Steel arch bridge arch rib linear extraction method and system based on improved DBSCAN algorithm
The improved DBSCAN algorithm with RANSAC processing enhances the accuracy of steel arch bridge rib alignment extraction by integrating preprocessing techniques and clustering, addressing the limitations of existing methods in handling complex bridge geometries.
Patent Information
- Application Number
- CN202411211576.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-30
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2044-08-30
AI Technical Summary
Existing methods for monitoring the linear shape of steel arch bridges, such as total station measurement, photogrammetry, and laser scanning, struggle with complex structures and fail to achieve accurate extraction of arch rib alignment due to their limitations in handling diverse and intricate steel arch bridge geometries.
An improved DBSCAN algorithm integrated with RANSAC is used to process point cloud data, involving preprocessing steps like downsampling with an Octree algorithm and denoising with bilateral filtering, followed by PCA for normal direction estimation, and subsequent line fitting and clustering to extract the arch rib shape accurately.
The method achieves millimeter-level accuracy in extracting the arch rib shape of steel arch bridges, reducing manual labor and time consumption while effectively handling complex structures and interferences, thereby improving extraction precision.
Smart Images

Figure CN119089554A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the field of steel arch bridge monitoring, and in particular to a method and a system for extracting the linear shape of an arch rib of a steel arch bridge based on an improved DBSCAN algorithm. Background Art
[0002] As a structural form, steel arch bridges have a long history of development and wide application in bridge construction. The earliest steel arch bridges appeared in the mid-19th century. These bridges quickly gained widespread use due to their superior mechanical properties and long service life. In the mid-20th century, steel arch bridge technology gradually matured, especially in long-span bridges and complex terrain conditions. A large number of modern steel arch bridges were built around the world. In the 21st century, with the further advancement of material science and construction technology, steel arch bridges have achieved more optimized and efficient design, construction and maintenance.
[0003] The commonly used method for linear monitoring of steel arch bridges is to use total station measurement combined with GPS monitoring to perform high-precision three-dimensional coordinate measurement of the key nodes of the bridge, thereby monitoring the linear changes of the bridge. The advantages of total station measurement are high accuracy and wide applicability, which are suitable for long-term monitoring of bridges, but they are highly dependent on manual labor, have high sight requirements, and have limited coverage. In addition, there are two other methods: photogrammetry and laser scanning. Among them, photogrammetry uses high-resolution cameras to take photos of the bridge from different angles and reconstruct the three-dimensional model of the bridge to monitor the linear changes of the bridge. It is suitable for short-term monitoring and monitoring in complex environments, but the accuracy is limited by the photographic equipment and processing technology, and the data processing is complex. Laser scanning technology can quickly obtain high-density point cloud data on the surface of the bridge, and monitor the linear changes of the bridge by analyzing the point cloud data. It is suitable for large-scale and high-precision linear monitoring.
[0004] Although some scholars have successfully extracted bridge alignment from point cloud data, the above existing methods have certain limitations in dealing with the complex structure and diversity of steel arch bridges with more complex structures, and cannot achieve accurate extraction. Therefore, in order to solve the above problems, a steel arch bridge arch rib alignment extraction method and system based on an improved DBSCAN algorithm is needed, which can complete the steel arch bridge alignment extraction more accurately and effectively, and significantly improve the alignment extraction accuracy. Summary of the invention
[0005] In view of this, the purpose of the present invention is to overcome the defects in the prior art and provide a method and system for extracting the linear shape of steel arch bridge ribs based on an improved DBSCAN algorithm, which can more accurately and effectively complete the linear shape extraction of steel arch bridges and significantly improve the linear shape extraction accuracy.
[0006] The method for extracting the linear shape of the arch rib of a steel arch bridge based on the improved DBSCAN algorithm of the present invention comprises:
[0007] Collect point cloud data of steel arch bridge to obtain original point cloud data;
[0008] Preprocessing the original point cloud data to obtain preprocessed point cloud data;
[0009] Perform preliminary extraction on the preprocessed point cloud data to obtain preliminary extraction results;
[0010] The improved DBSCAN algorithm is used to perform secondary extraction on the preliminary extraction results to obtain the final extraction results.
[0011] Further, the preprocessing includes downsampling and denoising;
[0012] The downsampling specifically includes: downsampling the point cloud data using an Octree algorithm, taking the centroid of each Octant as a representative point, and constructing a new downsampled point cloud;
[0013] The denoising specifically includes: denoising the downsampled point cloud data using bilateral filtering.
[0014] Furthermore, the preprocessed point cloud data is initially extracted, including:
[0015] Use the PCA algorithm to estimate the point cloud normal, select the eigenvector corresponding to the maximum eigenvalue as the normal direction, and obtain the normal direction of each point in the point cloud data set;
[0016] Set the neighborhood point n=n0 and make the normal vector direction consistent;
[0017] According to the analysis results of the point cloud data, thresholds are set in the Z direction and the X direction to extract the point clouds at the bottom and both sides of the arch rib respectively.
[0018] Furthermore, the improved DBSCAN algorithm is used to perform secondary extraction on the preliminary extraction results, including:
[0019] a. Fit and segment the first line: Use the RANSAC algorithm to fit a line in the current point cloud, find the inner points of the line, extract these inner points and remove them from the original point cloud;
[0020] b. Continue fitting the remaining point clouds: Apply the RANSAC algorithm again on the remaining point clouds to fit the next straight line;
[0021] c. Repeat the above steps until the number of points in the remaining point cloud is less than min_num;
[0022] d. Determine the line to which the point belongs: In each RANSAC algorithm fitting, the distance from each point to the fitted line is calculated to determine whether it belongs to the current line. Only the inner points will be segmented and removed from the original point cloud;
[0023] e. Represent each straight line as a set of feature vectors;
[0024] f. Use the DBSCAN algorithm to cluster the feature vectors of these lines. In a given ∈-neighborhood, if the number of neighbors of a line is greater than or equal to MinPts, then this line is considered to be a core line. The core line and the lines that can be reached by it form a cluster.
[0025] g. Continue clustering the remaining lines until all lines are classified;
[0026] h. Filter the point clouds of the bridge part according to the classification results, delete all kinds of interference point clouds, and only keep the point clouds of the arch ribs;
[0027] i. Extract and fit the key point coordinates of the filtered point cloud.
[0028] A steel arch bridge arch rib linear extraction system based on an improved DBSCAN algorithm, comprising a collection unit, a preprocessing unit, a preliminary extraction unit and a secondary extraction unit;
[0029] The acquisition unit is used to collect point cloud data of the steel arch bridge to obtain original point cloud data;
[0030] The preprocessing unit is used to preprocess the original point cloud data to obtain preprocessed point cloud data;
[0031] The preliminary extraction unit is used to perform preliminary extraction on the preprocessed point cloud data to obtain a preliminary extraction result;
[0032] The secondary extraction unit is used to perform secondary extraction on the preliminary extraction result by adopting the improved DBSCAN algorithm to obtain the final extraction result.
[0033] Further, the preprocessing includes downsampling and denoising;
[0034] The downsampling specifically includes: downsampling the point cloud data using an Octree algorithm, taking the centroid of each Octant as a representative point, and constructing a new downsampled point cloud;
[0035] The denoising specifically includes: denoising the downsampled point cloud data using bilateral filtering.
[0036] Furthermore, the preprocessed point cloud data is initially extracted, including:
[0037] Use the PCA algorithm to estimate the point cloud normal, select the eigenvector corresponding to the maximum eigenvalue as the normal direction, and obtain the normal direction of each point in the point cloud data set;
[0038] Set the neighborhood point n=n0 and make the normal vector direction consistent;
[0039] According to the analysis results of the point cloud data, thresholds are set in the Z direction and the X direction to extract the point clouds at the bottom and both sides of the arch rib respectively.
[0040] Furthermore, the improved DBSCAN algorithm is used to perform secondary extraction on the preliminary extraction results, including:
[0041] a. Fit and segment the first line: Use the RANSAC algorithm to fit a line in the current point cloud, find the inner points of the line, extract these inner points and remove them from the original point cloud;
[0042] b. Continue fitting the remaining point clouds: Apply the RANSAC algorithm again on the remaining point clouds to fit the next straight line;
[0043] c. Repeat the above steps until the number of points in the remaining point cloud is less than min_num;
[0044] d. Determine the line to which the point belongs: In each RANSAC algorithm fitting, the distance from each point to the fitted line is calculated to determine whether it belongs to the current line. Only the inner points will be segmented and removed from the original point cloud;
[0045] e. Represent each straight line as a set of feature vectors;
[0046] f. Use the DBSCAN algorithm to cluster the feature vectors of these lines. In a given ∈-neighborhood, if the number of neighbors of a line is greater than or equal to MinPts, then this line is considered to be a core line. The core line and the lines that can be reached by it form a cluster.
[0047] g. Continue clustering the remaining lines until all lines are classified;
[0048] h. Filter the point clouds of the bridge part according to the classification results, delete all kinds of interference point clouds, and only keep the point clouds of the arch ribs;
[0049] i. Extract and fit the key point coordinates of the filtered point cloud.
[0050] The beneficial effects of the present invention are as follows: the present invention discloses a method and system for extracting the linear shape of the arch ribs of a steel arch bridge based on an improved DBSCAN algorithm, which takes a large-span steel arch bridge with a complex structure as the analysis object, and uses a three-dimensional laser scanning technology to scan the steel arch bridge in view of the problems of complex steel arch structures and numerous interferences. The structural characteristics of the steel arch bridge are utilized, and a PCA algorithm is used to perform preliminary extraction of the key structural point cloud of the steel arch bridge. For structures such as the bridge deck, cross brace, cross-connection and web member in the preliminary extraction results that interfere with the linear extraction results, the improved DBSCAN algorithm integrated with RANSAC is used to achieve accurate extraction of the arch ribs of the steel arch bridge. BRIEF DESCRIPTION OF THE DRAWINGS
[0051] The present invention will be further described below in conjunction with the accompanying drawings and embodiments:
[0052] Figure 1 This is a schematic diagram of the original point cloud data of the steel arch bridge of the present invention;
[0053] Figure 2 This is a schematic diagram of the original point cloud data of a steel arch bridge after preprocessing of the present invention;
[0054] Figure 3 This is a schematic diagram of the point cloud of the key part of the arch rib after preliminary extraction of the present invention;
[0055] Figure 4a It is a schematic diagram of the point cloud screening results on both sides of the arch rib of the present invention;
[0056] Figure 4b It is a schematic diagram of the point cloud screening result of the arch rib bottom of the present invention;
[0057] Figure 5a This is a schematic diagram of the error distribution of the point cloud extracted from the bottom of the arch rib of the present invention;
[0058] Figure 5b This is a schematic diagram of the error distribution of the point cloud extracted on both sides of the arch rib of the present invention;
[0059] Figure 6 It is a schematic diagram of the process of the linear extraction method of the steel arch bridge arch rib of the present invention;
[0060] Figure 7a This is a schematic diagram of the clustering effect of the DBSCAN algorithm without any improvement;
[0061] Figure 7b It is a schematic diagram of the clustering effect of the improved DBSCAN algorithm integrated with RANSAC of the present invention. DETAILED DESCRIPTION
[0062] The present invention is further described below in conjunction with the accompanying drawings, as shown in the drawings:
[0063] This embodiment discloses a method for extracting the linear shape of the arch rib of a steel arch bridge based on an improved DBSCAN algorithm, comprising the following steps:
[0064] Collect point cloud data of steel arch bridge to obtain original point cloud data;
[0065] Preprocessing the original point cloud data to obtain preprocessed point cloud data;
[0066] Perform preliminary extraction on the preprocessed point cloud data to obtain preliminary extraction results;
[0067] The improved DBSCAN algorithm is used to perform secondary extraction on the preliminary extraction results to obtain the final extraction results.
[0068] In this embodiment, a RiGel station-type scanner is used to perform 3D laser scanning on a steel arch bridge under construction. During the scanning, two stations are set up on both sides of the bridge and three stations are set up on the bridge. The original point cloud data totals 178,946,265 points. Figure 1 Among them, the three-dimensional laser point cloud data has the advantages of comprehensiveness and detail, and can fully present the shape and deformation information (displacement, deformation, settlement and other parameters) of the bridge structure.
[0069] Since the amount of original point cloud data is too large and there are a small number of noise points, the original point cloud data is first downsampled and denoised. For downsampling: the Octree algorithm is used to downsample the point cloud. The Octree algorithm is an algorithm commonly used in point cloud data processing and analysis. It is mainly used to reduce the density and complexity of point cloud data while maintaining the main features and structure of point cloud data. The Octree algorithm effectively segments and organizes point cloud data by recursively dividing the three-dimensional space into eight sub-cubes (Octants). Since Octree is based on spatial division, it is able to retain the main structural features of point cloud data. This is because the center of gravity of each Octant is calculated from all points in the area, which can better retain the distribution and structural characteristics of the original data. The algorithm steps are as follows:
[0070] Octree is constructed to recursively divide the entire point cloud data space into eight sub-cubes until the number of point clouds in each sub-cube is less than a predetermined threshold or the maximum depth is reached.
[0071]
[0072] Where: P sub is a point within the current Octant.
[0073] For each non-empty Octant, calculate the centroid of all points inside it as the representative point.
[0074]
[0075] Where: p i ——points in Octant; n——the number of points in Octant;
[0076] Down-sampling is performed, and the centroid of each Octant is used as the representative point to construct a new down-sampled point cloud.
[0077] For denoising: Considering the structural characteristics of the bridge, bilateral filtering, which is more adaptable to the complex structural details of the bridge, is used to denoise the downsampled point cloud data, such as Figure 2 The number of raw point clouds after preprocessing is shown. Bilateral filtering is a nonlinear filtering technology in point cloud processing. It can process nonlinearly changing point cloud data and has better robustness to nonlinear noise. Since it considers the distance in both the spatial domain and the intensity domain, the edges and details of the filtered point cloud are better preserved.
[0078] In this embodiment, preliminary extraction is performed on the preprocessed point cloud data, specifically including:
[0079] Use the PCA algorithm to estimate the point cloud normal and make the normal vector consistent in direction:
[0080] The PCA algorithm regards the local point cloud of each point as a smooth surface, and then uses the least squares algorithm to estimate an optimal tangent plane in the neighborhood. Finally, the normal vector of this optimal tangent plane is set as the normal vector of the sampling point. The algorithm steps are as follows:
[0081] For a point cloud dataset containing n 3D points, each point consists of three coordinate values (x i ,y i ,z i ), then these points can be represented as a 3×n matrix X, whose column vectors are all the points in the point cloud. First, calculate the mean vector of the point cloud dataset
[0082]
[0083] Where: X i ——The i-th point in the point cloud.
[0084] Decentralize the point cloud dataset:
[0085]
[0086] Calculate the covariance matrix C:
[0087]
[0088] The covariance matrix C is a symmetric matrix whose eigenvector is the normal direction. For each point in the point cloud, the PCA algorithm calculates the normal by solving the eigenvalues and eigenvectors of the covariance matrix. The eigenvalue corresponding to the eigenvector represents the variance in that direction. Therefore, the eigenvector corresponding to the maximum eigenvalue is the main direction of the point cloud data, that is, the normal direction. Solve the eigenvalues and eigenvectors of the covariance matrix:
[0089] Cv=λv;
[0090] Where: v is the eigenvector; λ is the eigenvalue.
[0091] The eigenvector corresponding to the largest eigenvalue is selected as the normal direction.
[0092] Through the above steps, the PCA algorithm can calculate the normal direction of each point in the point cloud dataset.
[0093] Set the neighborhood point n=5 and perform direction consistency processing on the normal vector.
[0094] According to the analysis results of the point cloud data, the threshold is set to 0.85~1.000 in the Z direction, and the threshold in the X direction is set to -1.00~-0.99 and 0.99~1.00. The point clouds at the bottom and both sides of the arch rib are extracted respectively. After processing, the point cloud data has 865105 points, as shown in Figure 3 shown.
[0095] In this embodiment, the initial extraction of the point clouds at the bottom and both sides of the arch rib has been achieved based on the normal features of the point cloud. However, due to the complex structure of the steel arch, there are still many discrete points in the extraction results, points on the arch bridge deck, cross braces, cross links and web members, and points on the retaining wall, which will affect the linear extraction of the arch bridge. In order to remove the interference of these points on the extraction results, the present invention uses the RANSAC algorithm to perform straight line fitting classification and remove some discrete points, and performs DBSCAN clustering on the fitted straight line to remove the points on the arch bridge deck, cross braces, cross links and web members, and points on the retaining wall.
[0096] The improved DBSCAN algorithm is used to perform secondary extraction on the preliminary extraction results, including:
[0097] a. Fit and segment the first straight line: Use the RANSAC algorithm to fit a straight line in the current point cloud, find the inner points of the straight line (i.e., points that conform to the straight line model), extract these inner points and remove them from the original point cloud; where the straight line can be expressed as: L(A, B): y = A + Bx;
[0098] b. Continue fitting the remaining point clouds: Apply the RANSAC algorithm again on the remaining point clouds to fit the next straight line;
[0099] c. Repeat the above steps until the number of points in the remaining point cloud is less than min_num;
[0100] d. Determine the line to which the point belongs: In each RANSAC algorithm fitting, the distance from each point to the fitted line is calculated to determine whether it belongs to the current line (i.e., whether it is an inlier). Only inliers will be segmented and removed from the original point cloud;
[0101] e. Represent each line as a set of eigenvectors; for example, the eigenvectors are as follows:
[0102]
[0103] f. Use the DBSCAN algorithm to cluster the feature vectors of these lines. In a given ∈-neighborhood, if the number of neighbors of a line is greater than or equal to MinPts, then this line is considered to be a core line. The core line and the lines that can be reached by it form a cluster.
[0104] g. Continue to cluster the remaining straight lines until all the straight lines are classified. The comparison of the effects after clustering is shown in FIG7 .
[0105] h. Screen the point cloud of the bridge part according to the classification result, delete all kinds of interference point clouds, and only keep the point cloud of the arch rib; wherein the screening result can be shown in FIG4.
[0106] i. Extract and fit the coordinates of key points of the filtered point cloud. Key points refer to points with special meanings in the point cloud, such as corner points, edge points or feature points. Extract these key points and fit their coordinates, identify these key points from the point cloud, and use mathematical models or fitting methods (such as curve fitting or surface fitting, etc.) to estimate their precise positions.
[0107] Among them, the RANSAC algorithm is an iterative algorithm for estimating mathematical model parameters, especially suitable for situations where there is a lot of noisy data. It is a robust parameter estimation algorithm that can estimate model parameters from a set of data containing a large number of outliers. Its basic principles are as follows:
[0108] Given a two-dimensional data set containing N points Randomly select two points from the data set, because the smallest subset of data for a straight line fit is two points, and these two points define a candidate straight line. For these two randomly selected points (x1, y1) and (x2, y2), the equation of the line can be written as:
[0109] y=mx+c
[0110] Among them, the slope m and intercept c can be calculated by the following formula:
[0111]
[0112] c=y1-mx1
[0113] For each data point (x i ,y i ) to evaluate the internal point and calculate its vertical distance to the fitted line:
[0114]
[0115] Set a threshold d th , if d i ≤d th , then the point is considered an inlier. Then the fitting model is evaluated and the number of inliers is calculated. If the number of inliers exceeds the preset threshold (such as the proportion of inliers exceeds a certain percentage), it is considered that a reasonable straight line has been found. The straight line model with the largest number of inliers is retained. Repeat the above steps k times (the number of iterations can be adjusted according to the actual situation and the size of the data set), and select the straight line model with the most inliers as the final fitting result.
[0116] The DBSCAN algorithm is a density-based clustering algorithm, which is particularly suitable for discovering clusters of arbitrary shapes and processing noisy data. The basic idea of DBSCAN is to define clusters by density, that is, within a cluster, the density of points must be greater than a certain threshold. Its basic principle and algorithm flow are as follows:
[0117] Given a point p and radius ∈, the ∈-neighborhood of p consists of all points whose distance from p is no more than ∈.
[0118] N ∈ (p)={q∈D|dist(p,q)≤∈}
[0119]
[0120] Where: D is the data set; dist(p,q) is the distance between point p and point q (usually Euclidean distance is used).
[0121] If the ∈-neighborhood of a point p contains at least MinPts points (including p itself), then p is called a core point; if a point p itself is not a core point, but it falls within the ∈-neighborhood of a core point, then p is called a boundary point; if a point p is neither a core point nor a boundary point, then p is called a noise point. Core point determination:
[0122] N ∈ (p)>MinPts
[0123] Select any unvisited point. If it is a core point, take it as the starting point and find all directly density-reachable points based on the ∈-neighborhood to form a cluster. If it is a boundary point or a noise point, mark it as noise (if it is included in the ∈-neighborhood of other points in subsequent steps, it may be re-marked as a boundary point). Starting from the core point, recursively add all directly density-reachable points to the current cluster. If point p is a core point and point q is in the ∈-neighborhood of p, then q is directly density-reachable from p. Repeat the above steps until all points are visited.
[0124] Through the above processing, (1) Structural segmentation: RANSAC can effectively fit and segment points with specific geometric shapes, thereby providing structural constraints during preliminary segmentation. (2) Reducing the impact of noise: RANSAC can effectively reduce its sensitivity to noise points through iteration and model verification.
[0125] For steel arch bridges with complex structures but obvious straight line structures, the RANSAC algorithm can effectively extract these structures and constrain the clustering results of the DBSCAN algorithm. The extracted straight lines are further segmented and clustered through DBSCAN. The final clustering results can effectively separate discrete points and the bridge deck, cross braces, cross links and web members, and accurately extract the point cloud of the key parts of the arch ribs of steel arch bridges.
[0126] The present invention also relates to a steel arch bridge arch rib linear shape extraction system based on an improved DBSCAN algorithm. The extraction system corresponds to the steel arch bridge arch rib linear shape extraction method of the above embodiment, and can be understood as a system for implementing the above steel arch bridge arch rib linear shape extraction method. The system includes a collection unit, a preprocessing unit, a preliminary extraction unit, and a secondary extraction unit.
[0127] The acquisition unit is used to acquire point cloud data of the steel arch bridge to obtain original point cloud data;
[0128] The preprocessing unit is used to preprocess the original point cloud data to obtain preprocessed point cloud data;
[0129] The preliminary extraction unit is used to perform preliminary extraction on the preprocessed point cloud data to obtain a preliminary extraction result;
[0130] The secondary extraction unit is used to perform secondary extraction on the preliminary extraction result by adopting the improved DBSCAN algorithm to obtain the final extraction result.
[0131] In order to better understand the technical solution of the present invention, the present invention is further described as follows:
[0132] If DBSCAN clustering is directly performed on the point cloud, the extraction of the point cloud of the arch rib cannot be realized. For the point cloud on both sides of the arch rib, only a very small number of discrete points are filtered out, but the connection components between the arch ribs are powerless; there is no effect at all on the point cloud at the bottom of the arch rib, and even some data points are filtered out. The improved DBSCAN algorithm of the present invention, which integrates RANSAC, can well achieve the purpose of deleting discrete points and the point cloud of the bridge deck and crossbar.
[0133] In the results of different parameters, for the point clouds on both sides of the arch rib, when the minimum number of fitting points of RANSAC straight line fitting is 25, a large number of useful data points are filtered out. When the minimum number of fitting points of RANSAC straight line fitting is 15 or 20, the extraction effect is better, and the point cloud of the arch rib part can be separated. For the point cloud at the bottom of the arch rib, when the minimum number of fitting points of RANSAC straight line fitting is 25, almost all point clouds are filtered out. When the minimum number of fitting points of RANSAC straight line fitting is 15 or 20, the extraction effect is better. Among them, when the minimum number of fitting points of RANSAC straight line fitting is 20, in addition to the cross brace, the cross link and the web can also be filtered out, and the effect is the best.
[0134] According to the point cloud extracted by the improved DBSCAN algorithm fused with RANSAC of the present invention, key point fitting is performed, and compared with the manual extraction results, the key point errors of the point cloud at the bottom and both sides of the arch rib are all at the millimeter level, wherein the maximum difference of the point cloud extracted at the bottom of the arch rib is 9.2mm, the minimum difference is 0.1mm, the absolute value of the difference is below 5.0mm accounts for 92.05%, and the absolute value of the difference is below 2.0mm accounts for 82.95%; the maximum difference of the point cloud extracted at both sides of the arch rib is 8.0mm, the minimum difference is 0.1mm, the absolute value of the difference is below 5.0mm accounts for 96.74%, and the absolute value of the difference is below 2.0mm accounts for 56.82%. The present invention can accurately and quickly realize the automatic extraction of the arch rib linear shape that overcomes the influence of the complex structure of the steel arch bridge, and the extraction accuracy can meet the actual production needs.
[0135] The present invention provides an objective and scientific method and system for extracting the linear shape of the arch rib of a steel arch bridge. The method uses three-dimensional laser scanning technology and PCA algorithm to obtain point cloud normal data. Based on the normal line characteristics and structural characteristics of the steel arch bridge, the three-dimensional point cloud data of the steel arch bridge collected by the three-dimensional laser scanner is processed to extract the normal line feature data. The point cloud is extracted based on the different vertical characteristics of the normal lines at different positions of the steel arch bridge, and the characteristic point cloud of the key parts of the arch rib is obtained. The linear shape extraction of the steel arch bridge can be completed more accurately and effectively. With the normal line characteristics and structural characteristics of the arch rib of the steel arch bridge itself, the linear shape extraction accuracy reaches the millimeter level by using the three-dimensional laser point cloud, which greatly reduces the manpower consumption and time consumption. The process is simpler, the results are more accurate, and the complex structure of the steel arch bridge has better robustness, which can well meet the actual production needs.
[0136] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.
Claims
1. A method for extracting the linear shape of steel arch bridge ribs based on an improved DBSCAN algorithm. Features: include: Collect point cloud data of steel arch bridge to obtain original point cloud data; Preprocessing the original point cloud data to obtain preprocessed point cloud data; Perform preliminary extraction on the preprocessed point cloud data to obtain preliminary extraction results; The improved DBSCAN algorithm is used to perform secondary extraction on the preliminary extraction results to obtain the final extraction results.
2. According to the method for extracting the linear shape of the steel arch bridge rib based on the improved DBSCAN algorithm in claim 1, Features: The preprocessing includes downsampling and denoising; The downsampling specifically includes: downsampling the point cloud data using an Octree algorithm, taking the centroid of each Octant as a representative point, and constructing a new downsampled point cloud; The denoising specifically includes: denoising the downsampled point cloud data using bilateral filtering.
3. According to the method for extracting the linear shape of the steel arch bridge rib based on the improved DBSCAN algorithm in claim 1, Features: Perform preliminary extraction on the preprocessed point cloud data, including: Use the PCA algorithm to estimate the point cloud normal, select the eigenvector corresponding to the maximum eigenvalue as the normal direction, and obtain the normal direction of each point in the point cloud data set; Set the neighborhood point n = n 0 , the direction of the normal vector is consistent; According to the analysis results of the point cloud data, thresholds are set in the Z direction and the X direction to extract the point clouds at the bottom and both sides of the arch rib respectively.
4. The method for extracting the linear shape of the steel arch bridge rib based on the improved DBSCAN algorithm according to claim 1, Features: The improved DBSCAN algorithm is used to perform secondary extraction on the preliminary extraction results, including: a. Fit and segment the first line: Use the RANSAC algorithm to fit a line in the current point cloud, find the inner points of the line, extract these inner points and remove them from the original point cloud; b. Continue fitting the remaining point clouds: Apply the RANSAC algorithm again on the remaining point clouds to fit the next straight line; c. Repeat the above steps until the number of points in the remaining point cloud is less than min_num; d. Determine the line to which the point belongs: In each RANSAC algorithm fitting, the distance from each point to the fitted line is calculated to determine whether it belongs to the current line. Only the inner points will be segmented and removed from the original point cloud; e. Represent each straight line as a set of feature vectors; f. Use the DBSCAN algorithm to cluster the feature vectors of these lines. In a given ∈-neighborhood, if the number of neighbors of a line is greater than or equal to MinPts, then this line is considered to be a core line. The core line and the lines that can be reached by it form a cluster. g. Continue clustering the remaining lines until all lines are classified; h. Filter the point clouds of the bridge part according to the classification results, delete all kinds of interference point clouds, and only keep the point clouds of the arch ribs; i. Extract and fit the key point coordinates of the filtered point cloud.
5. A steel arch bridge rib linear extraction system based on improved DBSCAN algorithm, Features: It includes a collection unit, a pre-processing unit, a primary extraction unit and a secondary extraction unit; The acquisition unit is used to acquire point cloud data of the steel arch bridge to obtain original point cloud data; The preprocessing unit is used to preprocess the original point cloud data to obtain preprocessed point cloud data; The preliminary extraction unit is used to perform preliminary extraction on the preprocessed point cloud data to obtain a preliminary extraction result; The secondary extraction unit is used to perform secondary extraction on the preliminary extraction result by adopting the improved DBSCAN algorithm to obtain the final extraction result.
6. According to claim 5, the steel arch bridge arch rib linear extraction system based on the improved DBSCAN algorithm, Features: The preprocessing includes downsampling and denoising; The downsampling specifically includes: downsampling the point cloud data using an Octree algorithm, taking the centroid of each Octant as a representative point, and constructing a new downsampled point cloud; The denoising specifically includes: denoising the downsampled point cloud data using bilateral filtering.
7. According to claim 5, the steel arch bridge arch rib linear extraction system based on the improved DBSCAN algorithm, Features: Perform preliminary extraction on the preprocessed point cloud data, including: Use the PCA algorithm to estimate the point cloud normal, select the eigenvector corresponding to the maximum eigenvalue as the normal direction, and obtain the normal direction of each point in the point cloud data set; Set the neighborhood point n = n 0 , the direction of the normal vector is consistent; According to the analysis results of the point cloud data, thresholds are set in the Z direction and the X direction to extract the point clouds at the bottom and both sides of the arch rib respectively.
8. According to claim 5, the steel arch bridge arch rib linear extraction system based on the improved DBSCAN algorithm, Features: The improved DBSCAN algorithm is used to perform secondary extraction on the preliminary extraction results, including: a. Fit and segment the first line: Use the RANSAC algorithm to fit a line in the current point cloud, find the inner points of the line, extract these inner points and remove them from the original point cloud; b. Continue fitting the remaining point clouds: Apply the RANSAC algorithm again on the remaining point clouds to fit the next straight line; c. Repeat the above steps until the number of points in the remaining point cloud is less than min_num; d. Determine the line to which the point belongs: In each RANSAC algorithm fitting, the distance from each point to the fitted line is calculated to determine whether it belongs to the current line. Only the inner points will be segmented and removed from the original point cloud; e. Represent each straight line as a set of feature vectors; f. Use the DBSCAN algorithm to cluster the feature vectors of these lines. In a given ∈-neighborhood, if the number of neighbors of a line is greater than or equal to MinPts, then this line is considered to be a core line. The core line and the lines that can be reached by it form a cluster. g. Continue clustering the remaining lines until all lines are classified; h. Filter the point clouds of the bridge part according to the classification results, delete all kinds of interference point clouds, and only keep the point clouds of the arch ribs; i. Extract and fit the key point coordinates of the filtered point cloud.
Citation Information
Patent Citations
Three-dimensional curved surface reconstruction method for coal mine tunnel arch surface
CN114399603A
Tea leaf tender shoot positioning method based on 3D point cloud topological structure
CN117078926A
Pavement disease three-dimensional detection method based on grid mapping
CN117974597A
Single building three-dimensional reconstruction method based on point cloud semantic segmentation and structure fitting
WO2024077812A1
