furthermore, the installation instruction set is executed, and feedback data during the installation process is collected, including:

[0035] Control the installation equipment to execute installation instructions from the installation instruction set;

[0036] The actual installation point pose data is collected during the installation process using displacement sensors.

[0037] Installation force data during the installation process is collected using force sensors;

[0038] The actual installation point position data and installation force data are recorded as feedback data.

[0039] Furthermore, the installation errors in the feedback data are analyzed to distinguish between systematic and random errors, and the installation point adjustment threshold is corrected according to the error type until the installation error is less than the preset tolerance, including:

[0040] The deviation between the actual installation point pose data and the theoretical installation point coordinates is extracted from the feedback data as the installation error;

[0041] The time series stationarity test was performed on the installation error to distinguish between systematic and random errors;

[0042] The installation point adjustment threshold is corrected by a global offset compensation method for systematic errors and by a local smoothing filter method for random errors.

[0043] Repeat the threshold correction process until the installation error is less than the preset tolerance.

[0044] On the other hand, the present invention provides an installation positioning data processing system for curved curtain walls, comprising the following modules:

[0045] The synchronous data module is used to acquire the design model data and real-time scanned point cloud data of the curved curtain wall, perform spatiotemporal alignment on the design model data and real-time scanned point cloud data, and generate a synchronous dataset.

[0046] The distributed quantization module is used to quantize and generate the global position uncertainty distribution based on a synchronous dataset by analyzing the error propagation between the design model data and the real-time scanned point cloud data during the fusion process.

[0047] The threshold generation module is used to dynamically generate the installation point adjustment threshold based on the location uncertainty distribution, wherein the installation point adjustment threshold tightens non-linearly as the location uncertainty increases;

[0048] The instruction generation module is used to adjust the threshold and synchronize the dataset based on the installation point, calculate the actual installation point coordinates through an interpolation algorithm, and generate an installation instruction set.

[0049] The installation feedback module is used to execute the installation instruction set and collect feedback data during the installation process.

[0050] The error differentiation module is used to perform source analysis on the installation errors in the feedback data to distinguish between systematic errors and random errors, and to adjust the installation point adjustment threshold according to the error type until the installation error is less than the preset tolerance.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] 1. By constructing a synchronized dataset of design model and field point cloud data, a unified spatiotemporal reference and deep fusion of multi-source heterogeneous data were achieved. The position uncertainty distribution generated based on error propagation analysis can accurately quantify the positioning reliability of each region of the surface, providing data support for installation point adjustments. This dynamic threshold generation method based on uncertainty analysis enables installation point planning to adapt to changes in surface geometry, effectively overcoming the decision mismatch problem caused by insufficient data fusion in traditional methods.

[0053] 2. A closed-loop optimized installation accuracy control system was constructed through an intelligent source tracing and classification correction mechanism for installation errors. The system can automatically distinguish between systematic and random errors and adopt differentiated threshold correction strategies to continuously optimize the installation point adjustment threshold. This adaptive adjustment capability based on feedback data significantly improves the autonomous decision-making level of the curved curtain wall installation process, reduces the need for manual intervention, and ensures the synergistic improvement of installation accuracy and project quality. Attached Figure Description

[0054] Figure 1 This is a flowchart of an installation positioning data processing method for curved curtain walls according to the present invention;

[0055] Figure 2 This is a schematic diagram of the installation positioning data processing system for curved curtain walls according to the present invention. Detailed Implementation

[0056] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0057] Example 1: Figure 1 This invention provides a data processing method for installation and positioning of curved curtain walls, comprising the following steps:

[0058] S1. Obtain the design model data and real-time scanned point cloud data of the curved curtain wall, perform spatiotemporal alignment of the design model data and real-time scanned point cloud data, and generate a synchronous dataset.

[0059] S2. Based on the synchronous dataset, the error propagation between the design model data and the real-time scanned point cloud data during the fusion process is analyzed to quantify the distribution of the position uncertainty across the entire domain.

[0060] S3. Dynamically generate the installation point adjustment threshold based on the location uncertainty distribution, wherein the installation point adjustment threshold tightens non-linearly as the location uncertainty increases;

[0061] S4. Based on the installation point adjustment threshold and the synchronization dataset, calculate the actual installation point coordinates using an interpolation algorithm and generate an installation instruction set;

[0062] S5. Execute the installation instruction set and collect feedback data during the installation process;

[0063] S6. Perform source analysis on the installation errors in the feedback data to distinguish between systematic errors and random errors, and adjust the installation point adjustment threshold according to the error type until the installation error is less than the preset tolerance.

[0064] S1. Obtain the design model data and real-time scanned point cloud data of the curved curtain wall, perform spatiotemporal alignment of the design model data and real-time scanned point cloud data, and generate a synchronized dataset. The specific implementation is as follows:

[0065] After acquiring the design model data and real-time scanned point cloud data of the curved curtain wall, the design model data is first converted to a preset global coordinate system. This global coordinate system is defined based on at least three reference points set at the construction site. For example, the coordinates of the reference points are obtained by measuring with a total station, with the measurement accuracy controlled within 1 mm. The conversion process is achieved by applying a three-dimensional rotation matrix and translation vector. The parameters of the rotation matrix are calculated based on the angle difference between the original coordinate system of the design model and the global coordinate system. For example, the angle difference is obtained by vector cross product and dot product operations. The translation vector is determined based on the coordinate offset of the reference point. For example, the offset is calculated by the coordinate difference of the reference point in the two coordinate systems. This ensures that the coordinates of all points in the design model data are unified under the global coordinate system. The converted design model data is stored in the form of a floating-point array. Each point contains three-dimensional coordinates and normal vector information. For example, the three-dimensional coordinates include X-axis, Y-axis and Z-axis values, and the normal vector is obtained by geometric calculation of the point cloud surface.

[0066] Denoising is performed on real-time scanned point cloud data to remove noisy points. The denoising process employs a statistical outlier removal method. This method first calculates the average Euclidean distance between each point in the point cloud and its nearest neighbor. For example, if the number of neighboring points is set to 10, the average Euclidean distance is obtained by summing the distances between each point and its neighbors and then dividing by the number of neighboring points. Then, the global mean and standard deviation of the average distances of all points are calculated. For example, the global mean is calculated by summing the average distances of all points and dividing by the number of points, and the standard deviation is obtained by taking the square root of the average of the squared differences. A distance threshold is set as the global mean plus twice the standard deviation. For example, twice the standard deviation is used to identify outliers based on the principle of normal distribution. Any point whose average distance to its neighbors exceeds this threshold is identified as a noise point and removed. The denoised point cloud data retains the original color and intensity attributes and is stored in a point cloud list format to ensure data integrity.

[0067] The Iterative Nearest Point Algorithm (TNP) is used to register the denoised real-time scanned point cloud data with the transformed design model data. The registration process begins with initial alignment, which is based on the centroid matching of the point cloud and the design model. For example, the centroid is calculated using the average point coordinates. In each iteration, the TNP finds the nearest neighbor in the design model for each point in the point cloud, forming a set of point pairs. Then, the optimal rigid body transformation matrix between the point pairs is calculated. This matrix includes rotation and translation components. The average distance error between the point pairs is minimized using the least squares method. For example, the least squares method obtains the transformation parameters by solving a system of linear equations. The iteration termination condition is set to the change in the transformation matrix between two consecutive iterations being less than 0.001 mm or the number of iterations reaching 100. For example, the change is calculated using the matrix norm. The registered point cloud data and the design model data are aligned in the global coordinate system.

[0068] A synchronous dataset is generated based on the registration results. The synchronous dataset is created by merging the registered design model data and real-time scanned point cloud data. Each entry in the dataset contains the point's 3D coordinates, source identifier, timestamp, and accuracy index. For example, the source identifier is a string representing the design model or point cloud, the timestamp records the data acquisition time, and the accuracy index is calculated based on the registration residual. For example, the registration residual is obtained by averaging the distances between point pairs. The synchronous dataset is stored in the form of a structured database table. The table structure includes point index, coordinate column, attribute column, and metadata column. For example, the point index is an integer sequence, the coordinate column is a floating-point array, the attribute column stores color and intensity, and the metadata column records the registration parameters to ensure data traceability and consistency.

[0069] S2. Based on the synchronous dataset, by analyzing the error propagation between the design model data and the real-time scanned point cloud data during the fusion process, the global position uncertainty distribution is quantified and generated. The specific implementation is as follows:

[0070] Based on the synchronous dataset, point cloud density distribution and registration residuals are extracted from design model data and real-time scanned point cloud data. The point cloud density distribution is obtained by dividing the point cloud data into uniform grid cells and calculating the number of points in each grid cell. For example, the grid cell size is set to 10 mm by 10 mm. The number of points is obtained by counting the total number of points in each grid cell. The point cloud density distribution is stored in the form of a two-dimensional array, with each element corresponding to the density value of a grid cell. The registration residuals are extracted from the output of the iterative nearest point algorithm. The registration residuals represent the distance error of each point pair during the registration process. For example, the distance error is obtained by calculating the Euclidean distance between the point cloud points and the corresponding points in the design model. The registration residuals are stored in the form of a floating-point list, with each value corresponding to the residual size of a point pair. The extraction process ensures that the point cloud density distribution and registration residuals are aligned with the coordinate information in the synchronous dataset.

[0071] The local uncertainty of each region of the surface is calculated based on the point cloud density distribution and registration residuals. First, the point cloud density distribution is normalized to obtain the density uncertainty component. This normalization process maps the point cloud density values ​​to a range of 0 to 1 using a linear transformation; for example, the minimum point cloud density is mapped to 0, and the maximum point cloud density is mapped to 1. The density uncertainty component is calculated by subtracting the normalized density value from 1, resulting in a higher uncertainty component at lower densities. Then, the registration residuals are normalized to obtain the residual uncertainty component. This normalization process also maps the registration residual values ​​to a range of 0 to 1 using a linear transformation; for example, the minimum registration residual is mapped to 0, and the maximum registration residual is mapped to 1. The residual uncertainty component directly uses the normalized residual value, so that the larger the residual, the higher the uncertainty component. Finally, the density uncertainty component and the residual uncertainty component are weighted and fused based on preset weighting coefficients. The weighting coefficients are set by expert experience. For example, the weighting coefficient of the density uncertainty component is set to 0.6 and the weighting coefficient of the residual uncertainty component is set to 0.4. The weighted fusion is to obtain the local uncertainty value by multiplying the density uncertainty component of each region by the corresponding weighting coefficient and adding the residual uncertainty component by the corresponding weighting coefficient. The local uncertainty is stored in the same grid structure as the point cloud density distribution, and each grid corresponds to a local uncertainty value.

[0072] By combining sensor accuracy indicators, the local uncertainties are fused using the error propagation law to generate a position uncertainty distribution covering the entire curved curtain wall. Sensor accuracy indicators are obtained from sensor specifications; for example, the accuracy indicator of a laser scanner is 1 mm, representing the standard deviation of the measurement error. The error propagation law uses the variance addition principle to combine local uncertainties with sensor accuracy indicators. For example, local uncertainties are considered as one variance component, and sensor accuracy indicators are considered as another variance component. The total uncertainty is obtained by calculating the square root of the square of the local uncertainty plus the square of the sensor accuracy indicator. The position uncertainty distribution covers the entire curved curtain wall in a grid form, with the value of each grid representing the position uncertainty of that area. The generation process ensures that the local uncertainties of all areas are integrated. For example, the grid data is extended to the unsampled areas of the curved curtain wall using interpolation methods. The position uncertainty distribution is stored in the form of a floating-point matrix, with the matrix dimensions corresponding to the physical dimensions of the curved curtain wall.

[0073] S3. Dynamically generate the installation point adjustment threshold based on the location uncertainty distribution. The installation point adjustment threshold tightens non-linearly as the location uncertainty increases. Specifically, the implementation is as follows:

[0074] The installation point adjustment threshold is dynamically generated based on the location uncertainty distribution. First, the location uncertainty values ​​of each region are obtained from the location uncertainty distribution. The location uncertainty distribution is stored in the form of a two-dimensional matrix. The rows and columns of the matrix correspond to the grid division of the curved curtain wall. Each grid cell stores a location uncertainty value. For example, the grid cell size is 50 mm by 50 mm. The location uncertainty value is obtained by accessing the matrix element through the index. The value range is between 0 and 1, representing uncertainty from low to high. The acquisition process includes traversing all grid cells and reading the corresponding values ​​to ensure that the location uncertainty value of each region is completely extracted. The extracted values ​​are temporarily stored in the form of a floating-point number list. The list order is consistent with the grid index order.

[0075] The position uncertainty value is mapped to the corresponding installation point adjustment threshold based on a preset nonlinear function. The nonlinear function adopts an exponential function form, which is defined by a mathematical expression. The installation point adjustment threshold is equal to the base threshold multiplied by the negative scaling factor of the natural constant multiplied by the position uncertainty value raised to a power. For example, the base threshold is set to 10 mm according to the installation accuracy requirement, the scaling factor is set to 2 based on experience, and the natural constant is set to 2.718. The position uncertainty value is used as the input parameter. During the calculation, the position uncertainty value is first multiplied by the scaling factor to obtain the intermediate value, then the negative power of the intermediate value of the natural constant is calculated, and finally the base threshold is multiplied by the calculation result to obtain the installation point adjustment threshold. The exponential function is configured such that when the position uncertainty value increases, the installation point adjustment threshold decreases exponentially. For example, when the position uncertainty value increases from 0.1 to 0.5, the installation point adjustment threshold decreases from 8.2 mm to 3.7 mm. The deceleration rate gradually accelerates as the position uncertainty value increases. The mapping process calculates each position uncertainty value independently to ensure a one-to-one correspondence between the installation point adjustment threshold and the position uncertainty value.

[0076] Output a set of installation point adjustment thresholds corresponding to each area of ​​the curved curtain wall. The set of installation point adjustment thresholds is stored in the form of a two-dimensional matrix. The matrix dimension is the same as the position uncertainty distribution. Each grid cell stores an installation point adjustment threshold value. The output process includes filling the matrix with the calculated installation point adjustment threshold values ​​according to the original grid index. For example, the installation point adjustment threshold value of the first grid is stored in the first row and first column of the matrix. After the output is completed, data verification is performed to check whether all values ​​are within the preset range. For example, the installation point adjustment threshold value should not be less than 1 mm or greater than the basic threshold. The installation point adjustment threshold set is finally saved in the form of a file with comma-separated values, containing grid coordinates and corresponding installation point adjustment threshold values ​​to ensure that the data can be directly used in subsequent steps.

[0077] S4. Based on the installation point adjustment threshold and synchronization dataset, calculate the actual installation point coordinates using an interpolation algorithm and generate an installation instruction set. The specific implementation is as follows:

[0078] Based on the installation point adjustment threshold and the synchronous dataset, the theoretical installation point coordinates are first extracted from the design model data in the synchronous dataset. The synchronous dataset is stored in the form of a structured database table, which contains the registered design model data and real-time scanned point cloud data. The theoretical installation point coordinates are obtained by querying the records marked as installation points in the database table. For example, the installation point is identified by a specific attribute field in the design model, and the attribute field value is the installation point type. The theoretical installation point coordinates are stored in the form of three-dimensional vectors, each containing X-axis, Y-axis, and Z-axis coordinate values ​​in millimeters. The extraction process includes traversing all records in the database table, filtering out records marked as true installation points, and reading the corresponding coordinate values. The extracted theoretical installation point coordinates are temporarily stored in the form of a list, and the order of the list corresponds to the physical location of the installation point on the curved curtain wall.

[0079] The allowable adjustment range of theoretical installation point coordinates is determined based on the installation point adjustment threshold. The installation point adjustment threshold is obtained from the installation point adjustment threshold set generated in step S3. The installation point adjustment threshold set is stored in the form of a two-dimensional matrix, and the matrix elements correspond to the adjustment threshold values ​​of each region of the curved curtain wall. The allowable adjustment range is obtained by defining a spherical region with the theoretical installation point coordinates as the center and the installation point adjustment threshold as the radius. For example, the radius of the spherical region is equal to the installation point adjustment threshold value, in millimeters. The calculation of the allowable adjustment range includes finding the threshold value of the corresponding region in the installation point adjustment threshold matrix for each theoretical installation point coordinate, and then generating a spherical spatial region with the theoretical installation point coordinates as the center point and the threshold value as the radius. The spherical region is represented by the center point and radius parameters. The allowable adjustment range is stored in the form of a data structure containing the center coordinates and radius value of each installation point, ensuring that the adjustment range of all theoretical installation point coordinates is clearly defined.

[0080] Within the allowable adjustment range, the theoretical installation point coordinates are smoothly interpolated using a bicubic spline interpolation algorithm to obtain the actual installation point coordinates. The bicubic spline interpolation algorithm is implemented based on the theoretical installation point coordinate grid and the allowable adjustment range. The algorithm first arranges the theoretical installation point coordinates according to the grid structure, and the grid node spacing is set according to the size of the curved curtain wall, for example, the node spacing is set to 100 mm. Then, within each allowable adjustment range, the actual installation point coordinates are calculated using a bicubic spline function. The bicubic spline function is defined by 16 control points, which are selected from the theoretical installation point coordinate grid. The function calculation includes obtaining smooth coordinates for each interpolation point by polynomial interpolation based on the coordinate values ​​of adjacent control points and the first and second derivatives. The actual installation point coordinates are ensured to be within the allowable adjustment range, for example, by constraining the interpolation results not to exceed the boundary of the spherical region. After interpolation calculation, the actual installation point coordinates are stored in the form of a three-dimensional vector list, in millimeters, with the same dimension as the theoretical installation point coordinates.

[0081] The installation instruction set is generated by combining the actual installation point coordinates with the installation attitude parameters. The installation attitude parameters are extracted from the design model data, including the normal vector and rotation angle of the installation point. The normal vector is obtained by calculating the normal of the tangent plane of the surface at the installation point. The rotation angle is preset according to the installation process requirements, for example, the rotation angle is set to a range of 0 degrees to 360 degrees. The installation instruction set is generated by packaging each actual installation point coordinate and the corresponding installation attitude parameters into an instruction entry. The instruction entry contains a coordinate field and an attitude field. The coordinate field stores the three-dimensional value of the actual installation point coordinates, and the attitude field stores the normal vector and rotation angle. The installation instruction set is output in a structured file format, such as JSON format. The file contains a list of instructions, and each instruction entry corresponds to the complete information of an installation point, ensuring that the installation equipment can directly parse and execute it.

[0082] S5. Execute the installation instruction set and collect feedback data during the installation process. Specifically, the implementation is as follows:

[0083] The control system executes installation instructions from a set of installation instructions. This set is stored in a structured file format and includes the actual installation point coordinates and installation posture parameters. The actual installation point coordinates are represented by a three-dimensional vector, and the posture parameters include a normal vector and a rotation angle. The control process first parses the installation instruction set file, reading the coordinates and posture data of each instruction entry, and then converts them into control signals for the installation equipment. These control signals are sent to the actuators of the installation equipment, such as a robotic arm, via the equipment drive interface. The actuators move to the actual installation point coordinates and adjust to the specified posture according to the control signals. The moving speed is set according to the installation process requirements, for example, 50 millimeters per second. Posture adjustment is achieved through rotary joints, and the rotation angle is controlled according to the rotation angle value in the installation posture parameters. During execution, the equipment status is monitored in real time to ensure that each installation instruction is completed sequentially.

[0084] The actual installation point pose data is collected during the installation process using displacement sensors mounted on the end effector of the installation equipment. The displacement sensors are optical encoders with a measurement accuracy of 0.01 mm. The actual installation point pose data includes position coordinates and attitude angles. The position coordinates are obtained by measuring the three-dimensional coordinates of the end effector, such as the X, Y, and Z axes, using the displacement sensor. The attitude angles are obtained by measuring the angles around three axes, such as roll, pitch, and yaw, using the gyroscope built into the displacement sensor. The acquisition process is performed at a fixed frequency, for example, 100 Hz. The acquired actual installation point pose data is temporarily stored in time series format, with each data point containing a timestamp, coordinate value, and angle value to ensure that the data is synchronized with the execution of installation instructions.

[0085] Installation force data is collected during the installation process using a force sensor installed between the end effector of the installation equipment and the installation tool. The force sensor is a strain gauge type, with a measurement range of 0 to 100 Newtons and an accuracy of 0.1 Newtons. The installation force data includes force and torque components in three directions, such as forces and torques around the axes on the X, Y, and Z axes. The acquisition process is performed at a fixed frequency, for example, 100 Hz. The acquired installation force data is temporarily stored in time series form, with each data point containing a timestamp, force component, and torque component to ensure that the data is synchronized with the installation action. The force sensor is zero-point calibrated before acquisition to eliminate the influence of the equipment's own weight.

[0086] The actual installation point pose data and installation force data are recorded as feedback data. The feedback data undergoes data fusion processing to align the actual installation point pose data and installation force data by timestamp. The fused data is stored in a structured format, such as a table, which includes fields for timestamp, coordinate value, angle value, force component, and torque component. The recording process includes data verification to check whether the data is within a reasonable range, such as whether the coordinate value is within the boundary of the curved curtain wall and whether the force component is within the sensor range. The feedback data is finally saved as a data file in comma-separated value format, containing a complete record of all acquisition points to ensure that the data can be directly used in subsequent steps.

[0087] S6. Perform source analysis on the installation errors in the feedback data to distinguish between systematic errors and random errors, and adjust the installation point adjustment threshold according to the error type until the installation error is less than the preset tolerance. The specific implementation is as follows:

[0088] The deviation between the actual installation point pose data and the theoretical installation point coordinates is extracted from the feedback data as the installation error. The feedback data is stored in a structured table format, containing the actual installation point pose data and the theoretical installation point coordinates. The actual installation point pose data includes position coordinates and attitude angles, while the theoretical installation point coordinates are obtained from the design model data. The installation error is obtained by calculating the Euclidean distance between the position coordinates in the actual installation point pose data and the corresponding theoretical installation point coordinates. For example, the Euclidean distance is calculated by taking the square root of the sum of the squares of the differences in the three-dimensional coordinates, in millimeters. The extraction process includes aligning the actual installation point pose data and the theoretical installation point coordinates by timestamp, and then calculating the deviation value for each installation point. The installation error is stored in time series format, with each data point containing a timestamp and the corresponding deviation value, ensuring that the error of all installation points is completely extracted.

[0089] To distinguish between systematic and random errors, a time series stationarity test is performed on installation errors. The stationarity test employs a unit root test method, specifically the augmented Dickey-Fuller test. The process involves first arranging the installation error time series in chronological order, then calculating the test statistic. This statistic is obtained through regression analysis, with the regression model including lag terms, trend terms, and constant terms. For example, the number of lag terms is set to 2 based on the autocorrelation characteristics of the installation error series, and the trend term is set to a linear trend based on the time trend during installation. The test statistic is compared with a critical value obtained from a statistical distribution table, for example, with a significance level set to 0.05. If the test statistic is less than the critical value, the installation error series is considered stationary, corresponding to random errors. If the test statistic is greater than or equal to the critical value, the installation error series is considered non-stationary, corresponding to systematic errors. The results are stored as classification labels, with each installation point corresponding to an error type label.

[0090] For systematic errors, a global offset compensation method is used to correct the installation point adjustment threshold. Global offset compensation is achieved by calculating the average deviation of all systematic error installation points. The average deviation is obtained by summing the deviation values ​​of all systematic error installation points and dividing by the number of systematic error installation points, in millimeters. The correction process subtracts the average deviation value from the installation point adjustment threshold. For example, if the original installation point adjustment threshold is 5 millimeters and the average deviation is 1 millimeter, the corrected installation point adjustment threshold is 4 millimeters. The corrected installation point adjustment threshold is ensured to be no less than the minimum safety threshold, for example, the minimum safety threshold is set to 1 millimeter. For random errors, a local smoothing filter method is used to correct the installation point adjustment threshold. The local smoothing filter uses a moving average method, and the window size is set according to the fluctuation characteristics of the random error. For example, the window size is set to 5 adjacent installation points. During the correction process, for each random error installation point, the average value of its installation point adjustment threshold with that of its adjacent installation points is calculated, which is used as the corrected installation point adjustment threshold for that point. The corrected installation point adjustment threshold is updated and stored in matrix form, corresponding to each area of ​​the curved curtain wall.

[0091] Repeat the threshold correction process until the installation error is less than the preset tolerance. The preset tolerance is set according to the installation accuracy requirements, for example, the preset tolerance is set to 2 mm. The repeated process includes re-extracting the installation error, performing stability checks, classifying and correcting the installation point adjustment threshold. After each correction, calculate the maximum value of the current installation error and determine whether the maximum value is less than the preset tolerance. If it is less, terminate the repeated process; otherwise, continue to perform the correction. Set an upper limit for the number of repetitions, for example, set the upper limit to 10 times to prevent infinite loops. Finally, output the corrected installation point adjustment threshold set for use in subsequent installation processes.

[0092] Example 2: Figure 2 A schematic diagram of the installation positioning data processing system for curved curtain walls is provided according to the present invention. The installation positioning data processing system for curved curtain walls includes the following modules:

[0093] The synchronous data module is used to acquire the design model data and real-time scanned point cloud data of the curved curtain wall, perform spatiotemporal alignment on the design model data and real-time scanned point cloud data, and generate a synchronous dataset.

[0094] The distributed quantization module is used to quantize and generate the global position uncertainty distribution based on a synchronous dataset by analyzing the error propagation between the design model data and the real-time scanned point cloud data during the fusion process.

[0095] The threshold generation module is used to dynamically generate the installation point adjustment threshold based on the location uncertainty distribution, wherein the installation point adjustment threshold tightens non-linearly as the location uncertainty increases;

[0096] The instruction generation module is used to adjust the threshold and synchronize the dataset based on the installation point, calculate the actual installation point coordinates through an interpolation algorithm, and generate an installation instruction set.

[0097] The installation feedback module is used to execute the installation instruction set and collect feedback data during the installation process.

[0098] The error differentiation module is used to perform source analysis on the installation errors in the feedback data to distinguish between systematic errors and random errors, and to adjust the installation point adjustment threshold according to the error type until the installation error is less than the preset tolerance.

[0099] All calculations involved in the embodiments are dimensionless numerical calculations, and the preset parameters and thresholds in the calculations are set by those skilled in the art according to the actual situation.

[0100] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0101] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and inventive constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0102] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0103] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0104] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0105] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for processing installation positioning data for curved curtain walls, characterized in that, Includes the following steps: S1. Obtain the design model data and real-time scanned point cloud data of the curved curtain wall, perform spatiotemporal alignment of the design model data and real-time scanned point cloud data, and generate a synchronous dataset. S2. Based on the synchronous dataset, the error propagation between the design model data and the real-time scanned point cloud data during the fusion process is analyzed to quantify the distribution of the position uncertainty across the entire domain. S3. Dynamically generate installation point adjustment thresholds based on the location uncertainty distribution, wherein the installation point adjustment thresholds tighten non-linearly as the location uncertainty increases; S4. Based on the installation point adjustment threshold and the synchronization dataset, calculate the actual installation point coordinates using an interpolation algorithm and generate an installation instruction set; S5. Execute the installation instruction set and collect feedback data during the installation process; S6. Perform source analysis on the installation errors in the feedback data to distinguish between systematic and random errors, and adjust the installation point adjustment threshold according to the error type until the installation error is less than the preset tolerance, including: The deviation between the actual installation point pose data and the theoretical installation point coordinates is extracted from the feedback data as the installation error; The time series stationarity test was performed on the installation error to distinguish between systematic and random errors; The installation point adjustment threshold is corrected by a global offset compensation method for systematic errors and by a local smoothing filter method for random errors. Repeat the threshold correction process until the installation error is less than the preset tolerance.

2. The installation positioning data processing method for curved curtain walls according to claim 1, characterized in that, Acquire design model data and real-time scanned point cloud data of the curved curtain wall, perform spatiotemporal alignment on the design model data and real-time scanned point cloud data, and generate a synchronized dataset, including: Transform the design model data to a preset global coordinate system; Denoising is performed on real-time scanned point cloud data to remove noise points; The iterative nearest point algorithm is used to register the denoised real-time scan point cloud data with the converted design model data; And generate a synchronous dataset based on the registration results, which includes the registered design model data and real-time scanned point cloud data.

3. The installation positioning data processing method for curved curtain walls according to claim 1, characterized in that, Based on a synchronous dataset, by analyzing the error propagation during the fusion process of design model data and real-time scanned point cloud data, the global position uncertainty distribution is quantified, including: Extract the point cloud density distribution and registration residual between the design model data and the real-time scanned point cloud data from the synchronous dataset; The local uncertainty of each region of the surface is calculated based on the point cloud density distribution and registration residual. By combining sensor accuracy indicators and using the error propagation law, local uncertainties are fused to generate a position uncertainty distribution covering the entire curved curtain wall.

4. The installation positioning data processing method for curved curtain walls according to claim 3, characterized in that, The calculation of local uncertainty in each region of the surface based on point cloud density distribution and registration residual includes: normalizing the point cloud density distribution to obtain density uncertainty components; normalizing the registration residual to obtain residual uncertainty components; and weighting and fusing the density uncertainty components and residual uncertainty components based on preset weighting coefficients to calculate the local uncertainty in each region of the surface.

5. The installation positioning data processing method for curved curtain walls according to claim 1, characterized in that, The installation point adjustment threshold is dynamically generated based on the location uncertainty distribution. This threshold tightens non-linearly as the location uncertainty increases, and includes: Obtain the position uncertainty values ​​for each region from the position uncertainty distribution; The position uncertainty value is mapped to the corresponding installation point adjustment threshold based on a preset nonlinear function; The nonlinear function is configured such that when the position uncertainty increases, the installation point adjustment threshold decreases nonlinearly. Output the set of installation point adjustment thresholds corresponding to each area of ​​the curved curtain wall.

6. The installation positioning data processing method for curved curtain walls according to claim 5, characterized in that, The method of mapping the position uncertainty value to the corresponding installation point adjustment threshold based on the preset nonlinear function includes: using an exponential function as the preset nonlinear function, inputting the position uncertainty value into the exponential function to calculate the installation point adjustment threshold; wherein the exponential function is configured such that the installation point adjustment threshold decreases exponentially as the position uncertainty value increases.

7. The installation positioning data processing method for curved curtain walls according to claim 1, characterized in that, Based on the installation point adjustment threshold and synchronization dataset, the actual installation point coordinates are calculated using an interpolation algorithm, and an installation instruction set is generated, including: Extract the coordinates of the theoretical installation point from the design model data in the synchronized dataset; The allowable adjustment range of the theoretical installation point coordinates is determined based on the installation point adjustment threshold; Within the allowable adjustment range, the coordinates of the theoretical installation point are smoothly interpolated using a bicubic spline interpolation algorithm to obtain the coordinates of the actual installation point. The actual installation point coordinates are combined with the installation attitude parameters to generate an installation instruction set.

8. The installation positioning data processing method for curved curtain walls according to claim 1, characterized in that, Execute the installation instruction set and collect feedback data during the installation process, including: Control the installation equipment to execute installation instructions from the installation instruction set; The actual installation point pose data is collected during the installation process using displacement sensors. Installation force data is collected during the installation process using force sensors; The actual installation point position data and installation force data are recorded as feedback data.

9. A data processing system for installation positioning of curved curtain walls, used to implement the data processing method for installation positioning of curved curtain walls as described in any one of claims 1-8, characterized in that, Includes the following modules: The synchronous data module is used to acquire the design model data and real-time scanned point cloud data of the curved curtain wall, perform spatiotemporal alignment on the design model data and real-time scanned point cloud data, and generate a synchronous dataset. The distributed quantization module is used to quantize and generate the global position uncertainty distribution based on a synchronous dataset by analyzing the error propagation between the design model data and the real-time scanned point cloud data during the fusion process. The threshold generation module is used to dynamically generate the installation point adjustment threshold based on the location uncertainty distribution. The installation point adjustment threshold tightens non-linearly as the location uncertainty increases. The instruction generation module is used to adjust the threshold and synchronize the dataset based on the installation point, calculate the actual installation point coordinates through an interpolation algorithm, and generate an installation instruction set. The installation feedback module is used to execute the installation instruction set and collect feedback data during the installation process. The error differentiation module is used to perform source analysis on the installation errors in the feedback data to distinguish between systematic errors and random errors, and to adjust the installation point adjustment threshold according to the error type until the installation error is less than the preset tolerance.

Images