Bridge steel tower component size detection method and system based on three-dimensional laser scanning

Abstract

The invention discloses a bridge steel tower component size detection method and system based on three-dimensional laser scanning. The method comprises the following steps: acquiring overall point cloud data including a target steel tower component by adopting a three-dimensional laser scanner; obtaining effective point cloud data of the target steel tower component according to the overall point cloud data; performing clustering processing on the effective point cloud data by adopting a DBSCAN algorithm to obtain point cloud data of all units of the target steel tower component; processing the point cloud data of the steel pipe by adopting an RANSAC algorithm to obtain the diameter of the steel pipe; processing the ribbed steel plate point cloud data through a normal vector-based segmentation method, a DBSCAN algorithm, a voxelization sampling algorithm and an RANSAC algorithm to obtain a rib plate distance; processing the segment point cloud data by adopting an RANSAC algorithm to obtain the length and the width of the side panel, and performing slicing and two-dimensional plane projection after rotating the segment point cloud data to obtain the overall size of the segment; the size of the bridge steel tower component can be measured in a non-contact mode, and the problem that traditional manual measurement is low in efficiency is effectively solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of information technology and building safety management, and more specifically, relates to a method and system for detecting the size of bridge steel tower components based on three-dimensional laser scanning. Background Art

[0002] During the construction of bridge steel towers, due to restrictions on site, transportation, and hoisting, small components are often welded into segments first, and then transported to the construction site for assembly. Before assembly, the dimensions and quality of each component and segment need to be inspected. Traditional inspection methods are mostly for workers to use tools such as rulers and tape measures to complete the dimensional quality inspection of components and segments, which is inefficient and has large subjective errors among workers, making it impossible to accurately and efficiently control the dimensional quality of each processing stage. Summary of the invention

[0003] In view of the above defects or improvement needs of the prior art, the present invention provides a bridge steel tower component size detection method and system based on three-dimensional laser scanning. By combining multiple advanced technologies such as three-dimensional laser scanning technology, RANSAC algorithm, pass-through filtering algorithm, DBSCAN algorithm, Rodrigues formula and least squares method, a high-precision, automated and robust non-contact detection process for the size of bridge steel tower components is realized, which not only improves the accuracy and efficiency of measurement, but also enhances the reliability of data and the flexibility of processing, providing strong technical support for quality control, structural health monitoring, maintenance and repair of bridge projects, while reducing costs and improving the economic benefits of projects.

[0004] In order to achieve the above object, one aspect of the present invention provides a method for detecting the size of a bridge steel tower component based on three-dimensional laser scanning, comprising the following steps:

[0005] S1: Use a 3D laser scanner to scan a specific area containing the target steel tower components to obtain overall point cloud data containing the target steel tower components;

[0006] S2: using a straight-through filtering algorithm to process background noise in the overall point cloud data to obtain valid point cloud data of the target steel tower component;

[0007] S3: clustering the effective point cloud data using the DBSCAN algorithm to obtain all unit point cloud data of the target steel tower component; all unit point cloud data of the target steel tower component include steel pipe point cloud data, ribbed steel plate point cloud data and segment point cloud data;

[0008] S4: using the RANSAC algorithm to perform steel pipe axis fitting and noise removal on the steel pipe point cloud data to obtain steel pipe outer wall point cloud data; and obtaining the steel pipe diameter according to the steel pipe outer wall point cloud data;

[0009] S5: Processing the point cloud data of the ribbed steel plate by a normal vector-based segmentation method and a DBSCAN algorithm to obtain point cloud data of instantiated transverse plate units and longitudinal plate units; processing the point cloud data of the instantiated transverse plate units and longitudinal plate units by a voxelized sampling algorithm and a RANSAC algorithm to obtain the rib spacing;

[0010] S6: Use the RANSAC algorithm to perform multi-plane fitting extraction on the segment point cloud data to obtain the segment side panel point cloud data; use the cross product of the normal vector of the side panel and the Z axis as the front direction of the segment, rotate the original segment so that the front direction is parallel to the Y axis, and obtain the rotated segment point cloud data;

[0011] S7: obtaining the length and width of the segment side panel according to the segment side panel point cloud data; obtaining the overall size of the segment according to the rotated segment point cloud data;

[0012] S8: Obtaining a target value of the overall size of the segment of the target steel tower component according to the size of the target steel tower component processing drawing and the processing deformation of the target steel tower component;

[0013] S9: Obtain an error value of the segment's overall size based on the segment's overall size in step S7 and a target value of the segment's overall size in step S8; and mark and prompt the segment size that does not meet the manufacturing accuracy requirements based on a comparison of the error value and the accuracy value required by manufacturing.

[0014] Furthermore, the step S2 uses a straight-through filtering algorithm to process the background noise in the overall point cloud data containing the target steel tower component to obtain valid point cloud data of the target steel tower component; including:

[0015] Construct a bounding box larger than the target steel tower component, and confine the target steel tower component to a range in the three dimensions of x, y, and z;

[0016] The point cloud direct filtering algorithm is used to remove the noise data outside the bounding box, retaining the target steel tower components and some surrounding noise.

[0017] Furthermore, the step S4 uses the RANSAC algorithm to perform steel pipe axis fitting and noise removal on the steel pipe point cloud data in all unit point cloud data of the target steel tower component to obtain steel pipe outer wall point cloud data; including:

[0018] S411: Dividing the steel pipe point cloud data into a sample point set and a test point set;

[0019] S412: Using the RANSAC algorithm to perform steel pipe axis fitting on the sample point set data to obtain a steel pipe fitting axis equation;

[0020] S413: Calculate the distance between each point in the test point set data and the steel pipe fitting axis according to the steel pipe axis equation, set a distance threshold, remove the noise points on the inner wall of the steel pipe according to the distance threshold, and obtain the point cloud data of the outer wall of the steel pipe;

[0021] Furthermore, the step S4 of obtaining the diameter of the steel pipe according to the point cloud data of the outer wall of the steel pipe includes:

[0022] S421: Rotate the steel pipe axis to be parallel to the z-axis; calculate the rotation matrix of the steel pipe axis using Rodriguez formula;

[0023] S422: using the rotation matrix to transform the coordinates of each point in the steel pipe outer wall point cloud data from the coordinate system before rotation to the coordinate system after rotation;

[0024] S423: Projecting the rotated steel pipe axis onto the xoy plane to obtain two-dimensional circle data;

[0025] S424: Use the least squares method to perform circle fitting on the two-dimensional circle data to obtain the center coordinates and radius of the best fitting circle, and then obtain the diameter of the steel pipe.

[0026] Furthermore, step S5 processes the point cloud data of the ribbed steel plate by a segmentation method based on a normal vector and a DBSCAN algorithm to obtain point cloud data of instantiated transverse plate units and longitudinal plate units; including:

[0027] Distinguishing and extracting the point cloud data of the transverse plate and the longitudinal plate from the point cloud data of the ribbed steel plate by a segmentation method based on a normal vector;

[0028] The DBSCAN algorithm is used to instantiate and segment the point cloud data of the transverse plate and the longitudinal plate to obtain the point cloud data of the instantiated transverse plate unit and the longitudinal plate unit;

[0029] The step S5 uses a voxel sampling algorithm and a RANSAC algorithm to perform plane fitting on the point cloud data of the instantiated transverse plate unit and the longitudinal plate unit to obtain the rib spacing; including:

[0030] The plate units are processed by voxel sampling algorithm to obtain uniformly distributed plate units. The center coordinates of each plate unit are calculated, and the common normal vector of the plate unit is obtained after plane fitting by RANSAC algorithm. The distance between the center coordinates of each transverse plate and longitudinal plate along the common normal vector is taken as the rib spacing.

[0031] Furthermore, in step S6, the segment point cloud data is subjected to segment multi-plane fitting extraction using the RANSAC algorithm to obtain segment side panel data; including:

[0032] S611: Use the RANSAC algorithm to perform multi-plane fitting on the segment point cloud data. The RANSAC algorithm preferentially fits the plane containing the most point clouds. Each time a plane is fitted, the point cloud data of the plane is stored in a list, and the fitted plane data is deleted from the remaining segment point cloud data.

[0033] S612: Repeat step S61 to continue plane fitting until the number of planes in the list reaches a preset threshold, and then exit the fitting process, i.e., obtain the key plane panels (such as side panels, bottom panels, etc.) of each segment;

[0034] S613: Filter out side panels from the key plane panels of each segment.

[0035] Furthermore, in step S7, the length and width of the segment side panel are obtained according to the segment side panel point cloud data; including:

[0036] S711: using a statistical filtering algorithm to remove noise point clouds near the edge of the rectangular plate in the segment side panel point cloud data, rotating the side panel so that the normal vector of the reference plane is parallel to the Z axis and projecting it to obtain two-dimensional side panel data;

[0037] S712: Calculate the length and width of the side panel according to the two-dimensional side panel data;

[0038] Step S712 calculates the length and width of the side panel according to the two-dimensional side panel data; including:

[0039] An edge extraction algorithm is used to identify edge points in the two-dimensional side panel data; a RANSAC algorithm is used to perform straight line fitting on the edge points to obtain an edge line equation; an intersection of the fitted edge lines is calculated according to the edge line equation to determine the diagonal point of the rectangular panel; a distance formula is used to calculate the distance from the intersection to the edge of the rectangular panel to obtain the length and width of the side panel;

[0040] Furthermore, the step S7 of obtaining the overall size of the segment according to the rotated segment point cloud data includes:

[0041] S721: Slice the rotated segment point cloud data along the Y-axis direction, take partial slices at the front and rear segments, extract key structure point cloud data for segment slice size detection using a normal vector segmentation algorithm and a straight-through filtering algorithm, and project the key structure point cloud data onto the xoz plane to obtain segment two-dimensional point cloud data;

[0042] S722: Use the RANSAC algorithm to perform straight line fitting on the segment two-dimensional point cloud data, obtain the equation of each straight line, and calculate the coordinates of the intersection of each straight line equation; for data that needs to fit circular features, use the least squares method to estimate the coordinates of the center of the circle, and finally use the distance formula to calculate the distance between the key intersection points to obtain the overall size of the segment.

[0043] Furthermore, the step S8 of obtaining the target value of the overall size of the segment of the target steel tower component according to the size of the target steel tower component processing drawing and the processing deformation of the target steel tower component includes:

[0044] S81: establishing an accurate geometric model of the target steel tower component according to the processing drawing of the target steel tower component, meshing the geometric model using finite element analysis software, and constructing a finite element model of the target steel tower component;

[0045] S82: Determine the boundary constraint condition of the target steel tower component in the finite element model according to the placement state of the target steel tower component at the processing site;

[0046] S83: Calculate the influence of temperature change on material properties according to the temperature difference between the ambient temperature recorded during the on-site scanning and the temperature during processing, and apply a corresponding temperature load in the finite element model; apply the dead weight of the steel tower component in the finite element model;

[0047] S84: Run the finite element analysis software to calculate the deformation of the target steel tower components under the effects of temperature and deadweight;

[0048] S85: Add the deformation amount to the size on the target steel tower component processing drawing to obtain the target value of the overall size of the segment of the target steel tower component taking into account the deformation.

[0049] A second aspect of the present invention provides a bridge steel tower component size detection system based on three-dimensional laser scanning, which is used to implement the bridge steel tower component size detection method based on three-dimensional laser scanning, comprising:

[0050] The first main module is used to scan a specific area including a target steel tower component using a 3D laser scanner to obtain overall point cloud data including the target steel tower component;

[0051] The second main module is used to process the background noise in the overall point cloud data using a straight-through filtering algorithm to obtain effective point cloud data of the target steel tower component;

[0052] The third main module is used to perform clustering processing on the effective point cloud data using the DBSCAN algorithm to obtain all unit point cloud data of the target steel tower component; all unit point cloud data of the target steel tower component include steel pipe point cloud data, ribbed steel plate point cloud data and segment point cloud data;

[0053] The fourth main module is used to use the RANSAC algorithm to perform steel pipe axis fitting and noise removal on the steel pipe point cloud data to obtain the steel pipe outer wall point cloud data; and obtain the steel pipe diameter according to the steel pipe outer wall point cloud data;

[0054] The fifth main module is used to process the point cloud data of the ribbed steel plate by using a normal vector-based segmentation method and a DBSCAN algorithm to obtain point cloud data of instantiated transverse plate units and longitudinal plate units; and to process the point cloud data of the instantiated transverse plate units and longitudinal plate units by using a voxel sampling algorithm and a RANSAC algorithm to obtain the rib spacing;

[0055] The sixth main module is used to perform multi-plane fitting extraction of the segment point cloud data using the RANSAC algorithm to obtain the segment side panel point cloud data; the cross product of the normal vector of the side panel and the Z axis is used as the front direction of the segment, and the original segment is rotated so that the front direction is parallel to the Y axis to obtain the rotated segment point cloud data;

[0056] The seventh main module is used to obtain the length and width of the segment side panel according to the segment side panel point cloud data; and obtain the overall size of the segment according to the rotated segment point cloud data;

[0057] The eighth main module is used to obtain the target value of the segment overall size of the target steel tower component according to the size of the target steel tower component processing drawing and the processing deformation amount of the target steel tower component;

[0058] The ninth main module is used to obtain the error value of the overall size of the segment according to the overall size of the segment and the target value of the overall size of the segment; and to mark and prompt the segment size that does not meet the manufacturing accuracy requirements by comparing the error value with the accuracy value required by manufacturing.

[0059] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0060] (1) The present invention provides a bridge steel tower component size detection method and system based on 3D laser scanning. By comprehensively applying high-precision 3D laser scanning technology and combining powerful data processing algorithms such as RANSAC algorithm, direct filtering algorithm and DBSCAN algorithm, accurate non-contact measurement of the size of bridge steel tower components is achieved. Specifically, by fitting the steel pipe axis and removing noise through the RANSAC algorithm, more accurate steel pipe outer wall point cloud data can be obtained, thereby improving the accuracy of diameter measurement. The introduction of the RANSAC algorithm significantly improves the robustness to noise and outliers, ensuring that reliable geometric feature fitting can be obtained even under complex or imperfect data conditions. Background noise is effectively removed through direct filtering, and the target component is accurately separated through the DBSCAN algorithm, laying a solid foundation for subsequent size analysis. The rotation matrix is ​​calculated through the Rodriguez formula, and the scheme can rotate the steel pipe axis. The projected two-dimensional circle data is rotated to be parallel to the reference axis, and then the least squares method is used to perform circle fitting on the projected two-dimensional circle data, so as to accurately calculate the diameter of the steel pipe; the processing of the ribbed steel plate point cloud data is further optimized by the voxel sampling algorithm and the RANSAC algorithm, ensuring the accurate measurement of the rib spacing; the RANSAC algorithm is used to effectively extract accurate multi-faceted information from the complex segment point cloud data, especially for the identification of the side panel; by calculating the cross product of the side panel normal vector and the axis, the front direction of the segment can be clearly determined, which helps to standardize the measurement process and ensure that all dimensional measurements are performed in the correct reference frame; the original segment is rotated to the front direction parallel to the axis, and the rotated two-dimensional data is obtained to simplify the subsequent dimensional measurement work; the present invention can provide comprehensive dimensional information for the design verification, manufacture, installation and maintenance of bridge steel tower components, thereby ensuring that the manufacturing and installation quality of the components meet the design and safety standards.

[0061] (2) The present invention provides a bridge steel tower component dimension detection method and system based on three-dimensional laser scanning. By establishing a finite element model of the target steel tower component, the target steel tower component is subjected to deformation analysis under the action of temperature difference and deadweight. The corresponding deformation amount is added to the dimension on the processing drawing of the target steel tower component to obtain the target value of the overall dimension of the segment of the target steel tower component. The measured value of the overall dimension of the segment minus the target value of the overall dimension of the segment is used as the error value of the overall dimension of the segment given by the detection. This is more accurate and reasonable than the traditional method of simply taking the dimension on the processing drawing as the target value without considering the load deformation of the component, and improves the actual guiding value of the detection result.

[0062] (3) The present invention provides a method and system for detecting the size of bridge steel tower components based on three-dimensional laser scanning. Through automated data processing and analysis processes, manual operations are reduced, work efficiency and consistency of measurement results are improved, and the possibility of human error is reduced. Precise alignment of point cloud data obtained from different angles is ensured through fine registration technology and multi-site scanning strategies, thereby improving the accuracy of the overall data. The method of the present invention can handle bridge steel tower components of various shapes and types, and can provide comprehensive quality control and structural health monitoring means for all stages of bridge engineering, from design verification to construction monitoring, and then to maintenance and repair. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 It is an internal logic schematic diagram of a bridge steel tower component size detection method based on three-dimensional laser scanning according to an embodiment of the present invention;

[0064] Figure 2 The noise removal result of the inner wall of the steel pipe in the embodiment of the present invention is shown;

[0065] Figure 3 The ribbed steel plate segmentation result of the embodiment of the present invention;

[0066] Figure 4 The instantiation segmentation results of the horizontal plate and the vertical plate of the implementation example of the present invention;

[0067] Figure 5 It is a segment size measurement diagram of an embodiment of the present invention;

[0068] Figure 6 The side panel edge detection result of the embodiment of the present invention;

[0069] Figure 7 It is the segment two-dimensional projection result of the implementation example of the present invention;

[0070] Figure 8 This is a schematic structural diagram of a bridge steel tower component size detection system based on three-dimensional laser scanning according to an embodiment of the present invention;

[0071] Fig. 9 A schematic diagram of the structure of an electronic device according to an embodiment of the present invention;

[0072] Fig.10 The present invention is a flowchart of a method for detecting the size of a bridge steel tower component based on three-dimensional laser scanning according to an embodiment of the present invention.

[0073] It should be noted that:

[0074] DBSCAN is the abbreviation of Density-Based Spatial Clustering of Applications with Noise, which means density-based spatial clustering algorithm;

[0075] RANSAC is the abbreviation of RANdom SAmple Consensus, which means random sampling consensus algorithm; DETAILED DESCRIPTION

[0076] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0077] like Figure 1 As shown, one aspect of the present invention provides a bridge steel tower component size detection method based on three-dimensional laser scanning, which is used for automatic non-contact detection of unit size of bridge steel tower components, comprising the following steps:

[0078] S1: Use a 3D laser scanner to scan a specific area containing the target steel tower components to obtain overall point cloud data containing the target steel tower components;

[0079] S2: using a straight-through filtering algorithm to process the overall point cloud data of the target steel tower component to obtain valid point cloud data of the target steel tower component;

[0080] S3: clustering the effective point cloud data of the target steel tower component using the DBSCAN algorithm to obtain all unit point cloud data of the target steel tower component; all unit point cloud data of the target steel tower component include steel pipe point cloud data, ribbed steel plate point cloud data and segment point cloud data;

[0081] S4: using the RANSAC algorithm to perform steel pipe axis fitting and noise removal on the steel pipe point cloud data to obtain steel pipe outer wall point cloud data; and obtaining the steel pipe diameter according to the steel pipe outer wall point cloud data;

[0082] S5: Processing the point cloud data of the ribbed steel plate by a normal vector-based segmentation method and a DBSCAN algorithm to obtain point cloud data of instantiated transverse plate units and longitudinal plate units; processing the point cloud data of the instantiated transverse plate units and longitudinal plate units by a voxelized sampling algorithm and a RANSAC algorithm to obtain the rib spacing;

[0083] S6: Use the RANSAC algorithm to perform multi-plane fitting extraction on the segment point cloud data to obtain the segment side panel point cloud data; use the cross product of the normal vector of the side panel and the Z axis as the front direction of the segment, rotate the original segment so that the front direction is parallel to the Y axis, and obtain the rotated segment point cloud data;

[0084] S7: obtaining the length and width of the segment side panel according to the segment side panel point cloud data; obtaining the overall size of the segment according to the rotated segment point cloud data;

[0085] S8: Obtaining a target value of the overall size of the segment of the target steel tower component according to the size of the target steel tower component processing drawing and the processing deformation of the target steel tower component;

[0086] S9: Obtain an error value of the segment's overall size based on the segment's overall size in step S7 and a target value of the segment's overall size in step S8; and mark and prompt the segment size that does not meet the manufacturing accuracy requirements based on a comparison of the error value and the accuracy value required by manufacturing.

[0087] Furthermore, the three-dimensional laser scanner described in step S1 is used to scan the specific area containing the target steel tower component to obtain the overall point cloud data containing the target steel tower component; including: setting different stations around the target steel tower component, placing at least 3 public target balls between two adjacent stations, and completing the scanning of the point cloud data around the target steel tower component; aligning the cloud data of each station obtained by the scan, so that the target balls of the two adjacent stations are strongly corresponding to each other, ensuring that the position deviation is within 1mm, completing the fine alignment of the point cloud, and obtaining the overall point cloud data containing the target steel tower component.

[0088] Furthermore, there is a large amount of background noise data around the overall point cloud data containing the target steel tower component scanned in step S1, and most of the background noise data needs to be removed. The step S2 uses a straight-through filtering algorithm to process the background noise in the overall point cloud data containing the target steel tower component to obtain the effective point cloud data of the target steel tower component; including: using a point cloud straight-through filtering algorithm to pre-process the overall point cloud data of the target steel tower component. The basic idea of ​​the point cloud straight-through filtering algorithm is to limit the point cloud data to a certain range of the three dimensions of x, y, and z. This method can effectively remove most of the background noise data in the overall point cloud data of the target steel tower component and obtain the effective point cloud data of the target steel tower component; in specific implementation, a bounding box slightly larger than the target steel tower component is constructed (the target steel tower component is limited to a range by the three dimensions of x, y, and z), the noise data outside the bounding box is removed, and the target steel tower component and a small amount of noise around it are retained.

[0089] Furthermore, after the through-filtering in step S2 removes most of the background noise data, a small amount of noise data still exists around the target steel tower component; the DBSCAN algorithm is used in step S3 to cluster the effective point cloud data to obtain all unit point cloud data of the target steel tower component; including: according to the point cloud density, the DBSCAN algorithm is used to spatially cluster the target steel tower component point cloud data and the noise point cloud data in the effective point cloud data of the target steel tower component to obtain all unit point cloud data of the target steel tower component; all unit point cloud data of the target steel tower component include steel pipe point cloud data , ribbed steel plate point cloud data and segment point cloud data; wherein, the DBSCAN algorithm identifies the core point by detecting whether the number of points within the neighborhood distance ε of the data point reaches the minimum value MinPts. The core point and the points in its neighborhood form a cluster, and the points that do not meet the core point conditions may be marked as noise; this algorithm does not require a preset number of clusters, can effectively identify clusters of complex shapes and resist noise interference; in the specific implementation of the present invention, the neighborhood distance ε is set to 2, the minimum number of points is set to 1000, and the target component and the noise data are divided into different classes to obtain the target component data.

[0090] Furthermore, in step S4, the RANSAC algorithm is used to perform steel pipe axis fitting and noise removal on the steel pipe point cloud data in all unit point cloud data of the target steel tower component to obtain steel pipe outer wall point cloud data; including:

[0091] S411: Dividing the steel pipe point cloud data into a sample point set and a test point set;

[0092] S412: Using the RANSAC algorithm to perform steel pipe axis fitting on the sample point set data to obtain a steel pipe fitting axis equation;

[0093] S413: Calculate the distance between each point in the test point set data and the steel pipe fitting axis according to the steel pipe axis equation, set a distance threshold, remove the noise points on the inner wall of the steel pipe according to the distance threshold, and obtain the point cloud data of the outer wall of the steel pipe;

[0094] In step S4, the RANSAC algorithm randomly extracts a part of the steel pipe point cloud data as a sample point set, which usually contains two points, enough to define a straight line; the parameters of the axis fitting model are estimated according to the sample point set, and the model parameters will be the equation of the straight line, usually expressed as a two-point form or a parameter form; then the remaining point set in the steel pipe point cloud data is used to verify whether the required axis fitting model is optimal. Specifically, for each point, its distance to the fitting axis is calculated and compared with a preset distance threshold; those points whose distance is less than the threshold are considered to be inliers, that is, they conform to the model; repeat the above sampling, parameter estimation and verification. The verification process is iterated multiple times; each iteration may generate a new axis fitting model; the model supporting the largest number of internal points is selected as the optimal model; this model is considered to be the model that best represents the axis of the steel pipe, that is, the optimal axis fitting model is obtained; in the specific implementation of the present invention, the RANSAC algorithm is used for axis fitting, and the minimum distance sigma from the point to the cylindrical surface is set to 0.02, the number of randomly sampled points n is set to 5, and the maximum number of iterations is set to 1000, and the axis equation is obtained; a distance threshold is set, and the distance from all points in the steel pipe to the axis is calculated, and the points on the inner wall of the steel pipe are removed to obtain the point cloud data of the outer wall of the steel pipe.

[0095] Furthermore, if is a linear direction vector, and the coordinates of a point on the steel pipe fitting axis are B=(x0, y0, z0). Then, the steel pipe fitting axis equation in step S4 is:

[0096]

[0097] The coordinates of a point in the test point set data of the steel pipe point cloud data are A (x1, y1, z1), then: the calculation formula of the distance d from point A to the axis is:

[0098]

[0099] Furthermore, in step S4, the diameter of the steel pipe is obtained according to the point cloud data of the outer wall of the steel pipe, including:

[0100] S421: Rotate the steel pipe axis to be parallel to the z-axis; calculate the rotation matrix of the steel pipe axis using Rodriguez formula;

[0101] S422: using the rotation matrix to transform the coordinates of each point in the steel pipe outer wall point cloud data from the coordinate system before rotation to the coordinate system after rotation;

[0102] S423: Project the rotated steel pipe axis onto the xoy plane, ignore the Z coordinate, and obtain two-dimensional circle data;

[0103] S424: performing circle fitting on the two-dimensional circle data using the least squares method to obtain the center coordinates and radius of the best fitting circle, thereby obtaining the diameter of the steel pipe;

[0104] In step S421, if the axis normal vector Z-axis vector Then the rotation matrix R of the steel pipe axis is calculated by the following formula:

[0105]

[0106] θ=arccos(cosθ)(5)

[0107]

[0108] R=I+sinθK+(1-cosθ)K 2 (7)

[0109] Where I is the identity matrix;

[0110] In step S422, if the coordinates of a certain point in the steel pipe outer wall point cloud data before rotation are M0 (x0, y0, z0), the coordinates after rotation M2 (x2, y2, z2) are calculated using the following formula:

[0111] M2=RM0 (8);

[0112] In step S423, the least square method is used to perform circle fitting, and the circle diameter is the steel pipe diameter, as shown in Table 1, which is the intelligent detection value, target value and absolute error of the steel pipe diameter in the implementation example of the present invention;

[0113] Table 1

[0114]

[0115] Furthermore, step S5 processes the point cloud data of the ribbed steel plate by a segmentation method based on a normal vector and a DBSCAN algorithm to obtain point cloud data of instantiated transverse plate units and longitudinal plate units; including:

[0116] The point cloud data of the transverse plate and the longitudinal plate are distinguished and extracted from the point cloud data of the ribbed steel plate by a segmentation method based on a normal vector; the point cloud data of the transverse plate and the longitudinal plate are instantiated and segmented by a DBSCAN algorithm to obtain the point cloud data of the instantiated transverse plate unit and the longitudinal plate unit;

[0117] In a specific implementation of the present invention, the ribbed steel plate includes a transverse plate, a longitudinal plate and the steel plate itself, and the reference normal vectors of the three are perpendicular to each other. The transverse plate, the longitudinal plate and the steel plate itself are segmented from the ribbed steel plate point cloud data using a normal vector-based method (i.e., according to the direction of the normal vector), such as Figure 3 As shown;

[0118] The normal vector is calculated based on the principal component analysis (PCA) method, and the process is as follows:

[0119] First, for each point, find its neighbor point set P = {p1, p2, p3, ..., p k};

[0120] Next, calculate the covariance matrix C of the point set P;

[0121] Then, calculate the eigenvalues ​​and eigenvectors of the covariance matrix C;

[0122] Finally, sort the eigenvectors from large to small according to the corresponding eigenvalues, and select the largest eigenvector as the normal vector of the point;

[0123] In the specific implementation of the present invention, when the DBSCAN algorithm is used to instantiate and segment the point cloud data of the transverse plate and the longitudinal plate, the number of neighbor points neighbor_number=2000 is set to obtain the point cloud data of the instantiated transverse plate unit and the longitudinal plate unit, respectively. Figure 4 shown.

[0124] Furthermore, the step S5 uses a voxel sampling algorithm and a RANSAC algorithm to perform plane fitting on the point cloud data of the instantiated transverse plate unit and the longitudinal plate unit to obtain the rib spacing; including:

[0125] The plate elements are processed by voxel sampling algorithm to obtain uniformly distributed plate elements; the center coordinates of each plate element are calculated, and the common normal vector of the plate element is obtained after plane fitting by RANSAC algorithm, and the distance between the center coordinates of each transverse plate and the longitudinal plate along the common normal vector is taken as the rib spacing;

[0126] In the specific implementation of the present invention, voxel sampling is performed on each plate, and the voxel voxel_size is set to 0.005 to obtain uniformly distributed plate units; the central coordinates of each plate unit are calculated, that is, the average values ​​of x, y, and z are respectively calculated, and the RANSAC algorithm is used to perform plane fitting on each plate, and the minimum distance sigma from the point to the plane is set to 0.012, the number of randomly sampled points n is set to 5, and the maximum number of iterations is set to 1000, and the normal vector of each plate is obtained, and the average values ​​of the normal vectors of the transverse plate and the longitudinal plate are respectively calculated as the reference normal vectors of the transverse plate and the longitudinal plate; finally, the distances of the central coordinates of each plate in the transverse plate and the longitudinal plate along the reference normal vector are respectively calculated as the rib spacing; as shown in Table 2, the intelligent detection value, target value and absolute error of the rib spacing of the ribbed steel plate of the implementation example of the present invention are shown.

[0127] Table 2

[0128]

[0129] Furthermore, in step S6, the segment point cloud data is subjected to segment multi-plane fitting extraction using the RANSAC algorithm to obtain segment side panel data; including:

[0130] S611: Use the RANSAC algorithm to perform multi-plane fitting on the segment point cloud data. The RANSAC algorithm preferentially fits the plane containing the most point clouds. Each time a plane is fitted, the point cloud data of the plane is stored in a list, and the fitted plane data is deleted from the remaining segment point cloud data.

[0131] S612: Repeat step S61 to continue plane fitting until the number of planes in the list reaches a preset threshold, and then exit the fitting process, i.e., obtain the key plane panels (such as side panels, bottom panels, etc.) of each segment;

[0132] S613: Filter out side panels from the key plane panels of each segment;

[0133] Further, in a specific embodiment of the present invention, step S6 includes first marking the size of the segment as a whole that needs to be detected, such as Figure 5 As shown; the RANSAC algorithm is used to perform multi-plane fitting on the segment point cloud data, and the minimum distance from the point to the plane is set to sigma = 0.012, the number of randomly sampled points n = 5, and the maximum number of iterations = 1000. In the process of RANSAC plane fitting, the plane that can fit more point clouds will be selected first. Each time a plane is fitted, the current plane point cloud data is stored in a list, and the currently fitted plane data is deleted from the segment for plane fitting again. When the number of planes in the list reaches a certain threshold, the fitting is exited to obtain the key plane panels of each segment; the side panels are selected from the key plane panels obtained.

[0134] Furthermore, in step S7, the length and width of the segment side panel are obtained according to the segment side panel point cloud data; including:

[0135] S711: using a statistical filtering algorithm to remove noise point clouds near the edge of the rectangular plate in the segment side panel point cloud data, rotating the side panel so that the normal vector of the reference plane is parallel to the Z axis and projecting it to obtain two-dimensional side panel data;

[0136] S712: Calculate the length and width of the side panel according to the two-dimensional side panel data.

[0137] Furthermore, in a specific embodiment of the present invention, step S711 includes: setting the statistical filtering neighborhood parameter k=100 to complete the denoising of the side panel; rotating the side panel until the normal vector of the reference plane is parallel to the axis and projecting it to obtain two-dimensional side panel data.

[0138] Furthermore, in step S712, the length and width of the side panel are calculated according to the two-dimensional side panel data; including:

[0139] An edge extraction algorithm is used to identify edge points in the two-dimensional side panel data; a RANSAC algorithm is used to perform straight line fitting on the edge points to obtain an edge line equation; an intersection of the fitted edge line is calculated according to the edge line equation to determine the diagonal point of the rectangular plate; a distance formula is used to calculate the distance from the intersection to the edge of the rectangular plate to obtain the length and width of the side panel.

[0140] In a specific embodiment of the present invention, in step S712, since the side panel is a rectangle, an edge extraction algorithm is used to extract the edge of the side panel. i , construct p i The vector set Q(n1,n2,n3...n i ), randomly select any vector n from the set Q i As a reference vector; traverse the remaining vectors in set Q. If the current vector rotates clockwise to n i If the angle is smaller, the current rotation angle is saved in the clockwise angle set R1, otherwise the counterclockwise rotation angle is smaller and saved in the counterclockwise angle set R2. The maximum angle in R1 is α, and the maximum angle in R2 is β. Let S = α + β; if s is close to 180°, it is considered that the current point p i is a point near the edge line; if s is close to 90°, the current point p is considered i is a point near the corner point; the edge detection result of the rectangular plate is as follows Figure 6 As shown, the obtained edge points are fitted with straight lines to obtain four straight line equations, the intersection coordinates are obtained, and then the length and width of the rectangular plate are calculated according to the distance formula.

[0141] Furthermore, the step S7 of obtaining the overall size of the segment according to the rotated segment point cloud data includes:

[0142] S721: Slice the rotated segment along the Y-axis direction, extract the point cloud data of key components for size detection of the segment slices from the point cloud data of the rotated segment, and project the point cloud data of the key components to the xoz plane to obtain the two-dimensional point cloud data of the segment;

[0143] S722: Use the RANSAC algorithm to perform straight line fitting on the two-dimensional point cloud data of the segment, obtain the equation of each straight line, and calculate the coordinates of the intersection points of each straight line equation. These intersection points are the intersection points between the straight lines and represent the key size features of the segment; for data that need to fit circular features, use the least squares method to estimate the coordinates of the center of the circle, and finally use the distance formula to calculate the distance between the key intersection points. These distances represent the overall size of the segment.

[0144] Further, in a specific embodiment of the present invention, in step S721, the thickness of each slice is set to 0.02m, the segment straight line fitting key component is obtained by using the normal vector segmentation algorithm, the segment circle fitting key component is obtained by using the through filtering algorithm, and the point cloud is projected to the xoz plane, and each point is labeled, such as Figure 7 shown; for Figure 5 For each dimension in the example, dimension L1 can be obtained by detecting the length and width of the rectangular plate where L1 is located. Dimensions L2, L3, L4, L5, H1, B1, and B2 can be obtained by finding the coordinates of key points. L2 is the distance between points p6 and p9, and L3 is the distance between points p3 and p 10 The distance between points p5 and p8 is L4, L5 is the distance between points p4 and p8 minus the radius of the steel pipe, B1 is the distance between points p3 and p6, and B2 is the distance between points p9 and p 10 H1 is the distance between points p1 and p3 plus the height of the side panel.

[0145] Furthermore, in a specific embodiment of the present invention, in step S722, the RANSAC algorithm is used to perform straight line fitting, and the minimum distance from a point to a straight line is set to sigma = 0.005, the number of randomly sampled points n = 3, and the maximum number of iterations = 1000. The obtained point cloud data of each straight line are saved in a list, and the line segments are marked. Each line segment is then iteratively fitted 20 times, and 80% of the points are randomly sampled in each iteration. After each straight line equation is obtained, it is classified according to the characteristics of the straight line equation, and the intersection points are calculated. Finally, the least squares method is used to fit the center of the two-dimensional circle to obtain the center coordinates, and the distance of each intersection point is calculated to obtain the overall size of the segment.

[0146] Furthermore, the step S8 of obtaining the target value of the overall size of the segment of the target steel tower component according to the size of the target steel tower component processing drawing and the processing deformation of the target steel tower component includes:

[0147] S81: establishing an accurate geometric model of the target steel tower component according to the processing drawing of the target steel tower component, meshing the geometric model using finite element analysis software, and constructing a finite element model of the target steel tower component;

[0148] S82: Determine the boundary constraint condition of the target steel tower component in the finite element model according to the placement state of the target steel tower component at the processing site;

[0149] S83: Calculate the influence of temperature change on material properties according to the temperature difference between the ambient temperature recorded during the on-site scanning and the temperature during processing, and apply a corresponding temperature load in the finite element model; apply the dead weight of the steel tower component in the finite element model;

[0150] S84: Run the finite element analysis software to calculate the deformation of the target steel tower components under the effects of temperature and deadweight;

[0151] S85: Add the deformation amount to the size on the target steel tower component processing drawing to obtain the target value of the overall size of the segment of the target steel tower component taking into account the deformation.

[0152] Furthermore, step S9 obtains an error value of the overall size of the segment based on the overall size of the segment and the target value of the overall size of the segment; including: subtracting the target value of the overall size of the segment in S8 from the measured value of the overall size of the segment in S7 as the error value of the overall size of the segment given by the detection; finally, based on the comparison of the error value with the accuracy value required by the manufacturing, the segment size that does not meet the manufacturing accuracy requirement is marked and prompted.

[0153] In the implementation example of the present invention, the intelligent detection value, target value and absolute error of the overall size of the segment are shown in Table 3.

[0154] Table 3

[0155] Test items Measurement average value / mm Target value / mm Absolute error / mm <![CDATA[L1]]> 10794.0 10801.3 7.3 <![CDATA[L2-L3]]> 5.9 0 5.9 <![CDATA[L4-L5]]> 6.4 0 6.4 <![CDATA[B1]]> 13329.7 13330.7 1.0 <![CDATA[B2]]> 16130.6 16133.9 3.3 <![CDATA[H1]]> 5993.5 5998.9 5.4

[0156] From the above, it can be seen that the bridge steel tower component size detection method based on three-dimensional laser scanning provided by the present invention can complete the intelligent detection of the unit size of the bridge steel tower component, and can use a computer to calculate the steel pipe unit diameter, the rib spacing of the ribbed steel plate, and the overall size of the segment. The method of the present invention is practical and effective.

[0157] like Figure 8 As shown, the second aspect of the present invention provides a bridge steel tower component size detection system based on three-dimensional laser scanning, which is used to implement the above detection method, including a first main module, a second main module, a third main module, a fourth main module, a fifth main module, a sixth main module, a seventh main module, an eighth main module, and a ninth main module; wherein,

[0158] The first main module is used to scan a specific area including a target steel tower component using a 3D laser scanner to obtain overall point cloud data including the target steel tower component;

[0159] The second main module is used to process the background noise in the overall point cloud data using a straight-through filtering algorithm to obtain effective point cloud data of the target steel tower component;

[0160] The third main module is used to perform clustering processing on the effective point cloud data using the DBSCAN algorithm to obtain all unit point cloud data of the target steel tower component; all unit point cloud data of the target steel tower component include steel pipe point cloud data, ribbed steel plate point cloud data and segment point cloud data;

[0161] The fourth main module is used to use the RANSAC algorithm to perform steel pipe axis fitting and noise removal on the steel pipe point cloud data to obtain the steel pipe outer wall point cloud data; and obtain the steel pipe diameter according to the steel pipe outer wall point cloud data;

[0162] The fifth main module is used to process the point cloud data of the ribbed steel plate by using a normal vector-based segmentation method and a DBSCAN algorithm to obtain point cloud data of instantiated transverse plate units and longitudinal plate units; and to process the point cloud data of the instantiated transverse plate units and longitudinal plate units by using a voxel sampling algorithm and a RANSAC algorithm to obtain the rib spacing;

[0163] The sixth main module is used to perform multi-plane fitting extraction of the segment point cloud data using the RANSAC algorithm to obtain the segment side panel point cloud data; the cross product of the normal vector of the side panel and the Z axis is used as the front direction of the segment, and the original segment is rotated so that the front direction is parallel to the Y axis to obtain the rotated segment point cloud data;

[0164] The seventh main module is used to obtain the length and width of the segment side panel according to the segment side panel point cloud data; and obtain the overall size of the segment according to the rotated segment point cloud data;

[0165] The eighth main module is used to obtain the target value of the segment overall size of the target steel tower component according to the size of the target steel tower component processing drawing and the processing deformation amount of the target steel tower component;

[0166] The ninth main module is used to obtain the error value of the overall size of the segment according to the overall size of the segment and the target value of the overall size of the segment; and to mark and prompt the segment size that does not meet the manufacturing accuracy requirements by comparing the error value with the accuracy value required by manufacturing.

[0167] It should be noted that the bridge steel tower component size detection system based on three-dimensional laser scanning provided in this embodiment can be a computer program (including program code) running in a computer device. For example, the bridge steel tower component size detection system based on three-dimensional laser scanning is an application software; the bridge steel tower component size detection system based on three-dimensional laser scanning can be used to execute the corresponding steps in the above method provided in the embodiment of the present application.

[0168] In some feasible implementations, the bridge steel tower component size detection system based on three-dimensional laser scanning provided in this embodiment can be implemented by a combination of software and hardware. As an example, the bridge steel tower component size detection system based on three-dimensional laser scanning provided in this embodiment of the application can be a processor in the form of a hardware decoding processor, which is programmed to execute the bridge steel tower component size detection method based on three-dimensional laser scanning provided in the embodiment of the application. For example, the processor in the form of a hardware decoding processor can adopt one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field programmable gate arrays (FPGAs) or other electronic components.

[0169] In some feasible implementations, the bridge steel tower component size detection system based on three-dimensional laser scanning provided in this embodiment can be implemented in software form, which can be software in the form of programs and plug-ins, and includes a series of modules to implement the bridge steel tower component size detection method based on three-dimensional laser scanning provided in an embodiment of the present invention.

[0170] The bridge steel tower component size detection system based on three-dimensional laser scanning provided in this embodiment realizes a high-precision, automated, and robust non-contact detection process for the size of bridge steel tower components by combining multiple advanced technologies such as three-dimensional laser scanning technology, RANSAC algorithm, pass-through filtering algorithm, DBSCAN algorithm, Rodrigues formula, and least squares method. It not only improves the accuracy and efficiency of measurement, but also enhances the reliability of data and the flexibility of processing, providing strong technical support for quality control, structural health monitoring, maintenance, and repair of bridge projects, while reducing costs and improving the economic benefits of projects.

[0171] A third aspect of the present invention further provides an electronic device, Fig. 9 Schematic diagram of the structure of the electronic device of this embodiment. Fig. 9As shown, the electronic device 1000 in this embodiment may include: a processor 1001, a network interface 1004 and a memory 1005. In addition, the above-mentioned electronic device 1000 may also include: a user interface 1003, and at least one communication bus 1002. Among them, the communication bus 1002 is used to realize the connection and communication between these components. Among them, the user interface 1003 may include a display screen (Display), a keyboard (Keyboard), and the user interface 1003 may optionally include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 1004 may be a high-speed RAM memory, or it may be a non-volatile memory (non-volatile memory), such as at least one disk storage. The memory 1005 may optionally also be at least one storage device located away from the aforementioned processor 1001. As Fig. 9 As shown, the memory 1005 as a computer-readable storage medium may include an operating system, a network communication module, a user interface module, and a device control application program.

[0172] like Fig. 9 In the electronic device 1000 shown, the network interface 1004 can provide a network communication function; the user interface 1003 is mainly used to provide an input interface for the user; and the processor 1001 can be used to call the device control application stored in the memory 1005 to achieve:

[0173] A 3D laser scanner is used to scan a specific area including the target steel tower component to obtain overall point cloud data including the target steel tower component;

[0174] Using a straight-through filtering algorithm to process background noise in the overall point cloud data, and obtaining effective point cloud data of the target steel tower component;

[0175] The effective point cloud data are clustered using a DBSCAN algorithm to obtain all unit point cloud data of the target steel tower component; all unit point cloud data of the target steel tower component include steel pipe point cloud data, ribbed steel plate point cloud data and segment point cloud data;

[0176] The RANSAC algorithm is used to perform steel pipe axis fitting and noise removal on the steel pipe point cloud data to obtain the steel pipe outer wall point cloud data; the steel pipe diameter is obtained according to the steel pipe outer wall point cloud data;

[0177] The point cloud data of the ribbed steel plate is processed by a segmentation method based on a normal vector and a DBSCAN algorithm to obtain point cloud data of instantiated transverse plate units and longitudinal plate units; the point cloud data of the instantiated transverse plate units and longitudinal plate units are processed by a voxel sampling algorithm and a RANSAC algorithm to obtain the rib spacing;

[0178] The segment point cloud data is subjected to multi-plane fitting extraction by using the RANSAC algorithm to obtain the segment side panel point cloud data; the cross product of the normal vector of the side panel and the Z axis is used as the front direction of the segment, and the original segment is rotated so that the front direction is parallel to the Y axis to obtain the rotated segment point cloud data;

[0179] Obtaining the length and width of the segment side panel according to the segment side panel point cloud data; obtaining the overall size of the segment according to the rotated segment point cloud data;

[0180] Obtaining a target value of the overall size of the segment of the target steel tower component according to the size of the target steel tower component processing drawing and the processing deformation of the target steel tower component;

[0181] The error value of the segment's overall size is obtained according to the segment's overall size and the target value of the segment's overall size; and the segment size that does not meet the manufacturing accuracy requirement is marked and prompted according to the comparison between the error value and the accuracy value required for manufacturing.

[0182] It should be understood that in some feasible implementations, the processor 1001 may be a central processing unit (CPU), or other general-purpose processors, DSPs, ASICs, FPGAs or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc. The memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A portion of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.

[0183] In a specific implementation, the electronic device 1000 can execute the above-mentioned functions through its built-in functional modules. Fig.10 For the implementation methods provided in each step, please refer to the implementation methods provided in the above steps for details, which will not be repeated here.

[0184] The electronic device provided in this embodiment realizes a high-precision, automated, and robust non-contact detection process for the size of bridge steel tower components by combining multiple advanced technologies such as three-dimensional laser scanning technology, RANSAC algorithm, pass-through filtering algorithm, DBSCAN algorithm, Rodrigues formula, and least squares method. It not only improves the accuracy and efficiency of measurement, but also enhances the reliability of data and the flexibility of processing, providing strong technical support for quality control, structural health monitoring, maintenance, and repair of bridge projects, while reducing costs and improving the economic benefits of projects.

[0185] The present application also provides a computer-readable storage medium, which stores a computer program and is executed by a processor to implement Fig.10 For the methods provided in each step, please refer to the implementation methods provided in the above steps for details, which will not be repeated here.

[0186] The computer-readable storage medium provided in this embodiment realizes a high-precision, automated, and robust non-contact detection process for the size of bridge steel tower components by combining multiple advanced technologies such as three-dimensional laser scanning technology, RANSAC algorithm, pass-through filtering algorithm, DBSCAN algorithm, Rodrigues formula, and least squares method. It not only improves the accuracy and efficiency of measurement, but also enhances the reliability of data and the flexibility of processing, providing strong technical support for quality control, structural health monitoring, maintenance, and repair of bridge projects, while reducing costs and improving the economic benefits of projects.

[0187] Any reference to memory, storage, database or other medium used in the embodiments provided in the present application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0188] It will be easily understood by those skilled in the art that 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 and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for detecting the size of bridge steel tower components based on three-dimensional laser scanning, characterized in that: The steps include: S1: Use a 3D laser scanner to scan a specific area containing the target steel tower components to obtain overall point cloud data containing the target steel tower components; S2: using a straight-through filtering algorithm to process background noise in the overall point cloud data to obtain valid point cloud data of the target steel tower component; S3: clustering the effective point cloud data using the DBSCAN algorithm to obtain all unit point cloud data of the target steel tower component; all unit point cloud data of the target steel tower component include steel pipe point cloud data, ribbed steel plate point cloud data and segment point cloud data; S4: using the RANSAC algorithm to perform steel pipe axis fitting and noise removal on the steel pipe point cloud data to obtain steel pipe outer wall point cloud data; and obtaining the steel pipe diameter according to the steel pipe outer wall point cloud data; S5: Processing the point cloud data of the ribbed steel plate by a normal vector-based segmentation method and a DBSCAN algorithm to obtain point cloud data of instantiated transverse plate units and longitudinal plate units; processing the point cloud data of the instantiated transverse plate units and longitudinal plate units by a voxelized sampling algorithm and a RANSAC algorithm to obtain the rib spacing; S6: Use the RANSAC algorithm to perform multi-plane fitting extraction on the segment point cloud data to obtain the segment side panel point cloud data; use the cross product of the normal vector of the side panel and the Z axis as the front direction of the segment, rotate the original segment so that the front direction is parallel to the Y axis, and obtain the rotated segment point cloud data; S7: obtaining the length and width of the segment side panel according to the segment side panel point cloud data; obtaining the overall size of the segment according to the rotated segment point cloud data; S8: Obtaining a target value of the overall size of the segment of the target steel tower component according to the size of the target steel tower component processing drawing and the processing deformation of the target steel tower component; S9: obtaining an error value of the overall size of the segment according to the overall size of the segment and a target value of the overall size of the segment; and marking and prompting the segment size that does not meet the manufacturing accuracy requirement by comparing the error value with the accuracy value required by manufacturing.

2. The method for detecting the size of a bridge steel tower component based on three-dimensional laser scanning according to claim 1, characterized in that: The step S2 uses a straight-through filtering algorithm to process the background noise in the overall point cloud data containing the target steel tower component to obtain valid point cloud data of the target steel tower component; including: Construct a bounding box larger than the target steel tower component, and confine the target steel tower component to a range in the three dimensions of x, y, and z; The point cloud direct filtering algorithm is used to remove the noise data outside the bounding box, retaining the target steel tower components and some surrounding noise.

3. The method for detecting the size of a bridge steel tower component based on three-dimensional laser scanning according to claim 1 is characterized in that: The step S4 uses the RANSAC algorithm to perform steel pipe axis fitting and noise removal on the steel pipe point cloud data in all unit point cloud data of the target steel tower component to obtain steel pipe outer wall point cloud data; including: S411: Dividing the steel pipe point cloud data into a sample point set and a test point set; S412: Using the RANSAC algorithm to perform steel pipe axis fitting on the sample point set data to obtain a steel pipe fitting axis equation; S413: Calculate the distance between each point in the test point set data and the steel pipe fitting axis according to the steel pipe axis equation, set a distance threshold, remove the noise points on the inner wall of the steel pipe according to the distance threshold, and obtain the point cloud data of the outer wall of the steel pipe.

4. The method for detecting the size of a bridge steel tower component based on three-dimensional laser scanning according to claim 3 is characterized in that: The step S4 of obtaining the diameter of the steel pipe according to the point cloud data of the outer wall of the steel pipe comprises: S421: Rotate the steel pipe axis to be parallel to the z-axis; calculate the rotation matrix of the steel pipe axis using Rodriguez formula; S422: using the rotation matrix to transform the coordinates of each point in the steel pipe outer wall point cloud data from the coordinate system before rotation to the coordinate system after rotation; S423: Projecting the rotated steel pipe axis onto the xoy plane to obtain two-dimensional circle data; S424: Use the least squares method to perform circle fitting on the two-dimensional circle data to obtain the center coordinates and radius of the best fitting circle, and then obtain the diameter of the steel pipe.

5. A method for detecting the size of a bridge steel tower component based on three-dimensional laser scanning according to any one of claims 1 to 4, characterized in that: Step S5, in which the point cloud data of the ribbed steel plate is processed by a segmentation method based on a normal vector and a DBSCAN algorithm to obtain point cloud data of instantiated transverse plate units and longitudinal plate units, comprises: Distinguishing and extracting the point cloud data of the transverse plate and the longitudinal plate from the point cloud data of the ribbed steel plate by a segmentation method based on a normal vector; The DBSCAN algorithm is used to instantiate and segment the point cloud data of the transverse plate and the longitudinal plate to obtain the point cloud data of the instantiated transverse plate unit and the longitudinal plate unit; The step S5 uses a voxel sampling algorithm and a RANSAC algorithm to perform plane fitting on the point cloud data of the instantiated transverse plate unit and the longitudinal plate unit to obtain the rib spacing; including: The plate units are processed by voxel sampling algorithm to obtain uniformly distributed plate units. The center coordinates of each plate unit are calculated, and the common normal vector of the plate unit is obtained after plane fitting by RANSAC algorithm. The distance between the center coordinates of each transverse plate and longitudinal plate along the common normal vector is taken as the rib spacing.

6. A method for detecting the size of a bridge steel tower component based on three-dimensional laser scanning according to any one of claims 1 to 4, characterized in that: Step S6 uses the RANSAC algorithm to perform segment multi-plane fitting extraction on the segment point cloud data to obtain segment side panel data; including: S611: Use the RANSAC algorithm to perform multi-plane fitting on the segment point cloud data. The RANSAC algorithm preferentially fits the plane containing the most point clouds. Each time a plane is fitted, the point cloud data of the plane is stored in a list, and the fitted plane data is deleted from the remaining segment point cloud data. S612: Repeat step S61 to continue plane fitting until the number of planes in the list reaches a preset threshold, and then exit the fitting process, i.e., obtain the key plane panels (such as side panels, bottom panels, etc.) of each segment; S613: Filter out side panels from the key plane panels of each segment.

7. A method for detecting the size of a bridge steel tower component based on three-dimensional laser scanning according to any one of claims 1 to 4, characterized in that: Step S7, obtaining the length and width of the segment side panel according to the segment side panel point cloud data, comprises: S711: using a statistical filtering algorithm to remove noise point clouds near the edge of the rectangular plate in the segment side panel point cloud data, rotating the side panel so that the normal vector of the reference plane is parallel to the Z axis and projecting it to obtain two-dimensional side panel data; S712: Calculate the length and width of the side panel according to the two-dimensional side panel data; Step S712 calculates the length and width of the side panel according to the two-dimensional side panel data; including: An edge extraction algorithm is used to identify edge points in the two-dimensional side panel data; a RANSAC algorithm is used to perform straight line fitting on the edge points to obtain an edge line equation; an intersection of the fitted edge line is calculated according to the edge line equation to determine the diagonal point of the rectangular plate; a distance formula is used to calculate the distance from the intersection to the edge of the rectangular plate to obtain the length and width of the side panel.

8. A method for detecting the size of a bridge steel tower component based on three-dimensional laser scanning according to any one of claims 1 to 4, characterized in that: The step S7 of obtaining the overall size of the segment according to the rotated segment point cloud data includes: S721: Slice the rotated segment point cloud data along the Y-axis direction, take partial slices at the front and rear segments, extract key structure point cloud data for segment slice size detection using a normal vector segmentation algorithm and a straight-through filtering algorithm, and project the key structure point cloud data onto the xoz plane to obtain segment two-dimensional point cloud data; S722: Use the RANSAC algorithm to perform straight line fitting on the segment two-dimensional point cloud data, obtain the equation of each straight line, and calculate the coordinates of the intersection of each straight line equation; for data that needs to fit circular features, use the least squares method to estimate the coordinates of the center of the circle, and finally use the distance formula to calculate the distance between the key intersection points to obtain the overall size of the segment.

9. A method for detecting the size of a bridge steel tower component based on three-dimensional laser scanning according to any one of claims 1 to 4, characterized in that: The step S8 of obtaining the target value of the overall size of the segment of the target steel tower component according to the size of the target steel tower component processing drawing and the processing deformation of the target steel tower component comprises: S81: establishing an accurate geometric model of the target steel tower component according to the processing drawing of the target steel tower component, meshing the geometric model using finite element analysis software, and constructing a finite element model of the target steel tower component; S82: Determine the boundary constraint condition of the target steel tower component in the finite element model according to the placement state of the target steel tower component at the processing site; S83: Calculate the influence of temperature change on material properties according to the temperature difference between the ambient temperature recorded during the on-site scanning and the temperature during processing, and apply a corresponding temperature load in the finite element model; apply the dead weight of the steel tower component in the finite element model; S84: Run finite element analysis software to calculate the deformation of the target steel tower components under the effects of temperature and deadweight; S85: Add the deformation amount to the size on the target steel tower component processing drawing to obtain the target value of the overall size of the segment of the target steel tower component taking into account the deformation.

10. A bridge steel tower component size detection system based on three-dimensional laser scanning, characterized in that: A method for detecting the size of a bridge steel tower component based on three-dimensional laser scanning as described in any one of claims 1 to 9, comprising: The first main module is used to scan a specific area including a target steel tower component using a three-dimensional laser scanner to obtain overall point cloud data including the target steel tower component; The second main module is used to process the background noise in the overall point cloud data by using a straight-through filtering algorithm to obtain effective point cloud data of the target steel tower component; The third main module is used to perform clustering processing on the effective point cloud data using the DBSCAN algorithm to obtain all unit point cloud data of the target steel tower component; all unit point cloud data of the target steel tower component include steel pipe point cloud data, ribbed steel plate point cloud data and segment point cloud data; The fourth main module is used to use the RANSAC algorithm to perform steel pipe axis fitting and noise removal on the steel pipe point cloud data to obtain the steel pipe outer wall point cloud data; and obtain the steel pipe diameter according to the steel pipe outer wall point cloud data; The fifth main module is used to process the point cloud data of the ribbed steel plate by using a normal vector-based segmentation method and a DBSCAN algorithm to obtain point cloud data of instantiated transverse plate units and longitudinal plate units; and to process the point cloud data of the instantiated transverse plate units and longitudinal plate units by using a voxel sampling algorithm and a RANSAC algorithm to obtain the rib spacing; The sixth main module is used to perform multi-plane fitting extraction of the segment point cloud data using the RANSAC algorithm to obtain the segment side panel point cloud data; the cross product of the normal vector of the side panel and the Z axis is used as the front direction of the segment, and the original segment is rotated so that the front direction is parallel to the Y axis to obtain the rotated segment point cloud data; The seventh main module is used to obtain the length and width of the segment side panel according to the segment side panel point cloud data; and obtain the overall size of the segment according to the rotated segment point cloud data; The eighth main module is used to obtain the target value of the overall size of the segment of the target steel tower component according to the size of the target steel tower component processing drawing and the processing deformation amount of the target steel tower component; The ninth main module is used to obtain the error value of the overall size of the segment according to the overall size of the segment and the target value of the overall size of the segment; and to mark and prompt the segment size that does not meet the manufacturing accuracy requirements by comparing the error value with the accuracy value required by manufacturing.

Images