A steel pipe arch rib pre-assembly method, device, system and storage medium

Abstract

This invention provides a method, apparatus, system, and storage medium for pre-assembling steel pipe arch ribs, belonging to the field of steel pipe splicing technology. The method includes: scanning the steel pipe arch rib segment to be assembled and adjacent steel pipe arch rib segments using a 3D laser scanner to obtain original point cloud data and original adjacent point cloud data; performing control point analysis on the original point cloud data and original adjacent point cloud data to obtain an initial assembly surface control point 3D coordinate set and an initial adjacent assembly surface control point 3D coordinate set; and performing pre-assembly analysis on the initial assembly surface control point 3D coordinate set and the initial adjacent assembly surface control point 3D coordinate set to obtain the pre-assembly result of the steel pipe arch rib. This invention improves engineering efficiency, avoids errors from traditional manual measurement, saves manpower and material costs, reduces the number of on-site pre-assembly operations, shortens the construction period, and lays the foundation for subsequent overall alignment calculation and digital modeling of the steel pipe arch rib.

Description

Technical Field

[0001] This invention mainly relates to the field of steel pipe assembly technology, specifically to a method, device, system, and storage medium for pre-assembling steel pipe arch ribs. Background Technology

[0002] In the construction of steel-concrete composite arch bridges, the processing accuracy and assembly quality of the arch rib segments directly determine the bridge's alignment and structural safety. Traditional pre-assembly requires transporting components to a dedicated site for physical assembly, which has the following limitations: it consumes a lot of time and transportation costs, especially for long-span bridges, where the pre-assembly cycle can last for several weeks; due to site conditions, it is difficult to achieve full-size simulation; physical assembly relies on manual measurement, and the accuracy is affected by factors such as ambient temperature and instrument errors, leading to low efficiency due to repeated adjustments.

[0003] In recent years, 3D laser scanning technology has provided a new approach for the digital inspection of precast components. By acquiring point cloud data of the surface of steel arch rib segments and using computer technology to perform virtual pre-assembly checks on the overall arch rib alignment, this method avoids cumbersome physical pre-assembly and reduces site occupation. However, most current digital virtual assembly methods require adjusting the pre-assembly posture of the physical 3D point cloud model in conjunction with a BIM model, often resulting in multiple posture adjustments. Furthermore, poor matching between the theoretical model and the actual model leads to low accuracy in virtual pre-assembly. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a method, device, system and storage medium for pre-assembly of steel pipe arch ribs, which addresses the shortcomings of the prior art.

[0005] The technical solution of the present invention to solve the above-mentioned technical problems is as follows: A method for pre-assembling steel pipe arch ribs, comprising the following steps:

[0006] A 3D laser scanner is used to scan the steel pipe arch rib segment to be assembled and the neighboring steel pipe arch rib segment adjacent to the steel pipe arch rib segment to be assembled, thereby obtaining multiple original point cloud data of the steel pipe arch rib segment to be assembled and multiple original adjacent point cloud data of the neighboring steel pipe arch rib segment.

[0007] Control point analysis is performed on multiple original point cloud data of the steel pipe arch rib segment to be assembled and multiple original adjacent point cloud data of the neighboring steel pipe arch rib segment to obtain the three-dimensional coordinate set of the initial assembly surface control points of the steel pipe arch rib segment to be assembled and the three-dimensional coordinate set of the initial adjacent assembly surface control points of the neighboring steel pipe arch rib segment.

[0008] Pre-assembly analysis was performed on the three-dimensional coordinate set of the control points of the initial assembly surface and the three-dimensional coordinate set of the control points of the initial adjacent assembly surfaces to obtain the pre-assembly results of the steel pipe arch rib.

[0009] Another technical solution of the present invention to solve the above-mentioned technical problems is as follows: A steel pipe arch rib pre-assembly device, comprising:

[0010] The scanning module is used to scan the steel pipe arch rib segment to be assembled and the neighboring steel pipe arch rib segment adjacent to the steel pipe arch rib segment to be assembled using a 3D laser scanner, so as to obtain multiple original point cloud data of the steel pipe arch rib segment to be assembled and multiple original adjacent point cloud data of the neighboring steel pipe arch rib segment.

[0011] The control point analysis module is used to perform control point analysis on multiple original point cloud data of the steel pipe arch rib segment to be assembled and multiple original adjacent point cloud data of the neighboring steel pipe arch rib segment, respectively, to obtain the three-dimensional coordinate set of the initial assembly surface control points of the steel pipe arch rib segment to be assembled and the three-dimensional coordinate set of the initial adjacent assembly surface control points of the neighboring steel pipe arch rib segment.

[0012] The pre-assembly result acquisition module is used to perform pre-assembly analysis on the three-dimensional coordinate set of the control points of the initial assembly surface and the three-dimensional coordinate set of the control points of the initial adjacent assembly surfaces to obtain the pre-assembly result of the steel pipe arch rib.

[0013] Based on the above-mentioned method for pre-assembling steel pipe arch ribs, the present invention also provides a pre-assembly system for steel pipe arch ribs.

[0014] Another technical solution of the present invention to solve the above-mentioned technical problems is as follows: a pre-assembly system for steel pipe arch ribs, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the pre-assembly method for steel pipe arch ribs as described above is implemented.

[0015] Based on the above-mentioned method for pre-assembling steel pipe arch ribs, the present invention also provides a computer-readable storage medium.

[0016] Another technical solution of the present invention to solve the above-mentioned technical problems is as follows: a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steel pipe arch rib pre-assembly method described above is implemented.

[0017] The beneficial effects of this invention are as follows: By scanning the steel pipe arch rib segment to be assembled and adjacent steel pipe arch rib segments using a 3D laser scanner, original point cloud data and original adjacent point cloud data are obtained. Control point analysis of the original point cloud data and original adjacent point cloud data yields the initial assembly surface control point 3D coordinate set and the initial adjacent assembly surface control point 3D coordinate set. Pre-assembly analysis of the initial assembly surface control point 3D coordinate set and the initial adjacent assembly surface control point 3D coordinate set yields the pre-assembly result of the steel pipe arch rib. This improves engineering efficiency, avoids errors from traditional manual measurement, saves manpower and material costs, reduces the number of on-site pre-assembly operations, shortens the construction period, and lays the foundation for subsequent overall alignment calculation and digital modeling of the steel pipe arch rib. Attached Figure Description

[0018] Figure 1 A schematic flowchart of the pre-assembly method for steel pipe arch ribs provided in an embodiment of the present invention;

[0019] Figure 2 This is a module block diagram of the steel pipe arch rib pre-assembly device provided in an embodiment of the present invention. Detailed Implementation

[0020] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0021] Figure 1 This is a flowchart illustrating a method for pre-assembling steel pipe arch ribs according to an embodiment of the present invention.

[0022] like Figure 1 As shown, a method for pre-assembling steel pipe arch ribs includes the following steps:

[0023] A 3D laser scanner is used to scan the steel pipe arch rib segment to be assembled and the neighboring steel pipe arch rib segment adjacent to the steel pipe arch rib segment to be assembled, thereby obtaining multiple original point cloud data of the steel pipe arch rib segment to be assembled and multiple original adjacent point cloud data of the neighboring steel pipe arch rib segment.

[0024] Control point analysis is performed on multiple original point cloud data of the steel pipe arch rib segment to be assembled and multiple original adjacent point cloud data of the neighboring steel pipe arch rib segment to obtain the three-dimensional coordinate set of the initial assembly surface control points of the steel pipe arch rib segment to be assembled and the three-dimensional coordinate set of the initial adjacent assembly surface control points of the neighboring steel pipe arch rib segment.

[0025] Pre-assembly analysis was performed on the three-dimensional coordinate set of the control points of the initial assembly surface and the three-dimensional coordinate set of the control points of the initial adjacent assembly surfaces to obtain the pre-assembly results of the steel pipe arch rib.

[0026] It should be understood that the neighboring steel pipe arch rib segment can be a steel pipe arch rib segment adjacent to the steel pipe arch rib segment to be assembled.

[0027] It should be understood that scanning sites are planned and arranged, and 3D laser scanners and rotating triangular pyramid targets are set up according to the site locations. The 3D laser scanners scan the steel pipe arch rib segments (i.e., the steel pipe arch rib segments to be assembled and the adjacent steel pipe arch rib segments) to obtain on-site point cloud data (i.e., the original point cloud data and the original adjacent point cloud data).

[0028] Specifically, the planning and layout of scanning stations and target placement are as follows:

[0029] At least four stations should be set up for each arch rib segment;

[0030] One station is set up at each of the two splicing surfaces. The distance dz between the instrument and the end face of the arch rib segment should meet the requirement of 5m≤dz≤10m in order to better collect detailed end face data.

[0031] Stations are set on the sides of the arch rib segments according to the length of the arch rib. If the length of the arch rib segment is between 20m and 35m, a station can be set in the middle of it; and the scanning distance dz should satisfy 10m≤dz≤15m.

[0032] Set up a 3D laser scanner at the station. The triangular pyramid target should be placed in a position where it is visible from adjacent stations. When scanning station A and station B, rotate the target so that it is directly facing the scanner.

[0033] In the above embodiments, the original point cloud data and the original adjacent point cloud data are obtained by scanning the steel pipe arch rib segment to be assembled and the adjacent steel pipe arch rib segments using a 3D laser scanner. The control points of the original point cloud data and the original adjacent point cloud data are analyzed to obtain the three-dimensional coordinate set of the control points of the initial assembly surface and the three-dimensional coordinate set of the control points of the initial adjacent assembly surface. The pre-assembly analysis of the three-dimensional coordinate set of the control points of the initial assembly surface and the three-dimensional coordinate set of the control points of the initial adjacent assembly surface yields the pre-assembly result of the steel pipe arch rib. This improves engineering efficiency, avoids the errors of traditional manual measurement, saves manpower and material costs, reduces the number of pre-assembly operations on site, shortens the construction period, and lays the foundation for the subsequent overall alignment calculation and digital modeling of the steel pipe arch rib.

[0034] Optionally, as an embodiment of the present invention, the process of performing control point analysis on multiple original point cloud data of the steel pipe arch rib segment to be assembled and multiple original adjacent point cloud data of the neighboring steel pipe arch rib segment to obtain the three-dimensional coordinate set of the initial assembly surface control points of the steel pipe arch rib segment to be assembled and the three-dimensional coordinate set of the initial adjacent assembly surface control points of the neighboring steel pipe arch rib segment includes:

[0035] Registration analysis was performed on all the original point cloud data to obtain multiple initial flange plane point cloud data.

[0036] Preprocess all the initial flange plane point cloud data to obtain the first rotation matrix, the first centroid coordinates, and multiple target flange plane point cloud data;

[0037] Based on the first rotation matrix and the first centroid coordinates, control points are extracted from all the target flange plane point cloud data to obtain the initial assembly surface control point three-dimensional coordinate set of the steel pipe arch rib segment to be assembled;

[0038] Registration analysis was performed on all the original adjacent point cloud data to obtain multiple initial adjacent flange planar point cloud data;

[0039] Preprocess all the initial adjacent flange plane point cloud data to obtain the second rotation matrix, the second centroid coordinates, and multiple target adjacent flange plane point cloud data;

[0040] Based on the second rotation matrix and the second centroid coordinates, control points are extracted from the planar point cloud data of all adjacent flanges of the target to obtain the initial three-dimensional coordinate set of control points of the adjacent assembly surface of the neighboring steel pipe arch rib segment.

[0041] It should be understood that the process of performing registration analysis on all the original point cloud data is the same as the process of performing registration analysis on all the original adjacent point cloud data; the only difference is the data being processed.

[0042] It should be understood that the process of preprocessing all the initial flange plane point cloud data is the same as the process of preprocessing all the initial adjacent flange plane point cloud data; only the data being processed is different.

[0043] Specifically, the process of extracting control points from all target flange plane point cloud data based on the first rotation matrix and the first centroid coordinates is the same as the process of extracting control points from all adjacent target flange plane point cloud data based on the second rotation matrix and the second centroid coordinates; the only difference is the data being processed.

[0044] In the above embodiments, the control point analysis of the original point cloud data and the original adjacent point cloud data yields the three-dimensional coordinate set of the initial assembly surface control points and the three-dimensional coordinate set of the initial adjacent assembly surface control points, which improves engineering efficiency, avoids the errors of traditional manual measurement, saves manpower and material costs, and reduces the number of on-site pre-assembly operations.

[0045] Optionally, as an embodiment of the present invention, the process of performing registration analysis on all the original point cloud data to obtain multiple initial flange plane point cloud data includes:

[0046] Each of the original point cloud data A and any adjacent original point cloud data B are scanned to obtain multiple first original target point cloud data corresponding to each of the original point cloud data A and multiple second original target point cloud data corresponding to each of the original point cloud data B.

[0047] Multiple first original plane equations corresponding to each of the first original target point cloud data are extracted from each of the first original target point cloud data.

[0048] Multiple second original plane equations corresponding to each second original target point cloud data are extracted from each second original target point cloud data respectively;

[0049] The equations of the first and second primitive planes are defined as follows:

[0050] a0x1+b0y1+c0z1+d0=0,

[0051] Where a0 is the first X-axis normal vector component, b0 is the first Y-axis normal vector component, c0 is the first Z-axis normal vector component, and d0 is the first plane offset.

[0052] Using the random sampling consensus method and each of the first original target point cloud data, the parameters of each of the first original plane equations corresponding to each of the first original target point cloud data are calculated to obtain the first plane parameter set corresponding to each of the first original plane equations.

[0053] Each first original plane equation is updated according to each first plane parameter set to obtain multiple first updated plane equations corresponding to each first original target point cloud data.

[0054] Using the random sampling consensus method and each of the second original target point cloud data, the parameters of each of the second original plane equations corresponding to each of the second original target point cloud data are calculated to obtain the second plane parameter set corresponding to each of the second original plane equations.

[0055] Each second original plane equation is updated according to each second plane parameter set to obtain multiple second updated plane equations corresponding to each second original target point cloud data.

[0056] The intersection points of multiple first updated plane equations in each of the first original target point cloud data are calculated respectively, thereby obtaining the coordinates of the first feature points corresponding to each of the first original target point cloud data.

[0057] The intersection points of multiple second updated plane equations in each of the second original target point cloud data are calculated respectively, thereby obtaining the coordinates of the second feature points corresponding to each of the second original target point cloud data.

[0058] The coordinates of multiple first feature points corresponding to each original point cloud data A are respectively collected to obtain the first initial feature point coordinate set corresponding to each original point cloud data A;

[0059] The coordinates of multiple second feature points corresponding to each original point cloud data B are respectively collected to obtain the coordinate set of the second initial feature points corresponding to each original point cloud data B;

[0060] The fast four-point consensus set algorithm is used to perform coarse registration on each of the first initial feature point coordinate sets and the second initial feature point coordinate sets corresponding to each of the original point cloud data B, to obtain multiple coarsely registered feature point coordinate sets.

[0061] The ICP algorithm is used to perform fine registration processing on each set of coarsely registered feature point coordinates to obtain the initial rotation matrix and the translation vector corresponding to each set of coarsely registered feature point coordinates.

[0062] The Cloud Compare tool was used to segment and extract all the initial rotation matrices and all the translation vectors to obtain multiple initial flange plane point cloud data.

[0063] It should be understood that for the registration of point cloud data of arch ribs at multiple stations in the field (i.e., the original point cloud data), the registration of point cloud data of arch rib segments between stations can be achieved by using the spatial transformation matrix obtained by the registration of point cloud data of rotating triangular pyramid targets between stations.

[0064] Specifically, the registration process based on the rotating triangular pyramid target is as follows:

[0065] Assume that there should be at least two targets visible to each other between the two stations. There are a total of k targets between the two stations, and the target placement points cannot be collinear. Here, Ai and Bi are the point clouds of the i-th cone-shaped target in scan point cloud A and scan point cloud B, respectively (i.e., the first original target point cloud data and the second original target point cloud data).

[0066] Using the random sampling consensus method to fit the conical plane, the equations of the fitted plane (i.e., the first and second original plane equations) can be expressed as follows:

[0067] ax + by + cz + d = 0

[0068] In the formula: a, b, c, and d are all planar feature parameters. RANSAC sampling is performed on Ai (i.e., the first original target point cloud data). The parameters a, b, c, and d are calculated using the data from each sampling. The coordinates corresponding to the maximum probability density distribution of parameters aj, bj, cj, and dj (i.e., the first set of planar parameters) are taken as the deterministic solution of the planar equation coefficients. A target has three non-intersecting planes, and each plane parameter has three sets of peak values, j = 1, 2, 3. Similarly, for B... i (i.e., the second original target point cloud data) is used to calculate planar parameters.

[0069] Calculate the intersection points of the three planes fitted in scanning stations A and B, i.e., the coordinates ai of the feature point Ai (i.e., the coordinates of the first feature point). ) The feature point coordinates bi (i.e., the second feature point coordinates) of Bi are combined with the feature points of k targets. The feature point sets of the targets in station A and station B (i.e., the first initial feature point coordinate set and the second initial feature point coordinate set) are Ai = {i | i = 1: k} and Bi = {bi | i = 1: k}, respectively.

[0070] The fast four-point consistency set algorithm is used to analyze the target feature point set A. k B k The coordinates of the first and second initial feature points are registered to complete the coarse registration of the scanned point cloud data. Then, the ICP algorithm is used to perform fine registration on the target feature point set. The rotation matrix and translation vector are obtained, achieving efficient and fine registration of the two point cloud sites.

[0071] It should be understood that the data of the Nth arch rib segment after registration (i.e., the initial rotation matrix and translation vector) is imported into the Cloud Compare point cloud data processing software (i.e., the Cloud Compare tool), and the region clipping function is used to segment and extract the arch rib segment assembly surface (flange plane) and save it in PCD file format.

[0072] In the above embodiments, registration analysis is performed on all original point cloud data to obtain multiple initial flange plane point cloud data, avoiding the errors of traditional manual measurement and achieving millimeter-level accuracy.

[0073] Optionally, as an embodiment of the present invention, the process of preprocessing all the initial flange plane point cloud data to obtain the first rotation matrix, the first centroid coordinates, and multiple target flange plane point cloud data includes:

[0074] Denoising analysis was performed on all the initial flange plane point cloud data to obtain the first centroid coordinates, the fourth updated plane equation, and multiple processed flange plane point cloud data.

[0075] Based on the first centroid coordinates and the fourth updated plane equation, a rotational analysis is performed on all the processed flange plane point cloud data to obtain multiple target flange plane point cloud data.

[0076] In the above embodiments, all initial flange plane point cloud data are preprocessed to obtain the first rotation matrix, the first centroid coordinates, and multiple target flange plane point cloud data, which solves the problem of three-dimensional point feature point extraction and significantly improves the accuracy and reliability of control point positioning.

[0077] Optionally, as an embodiment of the present invention, the process of performing noise reduction analysis on all the initial flange plane point cloud data to obtain the first centroid coordinates, the fourth updated plane equation, and multiple processed flange plane point cloud data includes:

[0078] Define the equation of the third primitive plane as:

[0079] a1x2+b1y2+c1z2+d1=0,

[0080] Where a1 is the second X-axis normal vector component, b1 is the second Y-axis normal vector component, c1 is the second Z-axis normal vector component, and d1 is the second plane offset.

[0081] Construct a hybrid loss function, wherein the hybrid loss function is:

[0082]

[0083] Where L is the mixed loss value, a2 is the third X-axis normal vector component, b2 is the third Y-axis normal vector component, c2 is the third Z-axis normal vector component, d2 is the third plane offset, n is the number of initial flange plane point cloud data, and λ is the Lagrange multiplier.

[0084] Solving the hybrid loss function yields the third plane parameter set;

[0085] The third original plane equation is updated with parameters based on the third plane parameter set to obtain the third updated plane equation.

[0086] The first centroid coordinates are obtained by calculating all the initial flange plane point cloud data using the first formula, which is:

[0087]

[0088] in,

[0089] in, The first centroid coordinates, The average value of the initial flange planar point cloud data along the X-axis. The average value of the Y-axis of the initial flange plane point cloud data. The value of the initial flange plane point cloud data is the Z-axis average, where n is the number of initial flange plane point cloud data points, x4. i Let y4 be the X-axis coordinate of the i-th initial flange plane point cloud data. i Let z4 be the Y-axis coordinate of the i-th initial flange plane point cloud data. i The Z-axis coordinate of the i-th initial flange plane point cloud data;

[0090] The first centroid coordinates and all the initial flange plane point cloud data are decomposed using the covariance matrix eigenvalue decomposition algorithm to obtain a first eigenvalue, a first flange eigenvector corresponding to the first eigenvalue, a second eigenvalue, a second flange eigenvector corresponding to the second eigenvalue, a third eigenvalue, and a third flange eigenvector corresponding to the third eigenvalue.

[0091] The minimum value of the first feature value, the second feature value, and the third feature value is selected. After selection, the first flange feature vector, the second flange feature vector, or the third flange feature vector corresponding to the minimum value is used as the plane normal vector parameter group.

[0092] The fourth plane offset is obtained by solving the plane normal vector parameter set and the first centroid coordinates using the third original plane equation;

[0093] The third updated plane equation is updated with parameters based on the plane normal vector parameter set and the fourth plane offset to obtain the fourth updated plane equation.

[0094] Calculate the distance between each of the initial flange plane point cloud data and the fourth updated plane equation to obtain the target distance corresponding to each of the initial flange plane point cloud data;

[0095] The average error value is obtained by calculating the distances to all the targets using the second equation, which is:

[0096]

[0097] Where δ is the average error value, D i is the target distance corresponding to the i-th initial flange plane point cloud data, and n is the number of initial flange plane point cloud data;

[0098] If the target distance meets the determination condition, the initial flange plane point cloud data corresponding to the target distance is used as the flange plane point cloud data to be processed, thereby obtaining multiple flange plane point cloud data to be processed. The determination condition is:

[0099] D i >1.5δ,

[0100] Among them, D i δ represents the target distance corresponding to the i-th initial flange plane point cloud data, and δ is the average error value;

[0101] Determine whether the target distances all meet the determination conditions, and whether the absolute value of the difference between the average error value of the current iteration and the average error value of the previous iteration is less than a first preset threshold. If not, then reconstruct the hybrid loss function; if yes, then all the flange plane point cloud data to be processed are used as processed flange plane point cloud data, thereby obtaining multiple processed flange plane point cloud data.

[0102] It should be understood that the planar point distance constraint condition is used to perform noise reduction processing on the arch rib segment assembly surface (flange plane) (i.e., the initial flange plane point cloud data).

[0103] It should be understood that the optimal plane of the flange (i.e., the fourth updated plane equation) is fitted based on the least squares fitting plane method.

[0104] Specifically, the parameters of the plane equation are set to satisfy the normalization condition of the normal vector, i.e., a 2 +b 2 +c 2 =1, ensuring the uniqueness of the planar orientation; Objective function construction: minimizing the sum of squared distances from the point cloud to the plane is the optimization objective. A hybrid loss function is constructed using the Lagrange multiplier method, coupling the normalization condition of the plane normal vector with the error function; Eigenvalue decomposition solution: the eigenvector corresponding to the smallest eigenvalue is obtained through the covariance matrix eigenvalue decomposition method, which serves as the plane normal vector parameters a, b, c (i.e., the plane normal vector parameter set), combined with the centroid coordinates. Calculate the offset d (i.e., the fourth plane offset) and finally determine the optimal fitted plane equation (i.e., the fourth updated plane equation).

[0105] Specifically, a dynamic threshold iteration is set to remove point cloud noise data that deviates from the flange plane, and the average error δ (i.e., the average error value) is calculated:

[0106]

[0107] Set the dynamic threshold to 1.5δ and delete d. i Outliers >1.5δ, d i For point p i =(x i ,y i ,z iThe distance from the target plane to the fitted plane (i.e., the target distance) is used to repeatedly perform plane fitting and noise removal until the following condition is met: the average error difference between two iterations |δ k+1 -δ k |<ε,No new d i When the iteration reaches >1.5δ, the iteration terminates, and the denoised flange plane point cloud data is output (i.e., the processed flange plane point cloud data).

[0108] In the above embodiments, noise reduction analysis is performed on all initial flange plane point cloud data to obtain the first centroid coordinates, the fourth updated plane equation, and multiple processed flange plane point cloud data, effectively eliminating environmental interference and redundant data.

[0109] Optionally, as an embodiment of the present invention, the process of performing rotational analysis on all the processed flange plane point cloud data according to the first centroid coordinates and the fourth updated plane equation to obtain multiple target flange plane point cloud data includes:

[0110] By applying equations three, four, and five respectively to the fourth updated plane equation and each of the processed flange plane point cloud data, projected flange plane point cloud data corresponding to each of the processed flange plane point cloud data is obtained. Equation three is:

[0111] x1 j ′=

[0112] x5 j ×(b3 2 +c3 2 )-a3×(y5 j ×b3+z5 j ×c3+d3) / norm(|a3,b3,c3|), the fourth equation is:

[0113] y1′ j =

[0114] y5 j ×(a3 2 +c3 2 )-b3×(x5 j ×a3+z5 j ×c3+d3) / norm(|a3,b3,c3|), the fifth equation is:

[0115] z1 j ′=

[0116] z5 j ×(a3 2 +b3 2 )-c3×(x5 j ×a3+y5j ×b3+d3) / norm(|a3,b3,c3|),

[0117] Where x1 j y1' represents the X-axis coordinate of the projected flange plane point cloud data corresponding to the j-th processed flange plane point cloud data. j Let z1 be the Y-axis coordinate of the projected flange plane point cloud data corresponding to the j-th processed flange plane point cloud data. j ′ represents the Z-axis coordinate of the projected flange plane point cloud data corresponding to the j-th processed flange plane point cloud data, x5 j Let y5 be the X-axis coordinate of the j-th processed flange plane point cloud data. j Let z5 be the Y-axis coordinate of the j-th processed flange plane point cloud data. j Let a3 be the Z-axis coordinate of the j-th processed flange plane point cloud data, b3 be the fourth X-axis normal vector component, c3 be the fourth Z-axis normal vector component, d3 be the fourth plane offset, and norm(||) be the Euclidean length function.

[0118] According to the preset verification rules, the accuracy of all the projected flange plane point cloud data is verified. If the verification fails, the fourth updated plane equation and each of the processed flange plane point cloud data are re-projected using the third, fourth and fifth equations respectively. If the verification is successful, the projected flange plane point cloud data is used as the flange plane point cloud data to be rotated, thereby obtaining multiple flange plane point cloud data to be rotated.

[0119] The first included angle and the second included angle are extracted from the fourth updated plane equation. The first included angle is the angle between the normal vector of the fourth updated plane equation and the Z-axis under the projection of the YOZ coordinate system, and the second included angle is the angle between the normal vector of the fourth updated plane equation and the Z-axis under the projection of the XOZ coordinate system.

[0120] A first rotation matrix is ​​constructed using the first included angle and the second included angle. The first rotation matrix is:

[0121]

[0122] Where R1 is the first rotation matrix, α is the first included angle, and β is the second included angle;

[0123] Using the first centroid coordinate as the rotation center, the first rotation matrix is ​​used to rotate all the plane point cloud data of the flange to be rotated, thereby obtaining multiple target flange plane point cloud data.

[0124] It should be understood that the noise-reduced flange spatial plane point cloud (i.e., the processed flange plane point cloud data) is projected onto the plane coordinate system of the best-fit plane (i.e., the fourth updated plane equation), and the projected point cloud data (i.e., the projected flange plane point cloud data) is rotated to be parallel to the XOY plane.

[0125] It should be understood that the flange plane point cloud is fitted using the least squares fitting plane method to obtain the optimal fitting plane, where a, b, and c are the normal vector components corresponding to the optimal fitting plane, and d is the position parameter of the optimal fitting plane relative to the origin.

[0126] Specifically, the point cloud p of the flange plane is calculated using coordinate calculation formulas. i =(x i ,y i ,z i The new coordinates of the processed flange plane point cloud data are orthogonally projected onto the optimal fitting plane (i.e., the fourth updated plane equation), and the coordinate calculation formula is as follows:

[0127] x i ′=x i ×(b 2 +c 2 )-a×(y i ×b+z i ×c+d) / norm(|a,b,c|),

[0128] y i ′=y i ×(a 2 +c 2 )-b×(x i ×a+z i ×c+d) / norm(|a,b,c|),

[0129] z i ′=z i ×(a 2 +b 2 )-c×(x i ×a+y i ×b+d) / norm(|a,b,c|),

[0130] Wherein, the coordinates of the projected points (i.e., the projected flange plane point cloud data) are p′ i =(x′) i ,y′ i ,z′ i ), where norm(|a,b,c|) is the Euclidean length of the normal vector of the best-fitting plane;

[0131] The flange plane point cloud (i.e., the processed flange plane point cloud data) is projected onto the plane of the optimal fitting plane (i.e., the fourth updated plane equation), resulting in a set of projected point clouds of the flange plane (i.e., projected flange plane point cloud data). The accuracy of the projected point cloud dataset (i.e., the projected flange plane point cloud data) is verified to ensure that the error during the projection process is within a preset range, thus guaranteeing the accuracy of subsequent flange and bolt hole contour extraction. The optimal fitting plane (i.e., the fourth updated plane equation) is rotated until it is parallel to the XOY plane, and its rotation matrix (i.e., the first rotation matrix) is:

[0132]

[0133] In the formula: α is the angle between the plane normal vector projected onto the Z-axis in the YOZ coordinate system (i.e., the first angle), and β is the angle between the plane normal vector projected onto the Z-axis in the XOZ coordinate system (i.e., the second angle); sinα, cosα, sinβ, and cosβ can all be obtained from the optimally fitted plane normal vector n; the projection point cloud set of the flange plane (i.e., the projection point cloud data of the flange plane) is centered at the centroid. Using the rotation center as the rotation matrix R (i.e., the first rotation matrix), rotate until it is parallel to the XOY plane.

[0134] In the above embodiments, multiple target flange plane point cloud data are obtained by performing rotation analysis on the processed flange plane point cloud data based on the first centroid coordinates and the fourth updated plane equation. This avoids the errors of traditional manual measurement, saves manpower and material costs, and reduces the number of on-site pre-assembly operations.

[0135] Optionally, as an embodiment of the present invention, the process of extracting control points from all the target flange plane point cloud data based on the first rotation matrix and the first centroid coordinates to obtain the initial assembly surface control point three-dimensional coordinate set of the steel pipe arch rib segment to be assembled includes:

[0136] The adaptive Alpha-Shape algorithm is used to extract multiple flange contour points and multiple bolt hole contour points from all the target flange planar point cloud data, and the initial contour point dataset is obtained by combining all the flange contour points and all the bolt hole contour points.

[0137] The initial contour point dataset is segmented to obtain multiple subsets of initial contour point data;

[0138] The least squares algorithm is used to fit each of the initial contour point data subsets to obtain the circle equations corresponding to each of the initial contour point data subsets.

[0139] A system of error function equations is constructed, and the system of error function equations and each of the circle equations are solved to obtain the initial circle center coordinates corresponding to each subset of the initial contour point data. The system of error function equations is as follows:

[0140]

[0141] in, Let x be the x-axis coordinate of the k-th initial contour point in the a-th initial contour point data subset. Let be the Y-axis coordinate of the k-th initial contour point in the a-th initial contour point data subset. Let X be the initial center coordinate of the circle corresponding to the a-th initial contour point data subset. Let be the initial Y-axis coordinate of the center of the circle corresponding to the a-th initial contour point data subset, and E be the error function, r a Let be the radius corresponding to the a-th initial contour point data subset, and m be the number of initial contour points;

[0142] Each of the initial circle center coordinates is transformed using homogeneous coordinates to obtain the transformed circle center coordinates corresponding to each of the initial contour point data subsets.

[0143] The sixth equation is used to perform three-dimensional coordinate transformation on the first rotation matrix, each subset of initial contour point data, and the transformed center coordinates corresponding to each subset of initial contour point data, respectively, to obtain the target center coordinates corresponding to each subset of initial contour point data. Then, all the target center coordinates are combined to obtain the three-dimensional coordinate set of the initial assembly surface control points of the steel pipe arch rib segment to be assembled. The sixth equation is:

[0144]

[0145] in, Let X be the X-axis coordinate of the target circle center corresponding to the a-th initial contour point data subset. Let Y be the Y-axis coordinate of the target circle center corresponding to the a-th initial contour point data subset. Let Z be the Z-axis coordinate of the target circle center corresponding to the a-th initial contour point data subset. Let X be the initial center coordinate of the circle corresponding to the a-th initial contour point data subset. Let R1 be the initial Y-axis coordinate of the center of the circle corresponding to the a-th initial contour point data subset, and R1 be the first rotation matrix. The average value of the initial flange planar point cloud data along the X-axis. The average value of the Y-axis of the initial flange plane point cloud data. This represents the average Z-axis value of the initial flange plane point cloud data.

[0146] It should be understood that an adaptive alpha-shape algorithm is used to extract the flange and bolt hole contour point set (i.e., the three-dimensional coordinate set of the initial assembly surface control points).

[0147] Specifically, the adaptive alpha-shape algorithm is as follows:

[0148] Parameter initialization: Set the initial search radius parameter α0 of the adaptive Alpha-Shape algorithm as the benchmark scale for neighborhood partitioning;

[0149] Neighborhood construction: Based on the KD-tree spatial index structure, for each point p in the flange plane projection point cloud... i Perform a neighborhood search to obtain p i A spherical neighborhood set H centered at a radius of α0 i ;

[0150] Neighborhood density calculation: Calculate the point cloud density of each neighborhood, defined as the number of points H in the neighborhood. i The ratio of the volume of a spherical space to the volume of a spherical space.

[0151] Radius dynamic optimization: The search radius is dynamically adjusted based on the local density gradient; the radius is increased in sparse regions and decreased in dense regions by α. i The formula is: Where, ρ max This represents the global maximum density value.

[0152] Based on dynamic radius α i Construct a variable-scale rolling sphere model and traverse the point cloud to generate a topologically closed boundary;

[0153] Feature contour extraction: Identify the boundary point set covered by the Alpha-Shape model and output millimeter-precision contour data of the flange plane projection point cloud (i.e., flange contour points and bolt hole contour points), which are used for solving the geometric parameters of flange holes and bolt holes.

[0154] Understandably, the least squares method (i.e., the least squares algorithm) is used to perform circle fitting on the flange and bolt hole contours (i.e., the initial contour point dataset) to solve for the circle center coordinates (i.e., the initial circle center coordinates); and the obtained circle center coordinates (i.e., the initial circle center coordinates) are used to perform rotation inverse operation to obtain the bolt hole flange hole coordinates (XYZ) (i.e., the target circle center coordinates), and the three-dimensional coordinates of the assembly control points of the assembly surface (i.e., the initial assembly surface control point three-dimensional coordinate set) are obtained.

[0155] Specifically, the extracted flange and bolt hole contour points (i.e., the initial contour point dataset) are classified and segmented, and each contour point set (i.e., the subset of the initial contour point data) is denoted as U. iThere are a total of k contour point sets (i.e., subsets of the initial contour point data); contour point set U i The coordinates of the initial contour point data subset are: (x1, y1), (x2, y2), ..., (x...). i ,y i ), ..., (x n ,y n Let U be the set of contour points. i The center coordinates (i.e., the initial center coordinates) of the circle are (x0, y0), and the radius of the circle is r. According to the circle equation (x-x0)+(y-y0)... 2 =r is fitted to a circle using the least squares method. Based on the partial derivatives of the error function E with respect to the center coordinates (i.e., the initial center coordinates) and the radius, and setting the error function E to 0, the system of equations for the error function E is solved to obtain the values ​​of the center coordinates (x0, y0) (i.e., the initial center coordinates) and the radius r. The system of equations for the error function E is as follows:

[0156]

[0157] The center coordinates of the k contour points are collected into a point set G (i.e., the initial center coordinates corresponding to each subset of the initial contour point data). The fitted center coordinates p0 = (x0, y0, 0) are then converted into corresponding homogeneous coordinates: p0 = [x0, y0, 0, 1]. T Apply inverse transformation to recover the center coordinates p in the original coordinate system c =(x c ,y c ,z c (i.e., the coordinates of the center of the target circle), Where R T It is the inverse matrix of R; it will restore the center point p in the original coordinate system. ci (x ci ,y ci ,z ci The set of control points for the initial assembly surface is set to D (i.e., the three-dimensional coordinate set of the initial assembly surface control points). This is the set of control points for the digital pre-assembly (i.e., the three-dimensional coordinate set of the initial assembly surface control points). Repeat the above steps to solve for the set of control points for the assembly surfaces of adjacent arch rib segments (i.e., the three-dimensional coordinate set of the initial adjacent assembly surface control points), which are D respectively. n-1 (i.e., the initial three-dimensional coordinate set of adjacent assembly surface control points), D (i.e., the initial three-dimensional coordinate set of assembly surface control points).

[0158] In the above embodiments, the initial assembly surface control point three-dimensional coordinate set of the steel pipe arch rib segment to be assembled is obtained by extracting control points from the point cloud data of all target flange planes based on the first rotation matrix and the first centroid coordinates. This solves the problem of three-dimensional point feature point extraction and significantly improves the accuracy and reliability of control point positioning.

[0159] Optionally, as an embodiment of the present invention, the process of performing pre-assembly analysis on the three-dimensional coordinate set of the initial assembly surface control points and the three-dimensional coordinate set of the initial adjacent assembly surface control points to obtain the pre-assembly result of the steel pipe arch rib includes:

[0160] The fast four-point consistent set algorithm is used to perform coarse registration on the three-dimensional coordinate set of the initial assembly surface control points and the three-dimensional coordinate set of the initial adjacent assembly surface control points, respectively, to obtain the registered three-dimensional coordinate set of the assembly surface control points corresponding to the three-dimensional coordinate set of the initial assembly surface control points and the registered three-dimensional coordinate set of the adjacent assembly surface control points corresponding to the three-dimensional coordinate set of the initial adjacent assembly surface control points.

[0161] An ICP error function is constructed, and the singular value decomposition algorithm is used to calculate the ICP error function, the three-dimensional coordinate set of the control points of the registered assembly surface, and the three-dimensional coordinate set of the control points of adjacent registered assembly surfaces to obtain the third rotation matrix and the translation matrix. The ICP error function is:

[0162]

[0163] Where R2 is the third rotation matrix, T1 is the translation matrix, b is the number of 3D coordinates of the control points on the registered assembly surface or the number of 3D coordinates of adjacent control points on the registered assembly surface, and q u p represents the 3D coordinates of the u-th control point of the adjacent assembly surface after registration, within the 3D coordinate set of the adjacent assembly surface control points after registration. u This refers to the three-dimensional coordinates of the u-th control point of the registered assembly surface in the three-dimensional coordinate set of the control points after registration;

[0164] The three-dimensional coordinate set of the control points of the registered assembly surface is obtained by calculating the third rotation matrix and the translation matrix using the seventh equation. The seventh equation is:

[0165] D″=R2D′+T1,

[0166] Where D″ is the three-dimensional coordinate set of the control points of the assembly surface to be processed, R2 is the third rotation matrix, T1 is the translation matrix, and D′ is the three-dimensional coordinate set of the control points of the assembly surface after registration.

[0167] The average target distance is obtained by calculating the three-dimensional coordinate set of the control points of the assembly surface to be processed and the three-dimensional coordinate set of the control points of the adjacent assembly surfaces after registration using the eighth equation. The eighth equation is:

[0168]

[0169] in, q represents the average distance to the target, b represents the number of 3D coordinates of the control points on the registration surface or the number of 3D coordinates of adjacent control points on the registration surface, and q represents the distance to the target. u Let p′ be the 3D coordinate of the u-th control point of the adjacent assembly surface control points in the 3D coordinate set after registration. u Let be the 3D coordinates of the u-th control point of the assembly surface to be processed in the 3D coordinate set of the control points of the assembly surface to be processed;

[0170] Determine whether the average distance to the target is less than the second preset threshold. If not, then use the fast four-point consistent set algorithm to perform coarse registration on the three-dimensional coordinate set of the initial assembly surface control points and the three-dimensional coordinate set of the initial adjacent assembly surface control points respectively. If yes, then use the third rotation matrix and the translation matrix together as the pre-assembly result of the steel pipe arch rib.

[0171] It should be understood that the points in the control point set of the splicing surface of adjacent arch rib segments are registered, and the spatial transformation matrix (i.e., the third rotation matrix and the translation matrix) is output. The parameters of the spatial transformation matrix (i.e., the third rotation matrix and the translation matrix) are applied to the point cloud data of the complete arch rib segments, so that the adjacent arch rib segments can be digitally pre-stitched.

[0172] Specifically, the fast four-point consensus set algorithm is used to process the control point set D of the assembly surface of adjacent arch rib segments. n-1 (i.e., the initial three-dimensional coordinate set of control points for adjacent assembly surfaces), D n (i.e., the initial three-dimensional coordinate set of the assembly surface control points) is coarsely registered; the assembly control point set D of the pre-assembled N segments is... n (p1…p n (i.e., the three-dimensional coordinate set of the initial assembly surface control points) and the corresponding assembly control point set D on segment N-1. n-1 (q1…q n (i.e., the initial set of three-dimensional coordinates of adjacent assembly surface control points) is finely registered to construct the ICP error function (i.e., the ICP error function): Using singular value decomposition to analyze D n-1 (i.e., the initial three-dimensional coordinate set of control points for adjacent assembly surfaces) and D n (i.e., the initial set of 3D coordinates of the control points of the assembly surface) Matching point pairs calculates the rotation matrix R (i.e., the third rotation matrix) and the translation matrix T, resulting in a new point cloud set (i.e., the 3D coordinate set of the control points of the assembly surface to be processed) D′. n =RD n +T, calculate the point set D′ n With D n-1 The average distance (i.e., the average distance to the target) is: when (i.e., the average distance to the target) is less than the set value and the iteration stops at 1mm; the optimal rotation matrix R (i.e., the third rotation matrix) and translation vector T obtained by coarse and fine registration of the assembly control point set are used for spatial transformation of the overall point cloud data of the N segments of the arch rib, realizing the digital pre-assembly of adjacent steel arch rib segments.

[0173] In the above embodiments, the pre-assembly analysis of the three-dimensional coordinate set of the initial assembly surface control points and the three-dimensional coordinate set of the initial adjacent assembly surface control points yields the pre-assembly results of the steel pipe arch rib, which improves engineering efficiency, avoids errors in traditional manual measurement, saves manpower and material costs, reduces the number of on-site pre-assembly operations, shortens the construction period, and lays the foundation for subsequent overall alignment calculation and digital modeling of the steel pipe arch rib.

[0174] Optionally, as another embodiment of the present invention, the present invention relates to the field of steel-concrete composite arch bridge construction technology; the method includes: accurately acquiring point cloud data of steel-concrete composite arch ribs through multi-station scanning with a ground-based three-dimensional laser scanner, and extracting point clouds of the arch rib segment assembly surface; performing noise reduction processing on the arch rib segment assembly surface (flange plane) using planar point distance constraints; projecting the noise-reduced flange spatial planar point cloud onto the planar coordinate system of the best-fit plane, and rotating the projected point cloud data to be parallel to the XOY plane; extracting the flange and bolt hole contour point set using an adaptive alpha-shape algorithm; performing circle fitting on the flange and bolt hole contours using the least squares method to solve for the center coordinates of the circle; obtaining the bolt hole flange hole coordinates (XYZ) using rotation inverse operation, i.e., the three-dimensional coordinates of the assembly control points of the assembly surface; registering the points in the control point set of the splicing surface of adjacent arch rib segments, outputting a spatial transformation matrix, and applying the spatial transformation matrix parameters to the complete arch rib segment point cloud data, so that adjacent arch rib segments complete digital pre-splitting. The virtual pre-assembly method proposed in this invention improves engineering efficiency and saves manpower and material costs compared to physical pre-assembly, and lays the foundation for subsequent overall alignment calculation and digital modeling of steel pipe arch ribs.

[0175] Optionally, as another embodiment of the present invention, the present invention includes the following steps:

[0176] S1. Plan and arrange scanning sites, and set up 3D laser scanners and rotating triangular pyramid targets according to the site locations. The 3D laser scanners scan the steel pipe arch rib segments to obtain on-site point cloud data.

[0177] S2. Based on the multi-site point cloud data, register and construct a complete actual arch rib segment point cloud model, and segment and extract the arch rib segment assembly surface;

[0178] S3. Noise reduction treatment is applied to the assembly surface (flange plane) of the arch rib segment using planar point spacing constraints.

[0179] S4. Project the noise-reduced flange space plane point cloud onto the plane coordinate system of the best-fit plane, and rotate the projected point cloud data to be parallel to the XOY plane.

[0180] S5. Use the adaptive alpha-shape algorithm to extract the flange and bolt hole contour point set;

[0181] S6. Use the least squares method to perform circle fitting on the flange and bolt hole profiles and solve for the center coordinates; then use the obtained center coordinates to perform rotation inverse operation to obtain the bolt hole and flange hole coordinates (XYZ), and obtain the three-dimensional coordinates of the assembly control points of the assembly surface.

[0182] S7. Register the points in the control point set of the splicing surface of adjacent arch rib segments, output the spatial transformation matrix, and apply the parameters of the spatial transformation matrix to the point cloud data of the complete arch rib segments so that adjacent arch rib segments can be digitally pre-stitched.

[0183] Optionally, as another embodiment of the present invention, the present invention includes:

[0184] The data acquisition and preprocessing module plans and arranges scanning stations, sets up a 3D laser scanner and a rotating triangular pyramid target according to the station locations, and uses the 3D laser scanner to scan the steel pipe arch rib segments to obtain on-site point cloud data. Based on the on-site multi-station point cloud data, it registers and constructs a complete actual arch rib segment point cloud model, and segments and extracts the arch rib segment assembly surface.

[0185] The assembly control point extraction module uses planar point spacing constraints to denoise the assembly surface (flange plane) of the arch rib segment; it projects the denoised flange spatial plane point cloud onto the plane coordinate system of the best-fit plane, and rotates the projected point cloud data to be parallel to the XOY plane; it uses an adaptive alpha-shape algorithm to extract the flange and bolt hole contour point sets; it uses the least squares method to perform circle fitting on the flange and bolt hole contours to solve for the center coordinates of the circle; and it uses the obtained center coordinates to obtain the bolt hole and flange hole coordinates (XYZ) through rotation inverse operation, thus obtaining the three-dimensional coordinates of the assembly control points of the assembly surface.

[0186] The adjacent arch rib pre-assembly module registers the points in the control point set of the splicing surface of adjacent arch rib segments, outputs a spatial transformation matrix, and applies the parameters of the spatial transformation matrix to the point cloud data of the complete arch rib segments, so that the adjacent arch rib segments can complete digital pre-splitting.

[0187] Optionally, as another embodiment of the present invention, the beneficial effects of the present invention are as follows: By acquiring on-site point cloud data through three-dimensional laser scanning technology and combining it with multi-site point cloud registration to construct a complete arch rib model, the errors of traditional manual measurement are avoided, achieving millimeter-level accuracy. The data processing flow is optimized using planar projection, and the bolt hole contour is accurately extracted using the alpha-shape algorithm, solving the problem of three-dimensional point feature point extraction and significantly improving the accuracy and reliability of control point positioning. Intelligent noise reduction of the flange point cloud is achieved through planar point distance constraints, effectively eliminating environmental interference and redundant data. By extracting the three-dimensional coordinates of the assembly control points, "digital pre-assembly" of the arch rib segments is realized, reducing the number of on-site pre-assembly attempts and shortening the construction period. Simultaneously, repeated adjustments to traditional welding positioning are avoided, reducing labor costs and material waste. The entire process from data acquisition to control point generation is digitized, supporting reverse verification and quality traceability, providing a scientific basis for arch rib alignment deviation analysis and construction correction, and ensuring the overall structural safety of the bridge.

[0188] Figure 2 This is a module block diagram of a steel pipe arch rib pre-assembly device provided in an embodiment of the present invention.

[0189] Alternatively, as another embodiment of the present invention, such as Figure 2 As shown, a pre-assembly device for steel pipe arch ribs includes:

[0190] The scanning module is used to scan the steel pipe arch rib segment to be assembled and the neighboring steel pipe arch rib segment adjacent to the steel pipe arch rib segment to be assembled using a 3D laser scanner, so as to obtain multiple original point cloud data of the steel pipe arch rib segment to be assembled and multiple original adjacent point cloud data of the neighboring steel pipe arch rib segment.

[0191] The control point analysis module is used to perform control point analysis on multiple original point cloud data of the steel pipe arch rib segment to be assembled and multiple original adjacent point cloud data of the neighboring steel pipe arch rib segment, respectively, to obtain the three-dimensional coordinate set of the initial assembly surface control points of the steel pipe arch rib segment to be assembled and the three-dimensional coordinate set of the initial adjacent assembly surface control points of the neighboring steel pipe arch rib segment.

[0192] The pre-assembly result acquisition module is used to perform pre-assembly analysis on the three-dimensional coordinate set of the control points of the initial assembly surface and the three-dimensional coordinate set of the control points of the initial adjacent assembly surfaces to obtain the pre-assembly result of the steel pipe arch rib.

[0193] Optionally, another embodiment of the present invention provides a pre-assembly system for steel pipe arch ribs, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the pre-assembly method for steel pipe arch ribs as described above. This system can be a computer or similar system.

[0194] Optionally, another embodiment of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steel pipe arch rib pre-assembly method as described above.

[0195] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0196] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the above-described apparatus and unit can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0197] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0198] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of the present invention, depending on actual needs.

[0199] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0200] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0201] 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 pre-assembling steel pipe arch ribs, characterized in that, Includes the following steps: A 3D laser scanner is used to scan the steel pipe arch rib segment to be assembled and the neighboring steel pipe arch rib segment adjacent to the steel pipe arch rib segment to be assembled, thereby obtaining multiple original point cloud data of the steel pipe arch rib segment to be assembled and multiple original adjacent point cloud data of the neighboring steel pipe arch rib segment. Control point analysis is performed on multiple original point cloud data of the steel pipe arch rib segment to be assembled and multiple original adjacent point cloud data of the neighboring steel pipe arch rib segment to obtain the three-dimensional coordinate set of the initial assembly surface control points of the steel pipe arch rib segment to be assembled and the three-dimensional coordinate set of the initial adjacent assembly surface control points of the neighboring steel pipe arch rib segment. Pre-assembly analysis is performed on the three-dimensional coordinate set of the control points of the initial assembly surface and the three-dimensional coordinate set of the control points of the initial adjacent assembly surfaces to obtain the pre-assembly results of the steel pipe arch rib. The process of performing control point analysis on multiple original point cloud data of the steel pipe arch rib segment to be assembled and multiple original adjacent point cloud data of the neighboring steel pipe arch rib segment to obtain the three-dimensional coordinate set of the initial assembly surface control points of the steel pipe arch rib segment to be assembled and the three-dimensional coordinate set of the initial adjacent assembly surface control points of the neighboring steel pipe arch rib segment includes: The flange plane is the assembly surface of the arch rib segment. Registration analysis is performed on all the original point cloud data to obtain multiple initial flange plane point cloud data. Preprocess all the initial flange plane point cloud data to obtain the first rotation matrix, the first centroid coordinates, and multiple target flange plane point cloud data; Based on the first rotation matrix and the first centroid coordinates, control points are extracted from all the target flange plane point cloud data to obtain the initial assembly surface control point three-dimensional coordinate set of the steel pipe arch rib segment to be assembled; Registration analysis was performed on all the original adjacent point cloud data to obtain multiple initial adjacent flange planar point cloud data; Preprocess all the initial adjacent flange plane point cloud data to obtain the second rotation matrix, the second centroid coordinates, and multiple target adjacent flange plane point cloud data; Based on the second rotation matrix and the second centroid coordinates, control points are extracted from the plane point cloud data of all adjacent flanges of the target to obtain the initial three-dimensional coordinate set of control points of the adjacent assembly surface of the neighboring steel pipe arch rib segment; The process of preprocessing all the initial flange plane point cloud data to obtain the first rotation matrix, the first centroid coordinates, and multiple target flange plane point cloud data includes: Denoising analysis is performed on all the initial flange plane point cloud data to obtain the first centroid coordinates, the fourth updated plane equation, and multiple processed flange plane point cloud data. The first centroid coordinates are the average coordinates of the initial flange plane point cloud data. Based on the first centroid coordinates and the fourth updated plane equation, a rotational analysis is performed on all the processed flange plane point cloud data to obtain multiple target flange plane point cloud data. The first included angle and the second included angle are extracted from the fourth updated plane equation. The fourth updated plane equation is the best fitting plane of the initial flange plane point cloud data. The first included angle is the angle between the normal vector of the fourth updated plane equation and the Z-axis under the projection of the YOZ coordinate system. The second included angle is the angle between the normal vector of the fourth updated plane equation and the Z-axis under the projection of the XOZ coordinate system. A first rotation matrix is ​​constructed using the first included angle and the second included angle. The first rotation matrix is: ， in, Let be the first rotation matrix. The first included angle, This is the second included angle.

2. The pre-assembly method for steel pipe arch ribs according to claim 1, characterized in that, The process of performing registration analysis on all the original point cloud data to obtain multiple initial flange plane point cloud data includes: Each original point cloud data A and any adjacent original point cloud data B are scanned to obtain multiple first original target point cloud data corresponding to each original point cloud data A and multiple second original target point cloud data corresponding to each original point cloud data B. Multiple first original plane equations corresponding to each of the first original target point cloud data are extracted from each of the first original target point cloud data. Multiple second original plane equations corresponding to each second original target point cloud data are extracted from each second original target point cloud data respectively; The equations of the first and second primitive planes are defined as follows: ， in, The first X-axis normal vector component, The first Y-axis normal vector component, The first Z-axis normal vector component, This is the offset of the first plane; Using the random sampling consensus method and each of the first original target point cloud data, the parameters of each of the first original plane equations corresponding to each of the first original target point cloud data are calculated to obtain the first plane parameter set corresponding to each of the first original plane equations. Each first original plane equation is updated according to each first plane parameter set to obtain multiple first updated plane equations corresponding to each first original target point cloud data. Using the random sampling consensus method and each of the second original target point cloud data, the parameters of each of the second original plane equations corresponding to each of the second original target point cloud data are calculated to obtain the second plane parameter set corresponding to each of the second original plane equations. Each second original plane equation is updated according to each second plane parameter set to obtain multiple second updated plane equations corresponding to each second original target point cloud data. The intersection points of multiple first updated plane equations in each of the first original target point cloud data are calculated respectively, thereby obtaining the coordinates of the first feature points corresponding to each of the first original target point cloud data. The intersection points of multiple second updated plane equations in each of the second original target point cloud data are calculated respectively, thereby obtaining the coordinates of the second feature points corresponding to each of the second original target point cloud data. The coordinates of multiple first feature points corresponding to each original point cloud data A are respectively collected to obtain the first initial feature point coordinate set corresponding to each original point cloud data A; The coordinates of multiple second feature points corresponding to each original point cloud data B are respectively collected to obtain the coordinate set of the second initial feature points corresponding to each original point cloud data B; The fast four-point consensus set algorithm is used to perform coarse registration on each of the first initial feature point coordinate sets and the second initial feature point coordinate sets corresponding to each of the original point cloud data B, to obtain multiple coarsely registered feature point coordinate sets. The ICP algorithm is used to perform fine registration processing on each set of coarsely registered feature point coordinates to obtain the initial rotation matrix and the translation vector corresponding to each set of coarsely registered feature point coordinates. The Cloud Compare tool was used to segment and extract all the initial rotation matrices and all the translation vectors to obtain multiple initial flange plane point cloud data.

3. The pre-assembly method for steel pipe arch ribs according to claim 1, characterized in that, The process of performing noise reduction analysis on all the initial flange plane point cloud data to obtain the first centroid coordinates, the fourth updated plane equation, and multiple processed flange plane point cloud data includes: Define the equation of the third primitive plane as: ， in, The second X-axis normal vector component, The second Y-axis normal vector component, The second Z-axis normal vector component, This is the offset of the second plane; Construct a hybrid loss function, wherein the hybrid loss function is: ， in, For mixed loss values, The third X-axis normal vector component. The third Y-axis normal vector component, The third Z-axis normal vector component, This is the offset of the third plane. This represents the initial number of point cloud data points for the flange plane. For Lagrange multipliers; Solving the hybrid loss function yields the third plane parameter set; The third original plane equation is updated with parameters based on the third plane parameter set to obtain the third updated plane equation. The first centroid coordinates are obtained by calculating all the initial flange plane point cloud data using the first formula, which is: ， in, , , , in, The first centroid coordinates, The average value of the initial flange planar point cloud data along the X-axis. The average value of the Y-axis of the initial flange plane point cloud data. The average value of the Z-axis of the initial flange plane point cloud data. This represents the initial number of point cloud data points for the flange plane. For the first X-axis coordinates of the initial flange plane point cloud data. For the first Y-axis coordinates of the initial flange plane point cloud data. For the first Z-axis coordinates of the initial flange plane point cloud data; The first centroid coordinates and all the initial flange plane point cloud data are decomposed using the covariance matrix eigenvalue decomposition algorithm to obtain a first eigenvalue, a first flange eigenvector corresponding to the first eigenvalue, a second eigenvalue, a second flange eigenvector corresponding to the second eigenvalue, a third eigenvalue, and a third flange eigenvector corresponding to the third eigenvalue. The minimum value of the first feature value, the second feature value, and the third feature value is selected. After selection, the first flange feature vector, the second flange feature vector, or the third flange feature vector corresponding to the minimum value is used as the plane normal vector parameter group. The fourth plane offset is obtained by solving the plane normal vector parameter set and the first centroid coordinates using the third original plane equation; The third updated plane equation is updated with parameters based on the plane normal vector parameter set and the fourth plane offset to obtain the fourth updated plane equation. Calculate the distance between each of the initial flange plane point cloud data and the fourth updated plane equation to obtain the target distance corresponding to each of the initial flange plane point cloud data; The average error value is obtained by calculating the distances to all the targets using the second equation, which is: ， in, This is the average error value. For the first The target distance corresponding to the initial flange plane point cloud data. This represents the initial number of point cloud data points for the flange plane. If the target distance meets the determination condition, the initial flange plane point cloud data corresponding to the target distance is used as the flange plane point cloud data to be processed, thereby obtaining multiple flange plane point cloud data to be processed. The determination condition is: ， in, For the first The target distance corresponding to the initial flange plane point cloud data. This is the average error value; Determine whether the target distances all meet the determination conditions, and whether the absolute value of the difference between the average error value of the current iteration and the average error value of the previous iteration is less than a first preset threshold. If not, then reconstruct the hybrid loss function; if yes, then all the flange plane point cloud data to be processed are used as processed flange plane point cloud data, thereby obtaining multiple processed flange plane point cloud data.

4. The pre-assembly method for steel pipe arch ribs according to claim 1, characterized in that, The process of performing rotational analysis on all the processed flange plane point cloud data based on the first centroid coordinates and the fourth updated plane equation to obtain multiple target flange plane point cloud data includes: By applying equations three, four, and five respectively to the fourth updated plane equation and each of the processed flange plane point cloud data, projected flange plane point cloud data corresponding to each of the processed flange plane point cloud data is obtained. Equation three is: ， The fourth formula is: ， The fifth formula is: , in, For the first The X-axis coordinates of the projected flange plane point cloud data corresponding to the processed flange plane point cloud data. For the first The Y-axis coordinate of the projected flange plane point cloud data corresponding to the processed flange plane point cloud data. For the first The Z-axis coordinates of the projected flange plane point cloud data corresponding to the processed flange plane point cloud data. For the first The X-axis coordinates of the processed flange plane point cloud data For the first The Y-axis coordinate of the processed flange plane point cloud data. For the first The Z-axis coordinates of the processed flange plane point cloud data This is the fourth X-axis normal vector component. This is the fourth Y-axis normal vector component. This is the fourth Z-axis normal vector component. This is the offset of the fourth plane. It is a Euclidean length function; According to the preset verification rules, the accuracy of all the projected flange plane point cloud data is verified. If the verification fails, the fourth updated plane equation and each of the processed flange plane point cloud data are re-projected using the third, fourth and fifth equations respectively. If the verification is successful, the projected flange plane point cloud data is used as the flange plane point cloud data to be rotated, thereby obtaining multiple flange plane point cloud data to be rotated. Using the first centroid coordinate as the rotation center, the first rotation matrix is ​​used to rotate all the plane point cloud data of the flange to be rotated, thereby obtaining multiple target flange plane point cloud data.

5. The method for pre-assembling steel pipe arch ribs according to claim 1, characterized in that, The process of extracting control points from all the target flange plane point cloud data based on the first rotation matrix and the first centroid coordinates to obtain the initial assembly surface control point three-dimensional coordinate set of the steel pipe arch rib segment to be assembled includes: The adaptive Alpha-Shape algorithm is used to extract multiple flange contour points and multiple bolt hole contour points from all the target flange planar point cloud data, and the initial contour point dataset is obtained by combining all the flange contour points and all the bolt hole contour points. The initial contour point dataset is segmented to obtain multiple subsets of initial contour point data; The least squares algorithm is used to fit each of the initial contour point data subsets to obtain the circle equations corresponding to each of the initial contour point data subsets. A system of error function equations is constructed, and the system of error function equations and each of the circle equations are solved to obtain the initial circle center coordinates corresponding to each subset of the initial contour point data. The system of error function equations is as follows: ， in, For the first The first initial contour point data subset The initial X-axis coordinates of the contour points For the first The first initial contour point data subset The initial Y-axis coordinates of the contour points For the first The initial X-axis coordinates of the initial circle center corresponding to a subset of initial contour point data. For the first The initial Y-axis coordinates of the initial circle center corresponding to a subset of initial contour point data. Let be the error function. For the first The radius corresponding to a subset of initial contour point data. This represents the initial number of contour points; Each of the initial circle center coordinates is transformed using homogeneous coordinates to obtain the transformed circle center coordinates corresponding to each of the initial contour point data subsets. The sixth equation is used to perform three-dimensional coordinate transformation on the first rotation matrix, each subset of initial contour point data, and the transformed center coordinates corresponding to each subset of initial contour point data, respectively, to obtain the target center coordinates corresponding to each subset of initial contour point data. Then, all the target center coordinates are combined to obtain the three-dimensional coordinate set of the initial assembly surface control points of the steel pipe arch rib segment to be assembled. The sixth equation is: ， in, For the first The target circle center X-axis coordinates corresponding to a subset of initial contour point data. For the first The target circle center Y-axis coordinates corresponding to a subset of initial contour point data. For the first The Z-axis coordinates of the target circle center corresponding to a subset of initial contour point data. For the first The initial X-axis coordinates of the initial circle center corresponding to a subset of initial contour point data. For the first The initial Y-axis coordinates of the initial circle center corresponding to a subset of initial contour point data. Let be the first rotation matrix. The average value of the initial flange planar point cloud data along the X-axis. The average value of the Y-axis of the initial flange plane point cloud data. This represents the average Z-axis value of the initial flange plane point cloud data.

6. The method for pre-assembling steel pipe arch ribs according to claim 1, characterized in that, The process of performing pre-assembly analysis on the three-dimensional coordinate set of the initial assembly surface control points and the three-dimensional coordinate set of the initial adjacent assembly surface control points to obtain the pre-assembly result of the steel pipe arch rib includes: The fast four-point consistent set algorithm is used to perform coarse registration on the three-dimensional coordinate set of the initial assembly surface control points and the three-dimensional coordinate set of the initial adjacent assembly surface control points, respectively, to obtain the registered three-dimensional coordinate set of the assembly surface control points corresponding to the three-dimensional coordinate set of the initial assembly surface control points and the registered three-dimensional coordinate set of the adjacent assembly surface control points corresponding to the three-dimensional coordinate set of the initial adjacent assembly surface control points. An ICP error function is constructed, and the singular value decomposition algorithm is used to calculate the ICP error function, the three-dimensional coordinate set of the control points of the registered assembly surface, and the three-dimensional coordinate set of the control points of adjacent registered assembly surfaces to obtain the third rotation matrix and the translation matrix. The ICP error function is: ， in, This is the third rotation matrix. It is a translation matrix. This refers to the number of three-dimensional coordinates of control points on the assembled surface after registration, or the number of three-dimensional coordinates of control points on adjacent assembled surfaces after registration. The first of the three-dimensional coordinate sets of the control points of adjacent assembly surfaces after registration Three-dimensional coordinates of adjacent assembly surface control points after registration The first of the three-dimensional coordinate sets of the control points of the assembly surface after registration Three-dimensional coordinates of the control points of the assembly surface after registration; The three-dimensional coordinate set of the control points of the registered assembly surface is obtained by calculating the third rotation matrix and the translation matrix using the seventh equation. The seventh equation is: ， in, This is the three-dimensional coordinate set of the control points of the assembly surface to be processed. This is the third rotation matrix. It is a translation matrix. This is the three-dimensional coordinate set of the control points for the assembly surface after registration. The average target distance is obtained by calculating the three-dimensional coordinate set of the control points of the assembly surface to be processed and the three-dimensional coordinate set of the control points of the adjacent assembly surfaces after registration using the eighth equation. The eighth equation is: ， in, The average distance to the target. This refers to the number of three-dimensional coordinates of control points on the assembled surface after registration, or the number of three-dimensional coordinates of control points on adjacent assembled surfaces after registration. The first of the three-dimensional coordinate sets of the control points of adjacent assembly surfaces after registration Three-dimensional coordinates of adjacent assembly surface control points after registration The first of the three-dimensional coordinates of the control points of the assembly surface to be processed Three-dimensional coordinates of control points on the assembly surface to be processed; Determine whether the average distance to the target is less than the second preset threshold. If not, then use the fast four-point consistent set algorithm to perform coarse registration on the three-dimensional coordinate set of the initial assembly surface control points and the three-dimensional coordinate set of the initial adjacent assembly surface control points respectively. If yes, then use the third rotation matrix and the translation matrix together as the pre-assembly result of the steel pipe arch rib.

7. A pre-assembly device for steel pipe arch ribs, characterized in that, include: The scanning module is used to scan the steel pipe arch rib segment to be assembled and the neighboring steel pipe arch rib segment adjacent to the steel pipe arch rib segment to be assembled using a 3D laser scanner, so as to obtain multiple original point cloud data of the steel pipe arch rib segment to be assembled and multiple original adjacent point cloud data of the neighboring steel pipe arch rib segment. The control point analysis module is used to perform control point analysis on multiple original point cloud data of the steel pipe arch rib segment to be assembled and multiple original adjacent point cloud data of the neighboring steel pipe arch rib segment, respectively, to obtain the three-dimensional coordinate set of the initial assembly surface control points of the steel pipe arch rib segment to be assembled and the three-dimensional coordinate set of the initial adjacent assembly surface control points of the neighboring steel pipe arch rib segment. The pre-assembly result acquisition module is used to perform pre-assembly analysis on the three-dimensional coordinate set of the control points of the initial assembly surface and the three-dimensional coordinate set of the control points of the initial adjacent assembly surfaces to obtain the pre-assembly result of the steel pipe arch rib. The control point analysis module is specifically used for: The flange plane is the assembly surface of the arch rib segment. Registration analysis is performed on all the original point cloud data to obtain multiple initial flange plane point cloud data. Preprocess all the initial flange plane point cloud data to obtain the first rotation matrix, the first centroid coordinates, and multiple target flange plane point cloud data; Based on the first rotation matrix and the first centroid coordinates, control points are extracted from all the target flange plane point cloud data to obtain the initial assembly surface control point three-dimensional coordinate set of the steel pipe arch rib segment to be assembled; Registration analysis was performed on all the original adjacent point cloud data to obtain multiple initial adjacent flange planar point cloud data; Preprocess all the initial adjacent flange plane point cloud data to obtain the second rotation matrix, the second centroid coordinates, and multiple target adjacent flange plane point cloud data; Based on the second rotation matrix and the second centroid coordinates, control points are extracted from the plane point cloud data of all adjacent flanges of the target to obtain the initial three-dimensional coordinate set of control points of the adjacent assembly surface of the neighboring steel pipe arch rib segment; In the control point analysis module, the process of preprocessing all the initial flange plane point cloud data to obtain the first rotation matrix, the first centroid coordinates, and multiple target flange plane point cloud data includes: Denoising analysis is performed on all the initial flange plane point cloud data to obtain the first centroid coordinates, the fourth updated plane equation, and multiple processed flange plane point cloud data. The first centroid coordinates are the average coordinates of the initial flange plane point cloud data. Based on the first centroid coordinates and the fourth updated plane equation, a rotational analysis is performed on all the processed flange plane point cloud data to obtain multiple target flange plane point cloud data. The first included angle and the second included angle are extracted from the fourth updated plane equation. The fourth updated plane equation is the best fitting plane of the initial flange plane point cloud data. The first included angle is the angle between the normal vector of the fourth updated plane equation and the Z-axis under the projection of the YOZ coordinate system. The second included angle is the angle between the normal vector of the fourth updated plane equation and the Z-axis under the projection of the XOZ coordinate system. A first rotation matrix is ​​constructed using the first included angle and the second included angle. The first rotation matrix is: ， in, Let be the first rotation matrix. The first included angle, This is the second included angle.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the pre-assembly method for steel pipe arch ribs as described in any one of claims 1 to 6.

Images